Ambient and high pressure spectroscopy of Ce<sup>3+</sup> doped yttrium gallium garnet

A. Wittlin; H. Przybylińska; M. Berkowski; A. Kamińska; P. Nowakowski; P. Sybilski; Chong-Geng Ma; M.G. Brik; A. Suchocki

doi:10.1364/OME.5.001868

1. Introduction

Yttrium gallium garnet, with the chemical formula Y₃Ga₅O₁₂ (YGG), doped with several rare-earth ions exhibits efficient luminescence and can be used in phosphor applications [1]. Multicomponent garnets containing yttrium are used for preparation of the most successful scintillators, serving as materials for band gap engineering [2, 3 ].

YGG single crystals are produced by the Czochralski method, however they are relatively hard to grow by this method due to the relatively high volatility of Ga₂O₃. YGG can be also synthesised in a powder form, for example by the sol-gel method [4].

The dodecahedral sites in garnets are occupied by trivalent rare-earth ions as components of the host and as dopants. Depending on the growth conditions and doping (and codoping), various types of luminescence centers of rare earth and transition metal ions occur in garnets [5, 6 ]. Besides of luminescence centers, created by substitution of rare earth or trivalent cations by various dopant ions, also additional centers exist in garnets. Very typical example of such luminescence centers, which may be found in bulk crystals produced by the Czochralski method, are the antisites i.e., rare-earth ions replacing trivalent cations at the sites with octahedral crystal field symmetry, which typically are occupied by other trivalent cations [7]. Antisite formation in garnets was first postulated on the bases of differences in lattice parameters between the crystals grown by the Czochralski method and powders produced by the solid state reaction [8–10 ]. The models of antisite defect formation in garnets were examined thoroughly in Refs [11–14 ].

Parity allowed inter-configurational 4f⁰5d¹→4f¹5d⁰ transitions are responsible for appearance of very efficient luminescence in many garnets such as yttrium aluminum (YAG) or lutetium aluminum (LuAG) garnets doped with Ce³⁺ ions. In other garnets, among them in gadolinium gallium garnet, the luminescence of Ce³⁺ ions is completely absent at ambient conditions. The reason for that is the location of the lowest 5d level of cerium in the latter case: it is resonant with the conduction band [15], which quenches the luminescence. Application of pressure restores Ce³⁺ luminescence in this garnet. The lack of luminescence at ambient pressure significantly hinders investigations of the optical properties of Ce³⁺ multicenters in these materials, limiting their observation only to the absorption studies. Application of pressure allows including luminescence to the available investigation techniques of these crystals.

The Ce-doped YGG crystals grown by the Czochralski method are the members of that family of garnets, which does not exhibit Ce³⁺ luminescence at ambient pressure. However, luminescence of Ce³⁺ has been observed previously at ambient pressure in YGG powders formed by solid state reaction [16]. This might be related to the slightly lower lattice parameter of the powder as compared to the bulk crystal. Such an effect has been observed previously for several garnets and it was associated with more efficient antisite formation in bulk crystals grown by the Czochralski method than in powders manufactured at much lower temperatures [8, 17 ]. Spectroscopic evidences of Ce³⁺ multicenters and existence of antisites in bulk crystals and epitaxial layers have been presented by us previously for the gadolinium gallium (GGG) [18], YAG [19], and LuAG garnets [20]. We have also studied the emergence of Ce³⁺ luminescence in GGG crystals under influence of hydrostatic pressure [15].

This paper is dedicated to low temperature infrared optical transmission studies of YGG single crystals for investigation of the Ce³⁺ multicenters. Also the effect of pressure-induced luminescence of Ce³⁺ is demonstrated in this paper.

2. Experimental details

The YGG:Ce bulk crystals used in our study were grown by the Czochralski method using of 5N raw materials. The crystals contained about 0.5 mol. % of Ce.

The infrared absorption spectra were measured with a BOMEM DA3 Fourier-Transform Infrared (FTIR) spectrophotometer, with resolution of 1 cm⁻¹. For low temperature measurements the samples were placed on a cold finger of Oxford Instruments CF-102 continuous-flow cryostat equipped with KBr windows.

High-pressure PL spectra were measured using a 457.9-nm line of the Coherent Inova 400 Ar-ion laser as the excitation source. The spectra were dispersed by a Horiba Jobin-Yvon FHR 1000 monochromator. The signal was detected by means of a liquid nitrogen cooled charge-coupled device (CCD) camera. The high-pressure measurements were performed using a low-temperature diamond anvil cell (CryoDAC LT, easyLab Technologies Ltd). Argon was used as a pressure transmitting medium. The DAC was mounted in an Oxford Optistat CF cryostat equipped with a temperature controller for low-temperature measurements. The samples, cut and polished down to a thickness of 20 μm, were loaded into the cell along with a small ruby crystal. The R1-line ruby luminescence was used for pressure calibration. The ambient-pressure absorption spectra were measured with a Cary 5000 UV-Vis-NIR spectrophotometer.

