Spectroscopic Properties of KPb₂Cl₅ and RbPb₂Br₅ Doped with Er³⁺ and Yb³⁺

Abstract

Alkali halide crystals doped with Er³⁺ exhibit several properties that make them promising for practical use in optical fibers and converters of the emission spectrum. In this work, we studied the luminescence spectra of crystals as a function of the excitation radiation wavelength and the temperature dependences of the luminescence intensity in the up-conversion mechanism of excitation of the most intense lines (530, 550, and 650 nm) and upon excitation at 980 nm. In KPb₂Cl₅:Er³⁺, RbPb₂Br₅:Er³⁺, and KPb₂Cl₅:Er³⁺ + Yb³⁺ crystals, the luminescence kinetics were measured upon excitation at 313 and 365 nm, and the theoretical estimates of lifetimes were made using the Judd–Ofelt theory. The results can be used in the development of laser-based systems and for estimating the degree of approximation of calculations by the Judd–Ofelt method for these crystals.

1 INTRODUCTION

Alkali lead halide crystals, (K-Rb)Pb₂(Cl-Br)₅, that are promising as laser media were investigated. These crystals are technologically advanced, have high chemical resistance and a vast transparency region, have a somewhat loose structure that enables rare earth ions (REIs) of type Dy³⁺, Tb³⁺, Yb³⁺, Er³⁺, or Nd³⁺ to be included in it. They can effectively serve as converters of light emission from one wavelength range to another because of the narrowness of their phonon spectrum (~200 cm^–1 for KPb₂Cl₅ and 144 cm^–1 for RbPb₂Br₅). The crystals can be grown large enough and have a high optical quality for use in solid-state lasers. Optical transitions in KPb₂Cl₅:RE³⁺ crystals are characterized by high oscillator strengths and a large number of radiative transitions [1].

Up-conversion pumping can be used to obtain laser radiation from high excited REI states when pumped by a laser diode (LD). Up conversion of the excitation energy during LD pumping is possible both as a result of stepwise excitation (successive absorption of the pump energy from the ground and excited states) and in the nonradiative interaction of the excited states of the activator ions. These effects yield laser radiation in the visible region [2] and significantly increase the efficiency of laser radiation in the IR region [3]. From this point of view, crystals with narrow phonon spectra, in which the loss for nonradiative relaxation is minimal, can be considered as the most effective with appropriate selection of the emission spectrum of laser diodes.

2 EXPERIMENTAL

KPb₂Cl₅ crystals were grown by the Stockbarger method in the Institute of Geology and Mineralogy of the Siberian Branch of the Russian Academy of Sciences (Novosibirsk). The initial potassium and lead chlorides (high-purity grade) were subjected to multiple purifications by zone melting in the presence of a chlorine agent. KPb₂Cl₅ crystals were grown in evacuated quartz ampoules. The growth unit was a vertical two-zone furnace operated with a temperature g-radient of approximately 50°C/cm. The crystal growth rate was 2–4 mm/week. The size of the crystals obtained was 15 mm in diameter and 30 mm in length [4].

The photoluminescence (PL) and photoluminescence excitation spectra were measured using a DDS-400W deuterium lamp and a DMR-4 double prism monochromator; a GaAs laser diode (980 nm, 100 mW) was used for the excitation of anti-Stokes luminescence (ASL). The photoluminescence was recorded using an MDR-23 monochromator and an R6358-10 photomultiplier (Hamamatsu). The photoluminescence excitation spectra were normalized to an equal number of photons incident on a sample using Lumogen Yellow, which has a constant quantum yield in the spectral range under study. The photoluminescence decay kinetics were measured using a pulsed xenon lamp (FWHM = 1 μs, frequency 5 Hz). X-ray luminescence (XRL) spectra were measured using a URS-55A apparatus (Cu, 30 kV, 10 mA), an MDR-23 monochromator, and an FEU-106 photomultiplier as a detector.

The application of the Judd–Opelt theory for calculating the luminescence times in KPb₂Cl₅ crystals doped with Er³⁺ ion was considered, and calculation corrections were introduced for RbPb₂Br₅ crystals (the crystals have a structure similar to KPb₂Cl₅, which simplifies the necessary corrections). The calculation of lifetimes for crystals doped simultaneously with Er³⁺ and Yb³⁺ is complicated, because it requires the application of the method of model quantum-mechanical calculation, namely, the determination of energy transfer between ions, which was not taken into account in the method used. A more detailed description of the method can be found in [5]. The oscillator strengths of the magnetic dipole interaction, which are necessary for applying the method, are taken from [6].

