Mg,Ce co-doped Lu₂Gd₁(Ga,Al)₅O₁₂ by micro-pulling down method and their luminescence properties

A stoichiometric mixture of 4N MgCO₃, CeO₂, α-Al₂O₃, β-Ga₂O₃, Lu₂O₃, and Gd₂O₃ powders was prepared as the starting material. Concerning the nominal composition, Lu³⁺ and Gd³⁺ dodecahedral sites were substituted by Ce³⁺ dopant and Mg²⁺ co-dopant according to the formula (Mg_zCe_0.005(Lu_2/3Gd_1/3)_{1−z−0.005})₃Ga_xAl_5−xO₁₂. The values of x were 0, 1, 2, 3, and 4, and z = 0.0002 for Mg co-doped Ce:LGGAG, and z = 0 for non co-doped Ce:LGGAG. For example, the sample when x = 2 and z = 0.0002 is described as Ga2; Mg co-doped, and the sample when x = 3 and z = 0.0 as Ga3; non co-doped. 1.5 wt % excess β-Ga₂O₃ was added for the ignition loss of Ga₂O₃ during growth due to the high melting point close to that of LuAG crystal. Mg co-doped and non co-doped Ce:LGGAG crystals were prepared by the micro-pulling down (µ-PD) method equipped with a radiofrequency-heating system and an Ir crucible under N₂ atmosphere. See Refs. 23 and 24 for a schematic view of the µ-PD growth apparatus. Typical pulling rates in the growth method were 0.06–0.12 mm/min, and 15–50 mm length crystals were grown around the diameter of 2–3 mm. 〈100〉 oriented YAG crystals were used as the seed crystals. The as-grown crystals were cut into plates of ϕ (2–3) mm × 1.2 mm size and polished. The cut samples were annealed for 24 h at 1700 °C in air using an electric furnace (Motoyama MS3448) in order to remove the oxygen vacancies and the thermal distortion in the as-grown crystals. The annealed crystals were mechanically polished on both sides to 1-mm size again for the absorption and luminescence spectra measurements. The remaining rods were used for phase characterization.

Rods of ϕ (2–3) mm × a few mm taken from the middle part of the grown crystals were crushed into powder by using an alumina pestle and mortar. Powder X-ray diffraction analysis was performed in the 2θ range of 15–75° by using a Bruker D8 DISCOVER-HS diffractometer (Cu Kα X-ray source: 40 kV, 40 mA).

Absorption spectra measurements for optical properties were carried out with a JASCO V550 spectrophotometer in the 190–750 nm range. Using a CCD digital camera (Andor DU420-OE) as the photon detector, the radioluminescence (RL) spectra measurements were conducted using an X-ray (40 kV, 40 mA) tube (Rigaku RINT2000) excitation source. For the gamma-ray response study, the pulse height spectra were collected under 662 keV γ-ray excitation of a ¹³⁷Cs source using a setup consisting of a photomultiplier tube (PMT; Hamamatsu R7600U-200, ultra bialkali photocathode, quartz window), a bias-supplying module (Ortec 556) for a PMT, a pre-amplifier (Ortec 113), a multichannel analyzer (MCA; Amptek 8000A), and a personal computer. Light output measurements were performed using a PMT (Hamamatsu R7600U-200). The sample plates were coupled to the PMT using silicon grease (Oken 6262A) and covered by Teflon tape. The shaping time was set to 2 µs. In addition, the light yield (LY) values were estimated by quantitative comparisons of the photo-peaks in pulse height spectra with high quality Ce:GAGG single crystal (LY = 46,000 photons/MeV) as the standard sample considering the quantum efficiency (QE) of the PMT at the emission wavelength of the Ce:GAGG standard and each LGGAG sample. The decay time was measured using almost the same setup used for the pulse height spectra measurement, where the signal output from the pre-amplifier was fed into a digital oscilloscope (Tektronix TDS3052B).

Mg co-doped and non co-doped Ce:LGGAG crystals with different Ga concentration were grown by the µ-PD method. Selected photographs of the grown Ga0, 1, 2, 3, and 4; Mg co-doped and non co-doped crystals are shown in Fig. 1. The left side of the pictures corresponds to the seed side of the crystal and the right side to the tail side. The grown crystals are transparent with yellow-to-green color, 2–3 mm in diameter and 15–50 mm in length. Some of the as-grown crystals show a rough surface due to thermal etching. However, the inner parts of the crystals were transparent.