3. Experimental results and discussion

3.1. Ce³⁺ energy levels structure in YGG

The energy structure of the Ce³⁺ ions with the 4f ¹ configuration consists of the ²F_5/2 ground state and the ²F_7/2 excited state, which originate from the ²F term due to spin-orbit interaction. These two states are split further into three (levels #1-3) and four (levels #4-7) energy levels, respectively, by the crystal field (CF) in the D₂ point group symmetry.

The low temperature infrared absorption spectrum of the YGG:Ce (0.5 mol. %) crystal taken at 13 K in the region of the 4f-4f transitions is depicted in Fig. 1 . Background absorption was subtracted from the spectra. The three groups of lines in Fig. 1 located at around 2175, 2347, and 3651 cm⁻¹ are assigned to the optical transitions between the lowest-lying level (#1) of the ground ²F_5/2 state and levels #4, #5, and #7 of the ²F_7/2 excited state of Ce³⁺. Several lines in each group testify that a few Ce³⁺ - related centers can be seen in the spectrum. The low intensity transitions between levels #1 and #6 are not visible in the spectra.

Fig. 1 YGG:Ce absorption spectrum measured at 13 K in the region of Ce³⁺ 4f → 4f transitions.

Download Full Size | PDF

Multicenter structure of Ce³⁺ dopant in YGG can be better detected in the vicinity of the strongest absorption associated with the transitions from the levels #1 to levels #7 of various Ce³⁺-related centers, shown in Fig. 2 .

Fig. 2 Absorption spectrum of the YGG:Ce crystal, measured at 13 K in the region of Ce³⁺ 4f → 4f transitions between levels #1 and #7.

Download Full Size | PDF

Except the major line peaked at 3652 cm⁻¹, which we associate with the Ce³⁺ ions at regular yttrium dodecahedral positions, at least two other lines are visible at the side of this line at 3685 and 3725 cm⁻¹. Full width at half maximum of the particular lines in the spectrum is between 6 cm⁻¹ for the line at 3725 cm⁻¹ and 14 cm⁻¹ for the major line at 3652 cm⁻¹.

Ce³⁺ multisites in the examined crystal occur with relatively large concentrations as compared to the concentration of the major Ce³⁺ center, i.e. single Ce³⁺ ions in dodecahedral sites. Although the value of the absorption coefficient alone cannot be treated as a measure of concentration of the particular center without knowledge of their relative oscillator strengths, the total absorption of the two lines associated with Ce³⁺ multisites to the absorption of the major line at 3652 cm⁻¹, equal to about 12%, is a relatively large value. Nevertheless, the numbers and concentrations of Ce multisites seem to be smaller that observed previously by us in GGG [18], YAG [19], and LuAG [20] doped with Ce, where even up to seven additional Ce³⁺ centers have been found.

The authors of Refs [11–14 ]. proposed some plausible models of the Ce-related defect centers in garnets. Recently, A. Lupei et al. [21] suggested that the origin of the Ce³⁺ multicenters may be associated with antisites of those rare-earth ions, which are constituents of the host crystal (and Ce³⁺ in regular dodecahedral positions). Large Ce³⁺ ions (ionic radius of 1.01 Å [22] in the VI-fold coordination) seem to be too large to substitute much smaller sized cations at the octahedral sites in garnet host. On the other hand it has been shown using the electron paramagnetic resonance technique that Ce³⁺ antisites are formed in Lu₃Al₅O₁₂ crystals with an even smaller lattice parameter than Y₃Ga₅O₁₂ [23]. In this material, in addition to Ce_Lu ³⁺ ions at regular Lu positions, Ce_Al ³⁺ are formed, apart from two other types of Ce_Lu ³⁺ centers strongly interacting with Lu_Al ³⁺ antisite defects [6]. However, based solely on spectroscopic results it is difficult to assign particular lines in the spectrum to the proposed model center. Additional studies will be required to identify the nature of the particular centers.

A smaller number of Ce³⁺ related antisites is observed in the examined YGG crystal than in the previously studied similar GGG:Ce system. This may be related partially to relatively low concentration of Ce in YGG as compared with GGG measured by us in the past. This would suggest that at least a part of Ce- related multicenters originate from Ce-pairs or larger complexes in the material with higher concentration of dopant. Possibly, systematic studies of Ce multicenter occurrence as a function of dopant concentration could help to identify such complexes in garnets.

Room temperature absorption spectrum of YGG:Ce in the region of 2000 – 4000 cm⁻¹ is shown in Fig. 3 .

Fig. 3 Room temperature absorption spectrum in the region of 4f-4f transitions of YGG:Ce(0.5%) crystal.

Download Full Size | PDF

A very strong absorption in the region of transitions to the level #7 of the excited ²F_7/2 state is visible. This absorption consists of two lines, separated by about 159 cm⁻¹, in which the lower energy line is thermally populated and associated with the transitions from the level #2 of the ground ²F_5/2 state to the level #7 of the excited state.

Our observations of the positions of the levels # 2, #5, and #7 are slightly different than previously reported by Herrmann et al. [24].