3 RESULTS AND DISCUSSION

The luminescence spectra of crystals (Fig. 1) were studied under excitation at a wavelength of 313 nm (excitation of the exciton states near the band gap) and 365 nm (transition ⁴I_15/2 → ²G_11/2) and upon anti-Stokes excitation at 980 nm (transitions ⁴I_15/2 → ⁴I_11/2, ⁴I_11/2 → ⁴F_3/2).

Fig. 1.

Photoluminescence spectrum of KPb₂Cl₅:Er³⁺ at λ_exc = (a) 313, (b) 365, and (c) 980 nm. The following transitions are indicated: (1) ⁴G_11/2 → ⁴I_15/2, (2) ²H(G)_9/2 → ⁴I_15/2, (3) ⁴F_7/2 → ⁴I_15/2, (4) ⁴G_11/2 → ⁴I_13/2, (5) ²H_11/2 → ⁴I_15/2, (6) ⁴S_3/2 → ⁴I_15/2, (7) ²H(G)_9/2 → ⁴I_13/2, (8) ⁴F_3/2 → ⁴I_13/2, (9) ⁴F_9/2 → ⁴I_15/2, and (10) ²H(G)_9/2 → ⁴I_11/2.

Excitation at 313 nm brings the Er³⁺ ion into the ²G_7/2 state, and therefore, radiative relaxation to the f levels becomes possible and nonradiative transitions to the underlying ⁴G_9/2, ⁴G_11/2, ²H(G)_9/2 levels may occur. It is seen in the corresponding luminescence spectrum that emission occurs only from the ²H(G)_9/2 level.

A significant change in the shape of the spectrum is observed upon excitation at 365 nm, which can be fully described in terms of multiphonon relaxation. The most intensive transitions are ⁴G_11/2 → ⁴I_15/2 and ⁴G_11/2 → ⁴I_13/2; excitation brings the ion to the ²H(G)_9/2 level. Nonradiative transitions are largely manifested because the density of levels near the ²H(G)_9/2 level is relatively high. These transitions virtually completely extinguish the transitions observed upon excitation at 313 nm.

The excitation by a GaAs diode (980 nm) leads to the anti-Stokes process, that is, ⁴I_15/2 → ⁴I_11/2 and ⁴I_11/2 → ⁴F_3/2. Multiphonon nonradiative transitions may occur due to the relative proximity of the ⁴F_3/2 level with the ²H(G)_9/2 and ⁴F_5/2 levels. Note that the inverse transition of ⁴F_3/2 → ⁴I_15/2, which should be observed because of a significant excitation of the ⁴F_3/2 level, proceeds to a minimal degree (see the peak at 430 nm). This can be explained by the substantial lifetime of the Er³⁺ ion in this state; therefore, radiative transitions are much less likely.

The excitation wavelengths were selected using the measured excitation spectra of the most significant lines at 550 and 650 nm (Fig. 2).

Fig. 2.

Luminescence excitation spectra of (a) KPb₂Cl₅:Er³⁺ and (b) RbPb₂Br₅:Er³⁺ crystals; excitation of luminescence at a wavelength of (solid curve) 650 and (dashed curve) 550 nm.

A comparison of the KPb₂Cl₅ and RbPb₂Br₅ spectra (Fig. 3) leads to the conclusion that the luminescence from the ⁴S_3/2 and ²H(G)_9/2 levels begins to compete with the luminescence caused by the transition of ²H_11/2 → ⁴I_15/2. This effect can be explained by the difference in the phonon energies in the matrices.

Fig. 3.

Luminescence spectra of (thin curve) KPb₂Cl₅:Er³⁺ and (bold curve) RbPb₂Br₅:Er³⁺ at (a) 313, (b) 365, and (c) 980 nm.

Upon excitation at 313 nm, the luminescence of the self-trapped exciton is observed in RbPb₂Br₅ crystals. The effect is associated with a smaller band gap in the RbPb₂Br₅ matrix compared with the band gap of the KPb₂Cl₅ matrix, caused by the replacement of the chlorine ion with the bromine ion. Since the levels of the self-trapped exciton are fixed relative to the bottom of the conduction band, this energy achieved at 313 nm is sufficient to excite an exciton in a RbPb₂Br₅ crystal. The exciton glow actually extinguishes the glow of Er³⁺.

A residual exciton emission is observed upon excitation at 365 nm. A relatively broad band at 650–700 nm is due to the ⁴F_3/2 → ⁴I_13/2, ⁴F_7/2 → ⁴I_13/2, and ⁴F_9/2 → ⁴I_15/2 transitions. The ⁴F_7/2 and ²H_11/2 levels are located close to each other; the ²H_11/2 level is populated by a multiphonon mechanism, and therefore, luminescence by transition ²H_11/2 → ⁴I_15/2 is observed (line 490 nm). In the case of KPb₂Cl₅, this transition is much less likely, which is explained by a smaller population of the ⁴F_7/2 level.