Theoretical calculations of the energy levels for the main Ce³⁺ center and crystal field parameters (CFP) for the regular (dodecahedral) site with D ₂ symmetry in GGG and YAG performed using the exchange charge model are published in Ref [18]. Here, to analyze the Ce³⁺ spectra in YGG, we use the same model, which is briefly described below. The energy levels of the 4f¹ and 5d¹ configurations of Ce³⁺ ions in YGG can be found as the eigenvalues of the following crystal field Hamiltonians:

\begin{array}{l} H {(4f}^{1}) = E_{a v g} + ζ_{4 f} s_{f} \cdot l_{f} + B_{0}^{2} (f) C_{0}^{(2)} + B_{2}^{2} (f) (C_{2}^{(2)} + C_{- 2}^{(2)}) \\ + B_{0}^{4} (f) C_{0}^{(4)} + B_{2}^{4} (f) (C_{2}^{(4)} + C_{- 2}^{(4)}) + B_{4}^{4} (f) (C_{4}^{(4)} + C_{- 4}^{(4)}), \\ + B_{0}^{6} (f) C_{0}^{(6)} + B_{2}^{6} (f) (C_{2}^{(6)} + C_{- 2}^{(6)}) + B_{4}^{6} (f) (C_{4}^{(6)} + C_{- 4}^{(6)}) \end{array}

and

\begin{array}{l} H {(5d}^{1}) = Δ_{E} (fd) + ζ_{5 d} s_{d} \cdot l_{d} + B_{0}^{2} (d) C_{0}^{(2)} + B_{2}^{2} (d) (C_{2}^{(2)} + C_{- 2}^{(2)}), \\ + B_{0}^{4} (d) C_{0}^{(4)} + B_{2}^{4} (d) (C_{2}^{(4)} + C_{- 2}^{(4)}) + B_{4}^{4} (d) (C_{4}^{(4)} + C_{- 4}^{(4)}) \end{array}

with all notations and meanings of various operators and parameters defined according to the standard practice (see Ref [18]. for more details). The values of the spin-orbit coupling constants ζ _4f and ζ _5d can be readily found in the literature. A rather low symmetry of the Y site occupied by the Ce ions implies a large number of non-zero CFP

B_{q}^{k}

. Fitting these parameters to the experimental energy levels in the particular case of Ce³⁺ is not possible, since the number of CFP (13) is greater than the total number of Ce energy levels (12). To overcome such a difficulty, the exchange charge model (ECM) [25] can be used, which allows to calculate the crystal field parameters from the known crystal structure data with a small number of fitting parameters, which merely model the effects of the overlap between the wave functions of the central ion and ligands. The CFP in the ECM framework are calculated as:

B_{q}^{k} (n l) = B_{q}^{(pc) k} (n l) + B_{q}^{(ec) k} (n l)

where

B_{q}^{(pc) k} (n l)

and

B_{q}^{(ec) k} (n l)

are the so called point charges and exchange charges contributions to the CFP. They are determined by the following expressions:

B_{q}^{(pc) k} (n l) = - e^{2} 〈 n l | r^{k} | n l 〉 \sum_{i} q_{i} β_{k} \cdot {(- 1)}^{k} C_{- q}^{k} (θ_{i}, φ_{i}) / R_{i}^{k + 1},

B_{q}^{(ec) k} (n l) = e^{2} \frac{2 (2 k + 1)}{2 l + 1} \sum_{L} S_{k}^{n l} (R_{L}) \cdot {(- 1)}^{k} C_{- q}^{k} (θ_{L}, φ_{L}) / R_{L} .

Here C^k _- _q is the spherical function with rank k and order –q, the indices i and L are used to enumerate all crystal lattice ions and ligand ions, respectively, in the nearest-neighbor coordination shell, and (R_i, θ_i, φ_i), (R_L, θ_L, φ_L) are the spherical coordinates of the ith crystal lattice ions with the net charge eq_i and the Lth ligand ions in the reference system centered at Ce³⁺ ions.

The <nl|r^k|nl> entry stands for the radial integral of the Ce electron coordinate r^k between nl orbitals, which can be calculated by using numerical 4f and 5d radial wave functions of free Ce³⁺ ions. The reduction factor β_k (k = 2,4,…2l) is defined as (1- σ_k) (where σ_k is the shielding constant) [25] for 4f electrons due to the screening effects produced by the outer 5s²5p⁶ shells. Obviously, for the 5d electron such screening is absent. The ion charges q_i were fixed as + 3 for Y and Ga and −2 for O, as follows from the chemical formula.

The exchange charge term, i.e., Eq. (5), is directly proportional to the quadratic forms of various overlap integrals between the nl orbitals of the central Ce³⁺ ion and the outer 2p and 2s orbitals of oxygen ions:

S_{k}^{n l} (R_{L}) = G_{s}^{n l} S_{s}^{n l} {(R_{L})}^{2} + G_{σ}^{n l} S_{σ}^{n l} {(R_{L})}^{2} + γ_{_{k}}^{n l} G_{π}^{n l} S_{π}^{n l} {(R_{L})}^{2}

where all numerical coefficints γ can be found in Refs [18, 25 ]. and the overlap integrals are denoted as

S_{s}^{n l} (R_{L})

= <nl0|2s0>,

S_{σ}^{n l} (R_{L})

= <nl0|2p0> and

S_{π}^{n l} (R_{L})

= <nl ± 1|2p ± 1> (here the |nlm> notation is employed, where n, l and m are respectively the principal, orbital and magnetic quantum numbers); the

G_{s}^{n l}

,

G_{σ}^{n l}

and

G_{π}^{n l}

entries are the dimensionless adjustable parameters, which in our case are approximated to a single value G. The above given overlap integrals between the Ce³⁺ and O^2- ions were calculated numerically and are published in Ref [18]. An additional parameter α is introduced to model the so called multipole (like dipole, quadrupole etc) contributions to the crystal field effects. We suggested a simple linear relation, which assumes that such a contribution is simply proportional to the point charge parameters:

B_{q}^{(c o r r) 2} (n l) = α B_{q}^{(p c) 2} (n l) .

So, altogether we have only three parameters to be varied freely: α and two parameters G, since in the nearest environment of the Ce³⁺ ions in YGG there are two groups of oxygen ions located at different distance O1 (2.340 Å) and O2 (2.440 Å). The calculated energies of the 4f levels of Ce³⁺ at dodecahedral sites in YGG performed with use of this method are listed in Table 1 , the appropriate parameters of the theoretical fit to the experimental data are given in Table 2 . The agreement between the results of calculations and experimentally observed positions of the energy levels is good, with the r.m.s deviation of 34 cm⁻¹. The results of ab initio calculations [26])] predict energy positions of Ce³⁺ with somewhat lower accuracy, although in a similar range of wavenumbers.

Table 1. The calculated (E_calc) and observed (E_exp) 4f crystal field energy levels of Ce³⁺ ions in YGG (cm⁻¹).

View Table | View all tables in this article

Table 2. The energy parameters (cm⁻¹) for Ce³⁺ ions in YGG in the framework of exchange-charge model (ECM) [18]

View Table | View all tables in this article

All crystal field energy states have the same Γ ₅ representations in D₂ symmetry. The symbol “I.R. D_2d” stands for the double-value irreducible representation of the D_2d point-group (a parent group of D₂), more appropriate for garnets (see Ref [18].).

3.2. 4f – 5d absorption spectra

The temperature dependence of the ultraviolet (UV) and visible (VIS) absorption spectra of YGG Czochralski-grown crystal (thickness of 1.26 mm) is shown in Fig. 4 . The bands peaking at 352 and 416 nm at T = 10 K are assigned to the optical transitions from the 4f state to the two lowest 5d levels of the Ce³⁺ dopant.

Fig. 4 Temperature dependence of the UV and VIS absorption spectra of the YGG:Ce(0.5%) crystal.

Download Full Size | PDF

The absorption spectra consist of two relatively broad bands, with maxima at 416 and 352 nm at 10 K. The integral of the absorption coefficient of the band peaked at 352 nm increases with increasing temperature, which is accompanied by the decrease the absorption band at 416 nm. The strong absorption at about 225 nm is a bandgap edge absorption in YGG. A temperature shift of the absorption peaks towards lower energies is observed. Figure 5 shows the temperature dependence of the dipole matrix element M calculated for the 416 and 352 nm bands with:

M = \int^{​} αωdω

where α − is the absorption coefficient (after subtraction of the background), and ω − is the light frequency.

Fig. 5 Temperature dependence of the dipole matrix element for the two lowest f-d transition bands of YGG:Ce (0.5 mol. %). The solid lines are computer fits of Eqs. (9) to the data.

Download Full Size | PDF

The deactivation of the 416 nm band and activation of the 352 nm band are related to the thermal population of the upper-lying sublevel of the ² F ₅ _/ ₂ ground state, according to equations [18]:

\begin{array}{l} I (352 n m) = A_{352} \frac{S_{19} ϑ_{19} + S_{29} ϑ_{29} \cdot \exp (- \frac{Δ E}{k_{B} T})}{1 + \exp (- \frac{Δ E}{k_{B} T})}; \\ I (416 n m) = A_{416} \frac{S_{18} ϑ_{18} + S_{28} ϑ_{28} \cdot \exp (- \frac{Δ E}{k_{B} T})}{1 + \exp (- \frac{Δ E}{k_{B} T})} . \end{array}

where S_kl andν_kl denote the 4f-5d transition strength and energetic separation of the levels k and l, respectively (here k and l label the energy levels: # 1-7 for 4f levels, whereas levels # 8 and #9 represent the two lowest 5d levels), ΔE is the energy difference between energy levels #1 and #2, k _B is the Boltzmann constant, and T is the temperature. A_i are adjustable proportionality parameters. The products of parameters S_kl and υ_kl in Eqs. (9) represent the relative intensities of the transitions between the two lowest levels (#1 and #2) of the ground ²F_5/2 state and levels #8 and #9 of the 5d state. The value of ΔE was taken as equal to 159 cm⁻¹, calculated from the spectroscopic data, presented in the previous part of the paper (see Table 1).