Upon excitation at 980 nm, the intensity of the ⁴F_7/2 → ⁴I_15/2 transition increases significantly, which is probably due to lower phonon energies in the RbPb₂Br₅ matrix (133 cm^–1) compared to KPb₂Cl₅ (200 cm^–1). Consequently, the ⁴F_3/2 → ⁴F_5/2 transition and the subsequent relaxation of ⁴F_5/2 → ⁴F_7/2 are less intense; the corresponding intensities of the spectral lines at 525 and 550 nm in the photoluminescence spectrum are lower for RbPb₂Br₅ than those for K-Pb₂Cl₅.

The additional doping of the KPb₂Cl₅:Er³⁺ matrix with Yb³⁺ ions significantly increases the quantum yield upon excitation at 980 nm and affects it only slightly at the other excitation energies. The integrated quantum yield upon excitation at 980 nm by a laser diode increases by approximately five to ten times; at different excitation wavelengths, the integrated quantum yield does not actually change. The luminescence spectrum of a KPb₂Cl₅:Er³⁺ + Yb³⁺ crystal (1 : 3) upon excitation at 980 nm by a laser diode actually coincides in shape with the luminescence spectrum of K-Pb₂Cl₅:Er³⁺ under the same excitation conditions (Fig. 4). This is due to the absence of levels above ⁴F_5/2 in the Yb³⁺ ion in the band gap in this matrix; the allowed energy transitions in the Yb³⁺ ion are represented only by the ⁴F_7/2 → ⁴F_5/2 transition (near 980 nm). Er³⁺ ions also have an allowed ⁴I_15/2 → ⁴I_11/2 transition, corresponding to an energy of 980 nm. Thus, if both types of ions are present in the matrix, nonradiative energy transfer from Yb³⁺ to Er³⁺ may occur. Since Yb³⁺ ions can absorb only phonons with an energy of 980 nm, their effect is only observed upon excitation at 980 nm.

Fig. 4.

Luminescence spectrum upon excitation of (thin curve) KPb₂Cl₅:Er³⁺ and (bold curve) KPb₂Cl₅:Er³⁺ + Yb³⁺.

We also note a smoother spectrum with additional doping with Yb³⁺; the smoothness is due to the significant amplification of the signal when Yb³⁺ is activated, so that the signal-to-noise ratio in the transmission path increases significantly (the integrated quantum yield increases approximately five to ten times).

The temperature dependences were measured for crystals in the temperature range of 90–420 K upon excitation by a laser diode at 980 nm. The results are shown in Fig. 5.

Fig. 5.

Luminescence spectra in KPb₂Cl₅:Er³⁺ and RbPb₂Br₅:Er³⁺ crystals at different temperatures: (a) KPb₂Cl₅:Er³⁺, (b) KPb₂Cl₅:Er³⁺ + Yb³⁺, and (c) RbPb₂Br₅:Er³⁺.

As the temperature increases, the luminescence line at 490 nm in KPb₂Cl₅ is quenched, and a corresponding rise is observed for the line at 530 nm, which is associated with the population of the overlying ⁴F_5/2 level; in this case, a smaller number of photons is released upon transition from ⁴F_7/2 to the ground state.

Note the pattern found for the line at 550 nm in KPb₂Cl₅:Er³⁺ (⁴S_3/2 → ⁴I_15/2): as the temperature rises from 90 to 190 K, the line flares up. Further heating affects the intensity of the line only slightly; in fact, the line stabilizes in the temperature range of 190–400 K.

When the RbPb₂Br₅ matrix is doped with Er³⁺ ions, the lines at 490 and 530 nm become less intense at low temperatures. In this case, the population of the ⁴F_7/2 level is less likely because the phonon energy in K-Pb₂Cl₅ is higher than that in RbPb₂Br₅ (200 and 144 cm^–1, respectively). The quantum yield of Er³⁺ luminescence in the KPb₂Cl₅ matrix is higher than that in the RbPb₂Br₅ matrix by approximately five to ten times.

The time dependences of the luminescence spectra on temperature were measured for the lines 530, 550, and 650 nm (Fig. 6).

Fig. 6.

Time dependences of the luminescence intensity upon excitation at (a, c) 313 and (b, d) 365 nm in (a, b) KPb₂Cl₅:Er³⁺ and (c, d) KPb₂Cl₅:Er³⁺ + Yb³⁺ crystal.

The spectra presented show that in the case of the excitation of Er³⁺ ions in the KPb₂Cl₅ matrix (the ⁴I_15/2 → ⁴G_11/2 transition), the luminescence undergoes a burning stage with a typical burning time of approximately 0.5 ms. This is explained by the transfer of energy between the levels through intermediate channels (presumably, through the ⁴F_7/2 level).