Both #1 and #2 sublevels of the ground ²F_5/2 state are the initial states for the 4f-5d transitions studied here. However, the transitions to the levels #8 and #9 from these two sublevels of the ² F ₅ _/ ₂ ground state take place with different probabilities, according to the selection rules in the D_2d point group symmetry [18]. Therefore, the thermal population of the higher-lying level #2 of the ² F ₅ _/ ₂ ground state changes the absorption coefficients in the appropriate absorption bands. The solid lines at Fig. 5 show the computer fits of Eqs. (9) to the data; the values of the S_kl coefficients are collected in Table 3 .

Table 3. The relative values of the S_kl coefficients in Eqs. (2) obtained from the computer fits.

View Table | View all tables in this article

The relative values of the transition probabilities between levels #1 → #8 and #2 → #9 are higher than the transition probabilities between levels #1 → #9 and #2 → #8. This explains the temperature changes of the absorption coefficients in the appropriate 5d bands.

4.3 High pressure studies

YGG:Ce crystals do not exhibit any luminescence at ambient pressure, which is associated with the location of the lowest 5d ¹ level within the conduction band of YGG. Application of hydrostatic pressure above 2.5 GPa restores a quite strong Ce³⁺ luminescence in Czochralski grown crystals, which is related both to the increase of the band gap of the material and the pressure-induced lowering of the 5d¹ energy level [27]. This resembles very much the GGG:Ce crystals studied by us earlier [15].

A further increase of the pressure increases the PL intensity. The PL spectra of Ce³⁺ in YGG at T = 10 K as a function of hydrostatic pressure in the range up to about 12 GPa are shown in Fig. 6 .

Fig. 6 The pressure dependence of the PL spectra of Ce³⁺ in the YGG crystal. The narrow emission lines around 14000 cm⁻¹ are due to the ruby pressure gauge. The luminescence transitions from the level #8 of the 5d state to the ² F ₇ _/ ₂ and ² F ₅ _/ ₂ sublevels of the 4f ground state are indicated by arrows.

Download Full Size | PDF

The spectra are composed of two overlapping broad bands, pertaining to two transitions from the lowest level (level #8) of the excited 5d state to the ² F ₇ _/ ₂ (lower energy band, around 17 000 cm⁻¹) and ² F ₅ _/ ₂ (higher energy band, around 19 200 cm⁻¹) sublevels of the 4f ground state. The strong coupling to the lattice of the transitions from the excited 5d state is reflected by the large full width at half of maximum (FWHM) of the measured luminescence bands. The strong sharp emission lines around 14 000 cm⁻¹ are due to the ruby pressure gauge used in the experiment. The PL spectra have been computer fitted by the two Gaussian bands. Obtained in this way the pressure dependence of the spectral positions of the two components of the luminescence band are shown in Fig. 7 . The spectral positions of the YGG:Ce luminescence bands depend non-monotonically on pressure, exhibiting initial increase of energy and after reaching maxima at about 4–5 GPa their energies decrease with increasing pressure. This behavior is a result of quantum anticrossing behavior associated with the emergence of the lowest 5d¹ state from the conduction band states of the YGG crystal.

Fig. 7 The pressure dependence of the PL peak energies of the YGG:Ce (0.5%) crystal. The solid line is the best fit of the model to the experimental data.

Download Full Size | PDF

At ambient pressure the lowest 5d level of Ce³⁺ is degenerated with the conduction band states of YGG, and it is located slightly above the bottom of the conduction band. Due to negative pressure coefficient of Ce³⁺ 5d¹ level and positive pressure coefficient associated with the bottom of the conduction band, at certain pressure their energies become equal to each other. Quantum anticrossing effect explains the pressure dependence of the energetic position of the Ce³⁺ 5d→4f luminescence, due to the mixing between the wavefunctions of the conduction band states and the 5d states of Ce³⁺ [28] Eq. (10) describes the observed luminescence energy as a function of pressure [15]:

E_{P L} (0) = E_{5 d_{1}} (0) + \frac{1}{2} [p (k_{5 d} + k_{C B}) - Δ E (0) - \sqrt{{[Δ E (0) + p (k_{5 d} + k_{C B})]}^{2} + 4 V^{2}]},

where k _CB is the linear pressure coefficient of the CB bottom energy, ΔE(0) the ambient pressure energy separation between the minimum of the unmixed 5d ¹ energy level and CB bottom, and V is the coupling energy between the conduction band and 5d ¹states. Using for k ₅ _d the value of −17.5 ± 0.5 cm⁻¹ /kbar obtained from previous data on GGG crystal and for E ₅ _d ₁ (0) the value (22 000 ± 100) cm⁻¹, equal approximately to (E _Abs(0) + E _Em(0))/2, we have fitted this model to the experimental pressure dependence of photoluminescence. The best fit, shown as a solid line in Fig. 7, has been obtained for the following values of parameters from Eq. (10): k _CB = (40 ± 10) cm⁻¹ /kbar, ΔE(0) = (500 ± 200) cm⁻¹, and V = (2600 ± 400) cm⁻¹. This is in very good agreement with the other parameters describing the YGG [29] and GGG:Ce [15] systems. The results shows that at ambient pressure the energy separation between the minimum of the 5d¹ and conduction band bottom is smaller for YGG than for GGG crystals doped with Ce³⁺.

The pressure dependence of the lower energy band should also follow similar dependence as fitted for the higher energy band. However lower intensity of this band and the details of the deconvolution of the luminescence spectrum into two bands (also observed for the full widths at half of maxima (FWHM) presented below in Fig. 8 ) do not allow establishing the position of the maxima of the lower energy bands with such accuracy as for the higher energy band. Therefore the fit of the Eq. (10) for the lower energy band was not performed.

Fig. 8 Full width at half of maxima (FWHM) of the fitted PL luminescence 5d-4f bands as a function of pressure. Squares - higher energy band, circles – lower energy band, triangles – average.

Download Full Size | PDF

The pressure dependencies of the full widths at half of maximum of the deconvoluted luminescence bands associated with the 5d¹4f⁰ → 5d⁰4f¹ transitions are presented in Fig. 8. The deconvolution process correlates the values of the FWHM for both band due to the very strong overlap of the luminescence band. Therefore on Fig. 8 also the average value of the FWHM for these two bands is shown. The decrease of this average value can be seen with the increase of pressure, which testify that the Huang –Rhys factor also decreases slightly with the increase of pressure, however accuracy of the deconvolution does not allow establishing unambiguously the exact character of this dependence.

5. Conclusions

Low temperature infrared absorption spectra of Czochralski grown crystals of YGG:Ce³⁺ provide the evidence of at least two additional Ce³⁺-related centers formation, in addition to the main Ce³⁺ center in YGG, i.e. Ce ion located in yttrium sites. Those centers are most probably due to replacement of some Ga ions in octahedral sites by the rare earth ions. The number and concentrations of the additional Ce³⁺ centers in YGG is smaller than in other garnet crystals doped with Ce³⁺, such as GGG, YAG, or LuAG. The theoretically calculated positions of the crystal field Ce³⁺ 4f levels for the major Ce³⁺ _Y center are in very good agreement with those observed experimentally. Temperature changes of the absorption coefficients of the absorption bands associated with the 4f→5d transitions correlate very well with the thermal population of the second level of the ground ²F_5/2 state.

Application of the hydrostatic pressure above 30 kbar leads to appearance of the Ce³⁺ luminescence in this non-emitting at the ambient pressure YGG:Ce crystal, whose intensity increases with increasing pressure. This effect is associated with the anticrossing behavior and mixing between the 5d ¹ state of Ce³⁺ and CB states induced by pressure. At higher pressures the lowest 5d state, degenerated with the conduction band states at ambient pressure, moves into the band-gap of YGG, which is justified by the appearance of radiative Ce³⁺ emission.

The optical properties of YGG:Ce³⁺ crystals resemble very much the properties of the GGG:Ce garnet. The smaller lattice parameter of YGG in comparison with GGG makes the position of the lowest 5d state of Ce³⁺ closer to the bottom of the conduction band than in the case of GGG:Ce.

Acknowledgments

The cooperation program between Estonian and Polish Academies of Sciences for the years 2013–2015 is kindly acknowledged. This work was partially supported by the European Union within the European Regional Development Fund through the Innovative Economy grant MIME (POIG.01.01.02-00-108/09), and Polish National Science Center (project DEC-2012/07/B/ST5/02080). M.G. Brik acknowledges the Recruitment Program of High-end Foreign Experts (Grant No. GDW20145200225), the Programme for the Foreign Experts offered by Chongqing University of Posts and Telecommunications, Guest Professorship at the Institute of Physics of Polish Academy of Sciences and the Ministry of Education and Research of Estonia, Project PUT430.

References and links

1. S. P. Keller and G. D. Tettit, “Visible luminescence of rare-earth yttrium gallium garnets,” Phys. Rev. 121(6), 1639–1648 (1961). [CrossRef]

2. K. Kamada, T. Yanagida, J. Pejchal, M. Nikl, T. Endo, K. Tsutumi, Y. Fujimoto, A. Fukabori, and A. Yoshikawa, “Scintillator-oriented combinatorial search in Ce-doped (Y,Gd)₃(Ga,Al)₅O₁₂ multicomponent garnet compounds,” J. Phys. D Appl. Phys. 44(50), 505104 (2011). [CrossRef]

3. Z. M. Seeley, N. J. Cherepy, and S. A. Payne, “Expanded phase stability of Gd-based garnet transparent ceramic scintillators,” J. Mater. Res. 29(19), 2332–2337 (2014). [CrossRef]