The corresponding luminescence times were found from the kinetics obtained; the typical value of the error in the calculations was on average 2%. The experimental luminescence times were compared with those calculated by the Judd–Ofelt method (Table 1); in the case of RbPb₂Br₅, the phonon energy was renormalized.

Table 1. Luminescence times of the centers, found experimentally, depending on the wavelength of excitation and luminescence

As can be seen from the data of Table 1, the luminescence times in a crystal additionally activated with Yb³⁺ ions are in all cases smaller than the corresponding luminescence times in a crystal without a sensitizer. This effect is in accordance with the idea of the existence of an additional channel for the transfer of energy between ligand ions, the effectiveness of which can be found using the equation

$$\eta = 1 - \frac{{{{\tau }_{{D - A}}}}}{{{{\tau }_{D}}}},$$

((1))

where τ_D–A is the luminescence time in a crystal doped with Er³⁺ and Yb³⁺, and τ_D is the luminescence time in a crystal doped with only Er³⁺.

In accordance with the above, the experimental energy transfer efficiency upon excitation at 365 nm is 12% for a luminescence wavelength of 550 nm and 24% for a luminescence wavelength of 650 nm.

The theoretical luminescence times for KP-b₂Cl₅:Er³⁺ crystals are taken from [7]; for RbPb₂Br₅:Er³⁺ crystals, the recalculation to phonon energy (200 cm^–1 for KPb₂Cl₅ and 133 cm^–1 for RbPb₂Br₅) was performed in accordance with the equations from [5].

To recalculate the theoretical values of the luminescence times from the Er³⁺ levels in RbPb₂Br₅ crystals, we used the following phenomenological parameters B and α: B = 4 × 10⁹ s^–1, α = 1.2 × 10^–2 cm^–1 [7].

The recalculation of the luminescence times for KPb₂Cl₅:Er³⁺ + Yb³⁺ is complex and has not been done in this work, since it requires taking into account energy transfer between ions by different mechanisms.

The calculation presented proves that the results given by the theory predict relatively correctly the behavior of luminescent centers for the excitation relaxation from the ⁴S_3/2 and ⁴F_9/2 levels and give incorrect predictions for the excitation relaxation from the ²H_11/2 level. This pattern is preserved when recalculated to the phonon energy, which was done in this work.

4 CONCLUSIONS

The luminescence spectra of crystals were measured at different excitations (zone–zone transitions, transitions between the energy states of doping Er³⁺ and Yb³⁺ ions) with full identification of all observed transitions in the impurity center. Based on the experimental data, we demonstrated that KPb₂Cl₅:Er³⁺ crystals could be used as efficient energy converters from one range to another. On the other hand, RbPb₂Br₅:Er³⁺ crystals do not exhibit high conversion efficiency due to the exciton luminescence, and their use as converters is not recommended.

The anti-Stokes luminescence spectra of doping ions were measured at various temperatures (excitation by a laser diode at 980 nm), and the temperature dependences of the spectra were found. It is shown that in the KPb₂Cl₅ matrix, the populations of the ⁴F_7/2 and ²H_11/2 levels are correlated. In KPb₂Cl₅:Er³⁺, the effect of stabilization of the luminescence intensity was observed at 550 nm in the temperature range of 190–420 K. No such effect was found for codoping with Er³⁺ and Yb³⁺.

The luminescence kinetics were measured for the most intense bands upon excitation at 313 and 365 nm; the corresponding luminescence times of center are obtained. In a KPb₂Cl₅:Er³⁺ crystal, the luminescence build-up was observed upon excitation at 365 nm, which is evidence of the population of the level through intermediate states (presumably, through ⁴F_7/2). In crystals additionally activated with Yb³⁺, a decrease in the luminescence time was detected, which is because of the presence of an additional channel of energy transfer from the sensitizer to the activator.

The Judd–Ofelt method and the theory of energy transfer in the form of multiphonon nonradiative relaxation are considered with respect to the lifetimes of the centers. The results calculated by the Judd–Ofelt method are similar to the experimental data on the order of magnitude for luminescence from the ⁴S_3/2 and ⁴F_9/2 levels. For luminescence from the ⁴F_7/2 level, the difference between the theory and experiment is significant, which may be due to a nonproper selection of the model parameters for the calculation. The values calculated for RbPb₂Br₅ correlate with the calculations for KPb₂Cl₅; the correction takes into account the difference in the phonon energy of the lattice: 200 cm^–1 for KPb₂Cl₅ and 133 cm^–1 for RbPb₂Br₅.