4. F. Yu, D. Yuan, X. Cheng, X. Duan, X. Wang, L. Kong, L. Wang, and Z. Li, “Preparation and characterization of yttrium garnet nanoparticles by citrate sol-gel method at low temperature,” Mater. Lett. 61(11–12), 2322–2324 (2007). [CrossRef]

5. W. Nie, G. Boulon, and A. Montail, “Zero-phonon lines and energy transfer between chromium (III) and neodymium (III) multisites in yttrium aluminium garnet (YAG),” Chem. Phys. Lett. 164(1), 106–112 (1989). [CrossRef]

6. V. V. Laguta, A. M. Slipenyuk, M. D. Glinchuk, I. P. Bykov, Y. Zorenko, M. Nikl, J. Rosa, and K. Nejezchleb, “Paramagnetic impurity defects in LuAG:Ce thick film scintillators,” Rad. Measur. 42(4–5), 835–838 (2007). [CrossRef]

7. C. R. Stanek, K. J. McClellan, M. R. Levy, C. Milanese, and R. W. Grimes, “The effect of intrinsic defects on RE₃Al₅O₁₂ garnet scintillator performance,” Nucl. Instrum. Methods Phys. Res. A 579(1), 27–30 (2007). [CrossRef]

8. S. Geller and Z. Kristallogr, “Crystal chemistry of the garnets,” 125(1–6), 1–47 (1967).

9. L. Suchow, M. Kokta, and V. J. Flynn, “New garnet compounds with trivalent rare-earth ions on both dodecahedral and octahedral sites,” J. Solid State Chem. 2(2), 137–143 (1970). [CrossRef]

10. C. D. Brandle and R. L. Barns, “Crystal stoichiometry of Czochralski grown rare-earth gallium garnets,” J. Cryst. Growth 26(1), 169–170 (1974). [CrossRef]

11. M. Kh. Ashurov, Yu. Voronko, V. V. Osiko, A. A. Sobol, and M. I. Timoshechkin, “Spectroscopic study of stoichiometry deviation in crystals with garnet structure,” Phys. Status Solidi42(1), 101–110 (1977) (a). [CrossRef]

12. Yu. K. Voronko, C. B. Gessen, A. B. Kudriavtsev, A. A. Sobol, E. B. Sorokin, C. N. Uszakov, and K. B. Tsymbal, Proc. of Institute of General Physics (IOFAN), 29, 3 (1991), in “Spectroscopy of oxide crystals for quantum electronics”, Moscow, Nauka, 1991.

13. N. I. Agladze, Ch. C. Bagdasarov, E. A. Vinogradov, W. I. Zhekov, T. M. Murina, M. H. Popowa, and E. A. Fedorov, “Shape of spectral lines in garnets (Y_1-xRE_x)₃Al₅O₁₂,” Кристаллография33(4), 912–919 (1988) (in Russian).

14. A. B. Munoz-Garcia, Z. Barandiaran, and L. Seijo, “Antisite defects in Ce-doped YAG (Y₃Al₅O₁₂): first-principles study on structure and 4f-5d transitions,” J. Mater. Chem. 22(37), 19888–19897 (2012). [CrossRef]

15. A. Kamińska, A. Duzynska, M. Berkowski, S. Trushkin, and A. Suchocki, “Pressure-induced luminescence of cerium-doped gadolinium gallium garnet crystal,” Phys. Rev. B 85(15), 155111 (2012). [CrossRef]

16. X. Liu, X. Wang, and W. Shun, “Luminescence properties of the Ce³⁺ ion in yttrium gallium garnet,” Phys. Status Solidi101(2), K161–K165 (1987) (a). [CrossRef]

17. I. I. Syvorotka, D. Sugak, A. Wierzbicka, A. Wittlin, H. Przybylińska, J. Barzowska, A. Barcz, M. Berkowski, J. Domagała, S. Mahlik, M. Grinberg, C.-G. Ma, M. G. Brik, A. Kamińska, Z. Zytkiewicz, and A. Suchocki, “Optical properties of pure and Ce³⁺ doped gadolinium gallium garnet crystals and epitaxial layers,” J. Lumin. 164, 31–37 (2015). [CrossRef]

18. H. Przybylińska, C.-G. Ma, M. G. Brik, A. Kamińska, J. Szczepkowski, P. Sybilski, A. Wittlin, M. Berkowski, W. Jastrzębski, and A. Suchocki, “Evidence of multicenter structure of cerium ions in gadolinium gallium garnet crystals studied by infrared absorption spectroscopy,” Phys. Rev. B 87(4), 045114 (2013). [CrossRef]

19. H. Przybylinska, C.-G. Ma, M. G. Brik, A. Kamińska, P. Sybilski, A. Wittlin, M. Berkowski, Yu. Zorenko, V. Gorbenko, H. Wrzesinski, and A. Suchocki, “Electronic structure of Ce³⁺ multicenters in yttrium aluminum garnets,” Appl. Phys. Lett. 102(24), 241112 (2013). [CrossRef]

20. H. Przybylińska, A. Wittlin, C.-G. Ma, M. G. Brik, A. Kamińska, P. Sybilski, Yu. Zorenko, M. Nikl, V. Gorbenko, A. Fedorov, M. Kučera, and A. Suchocki, “Rare-earth antisites in lutetium aluminum garnets: Influence on lattice parameter and Ce³⁺ multicenter structure,” Opt. Mater. 36(9), 1515–1519 (2014). [CrossRef]

21. A. Lupei, V. Lupei, C. Gheorghe, S. Hau, and A. Ikesue, “Multicenter in Ce³⁺ visible emission of YAG ceramics,” Opt. Mater. 37, 727–733 (2014). [CrossRef]

22. http://abulafia.mt.ic.ac.uk/shannon/ptable.php

23. M. Nikl, J. Pejchal, E. Mihokova, J. A. Mares, H. Ogino, A. Yoshikawa, T. Fukuda, A. Vedda, and C. D’Ambrosio, “Antisite defect-free Lu₃(Ga_xAl_1-x)₅O₁₂:Pr scintillator,” Appl. Phys. Lett. 88(14), 141916 (2006). [CrossRef]

24. G. F. Herrmann, J. J. Pearson, K. A. Wickersheim, and R. A. Buchanan, “Crystal FIELD EFFECTS for Ce³⁺ and Yb³⁺ in the Garnets,” J. Appl. Phys. 37(3), 1312 (1966). [CrossRef]

25. B. Z. Malkin, in: A.A. Kaplyanskii and R.M. Macfarlane (Eds.), Spectroscopy of Solids Containing Rare Earth Ions (North-Holland Publishing Company, 1987), p. 13.

26. L. Seijo and Z. Barandiarán, “Large splittings of the 4f shell of Ce³⁺ in garnets,” Phys. Chem. Chem. Phys. 16(8), 3830–3834 (2014). [CrossRef] [PubMed]

27. A. J. Wojtowicz, M. Kazmierczak, A. Lempicki, and R. H. Bartram, “Broadband impurity absorption and luminescence: experiment and line-shape calculations,” J. Opt. Soc. Am. B 6(6), 1106–1109 (1989). [CrossRef]

28. U. Hömmerich and K. L. Bray, “Direct observation of anticrossing behavior in a luminescent Ce³⁺-doped system,” Phys. Rev. B Condens. Matter 51(13), 8595–8598 (1995). [CrossRef] [PubMed]

29. V. Monteseguro, P. Rodríguez-Hernández, V. Lavín, F. J. Manjón, and A. Muñoz, “Electronic and elastic proeprties of yttrium gallium garnet under pressure from ab initio studies,” J. Appl. Phys. 113(18), 183505 (2013). [CrossRef]

Level No.	² L_J	I.R. D_2d	E _calc	E _exp
1	²F_5/2	Γ ₆	0	0
2		Γ ₇	190	159
3		Γ ₇	721
4	²F_7/2	Γ ₆	2209	2175
5		Γ ₆	2344	2347
6		Γ ₇	2454
7		Γ ₇	3666	3652

Contribution from
	Point charges	exchange charges	other corrections	total
$B_{0}^{2}$	−746	−9	926	171
$B_{2}^{2}$	415	26	−515	−74
$B_{0}^{4}$	−339	122		−217
$B_{2}^{4}$	1625	769		2394
$B_{4}^{4}$	−924	−600		−1524
$B_{0}^{6}$	−848	−610		−1458
$B_{2}^{6}$	288	218		506
$B_{4}^{6}$	369	358		727
$B_{6}^{6}$	287	342		629
G(O₁)	2.353
G(O₂)	3.502
α	−1.241
E _avg	1666
ξ	635.7

Initial level/final level	#8	#9
#1	582 ± 6	287 ± 3
#2	281 ± 5	465 ± 2

Level No.	² L_J	I.R. D_2d	E _calc	E _exp
1	²F_5/2	Γ ₆	0	0
2		Γ ₇	190	159
3		Γ ₇	721
4	²F_7/2	Γ ₆	2209	2175
5		Γ ₆	2344	2347
6		Γ ₇	2454
7		Γ ₇	3666	3652

Contribution from
	Point charges	exchange charges	other corrections	total
$B_{0}^{2}$	−746	−9	926	171
$B_{2}^{2}$	415	26	−515	−74
$B_{0}^{4}$	−339	122		−217
$B_{2}^{4}$	1625	769		2394
$B_{4}^{4}$	−924	−600		−1524
$B_{0}^{6}$	−848	−610		−1458
$B_{2}^{6}$	288	218		506
$B_{4}^{6}$	369	358		727
$B_{6}^{6}$	287	342		629
G(O₁)	2.353
G(O₂)	3.502
α	−1.241
E _avg	1666
ξ	635.7

Ambient and high pressure spectroscopy of Ce³⁺ doped yttrium gallium garnet

Abstract

1. Introduction

2. Experimental details