Zoom In Zoom Out Reset image size

The crystal phase of the grown crystals was identified by powder X-ray diffraction analysis. In addition to the powder diffraction file (PDF) data for Gd₃Al₅O₁₂ (PDF: 73-1371), the powder X-ray diffraction data of Ga0–4; Mg co-doped and non co-doped crystals and their magnifications of the wide angle range in the inset are shown in Fig. 2.

Zoom In Zoom Out Reset image size

The single cubic garnet phase was observed in all the grown crystals. The wide angle peaks obviously showed systematic shifts towards shorter angles with increase of Ga³⁺ concentration due to the increase of the lattice parameters mentioned below. Figure 3 shows the dependence of the lattice constant on the Ga³⁺ concentration. The lattice parameters showed an increasing trend with increasing Ga³⁺ concentration. The ionic radii of Ga³⁺ are bigger than those of Al³⁺, which is why this trend is reasonable and is in good agreement with Vegard's law.

Zoom In Zoom Out Reset image size

The absorption spectra of the Ga2; Mg co-doped and non co-doped samples are presented in Fig. 4. The difference of the absorption between Mg co-doped and non co-doped samples in the range 210–320 nm is also shown in the inset of Fig. 4. Furthermore, the dependences of the Ce³⁺ 4f–5d₁ and 4f–5d₂ absorption band maximum positions on Ga concentration are shown in Fig. 5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In addition to the 4f–5d₁ absorption bands of Ce³⁺ centered between 450–418 nm and 4f–5d₂ around 342–355 nm, enhancement of the absorption peak around 240 nm was observed in Fig. 4 for the Mg co-doped sample. The peak around 240 nm has been ascribed to the Ce³⁺-related absorption bands, namely 4f–5d_3,4,5 states formed by the crystal field splitting of the 5d state.^11,25) Similar absorption bands were shown in a previous report for Lu₃(Ga,Al)₅O₁₂ (LGAG)¹¹⁾ and LuAG³⁾ crystals and these states in LGGAG crystals were most probably shifted towards longer wavelength than in LGAG and LuAG crystals.

The enhancement of absorption below 300 nm has been explained as the Ce⁴⁺ charge transfer (CT) absorption upon Mg co-doping according to a previous study on alkali-earth co-doped samples.^8,10,26) Furthermore, a peak around 260 nm in absorption-difference spectra (inset of Fig. 4) was observed. This absorption peak is similar to that observed in Ca²⁺ co-doped Ce:LYSO and Mg²⁺ co-doped Ce:LuAG single crystals and has been ascribed to the CT transition between the O²⁻ levels at the top of valence band (VB) and the 4f ground state of Ce⁴⁺.^8,10,26)

The peak position of the Ce³⁺ 4f–5d₁ absorption band shifted to shorter wavelength and that of the Ce³⁺ 4f–5d₂ absorption band showed the opposite behavior (see Fig. 5). Therefore their separation decreased with the increase of Ga³⁺ concentration. The increase in the lattice parameters with increasing Ga contents causes both a decrease in the crystalline field at the dodecahedral site and increased covalent bonding between the Ga³⁺ and O²⁻ ions, resulting in the removal of outer electrons of O²⁻ ligands around the dodecahedral site. This explanation is supported by previous reports.^3,25)

The peak position of the Ce³⁺ 5d₁–4f transition in the radioluminescence spectra is shown in Fig. 6 as a function of Ga concentration. The RL spectra of Ga1 and 3; Mg co-doped and non co-doped samples are simultaneously presented in the inset of Fig. 6. The Ce³⁺ emission was observed between 539–513 nm in addition to the Gd³⁺ 4f–4f emission at ∼310 nm. The RL peak showed the same behavior, as it was shifted towards shorter wavelengths by Ga³⁺ admixture as the 4f–5d₁ absorption peak.

Zoom In Zoom Out Reset image size

The Ce³⁺ emission band was enhanced by Mg co-doping for Ga1 as seen from the RL spectra in Fig. 6. On the other hand, the opposite effect was observed for Ga3. This can be explained by the stability of the Ce tetravalent state and its converging to the conduction band related to the Ga concentration. With increasing Ga concentration, the Fermi level of the materials shifts to lower energies, which increases the separation between the Ce³⁺ 4f level and the Fermi level.^27,28) Consequently, the Ce tetravalent state became more stable and a larger amount of Ce⁴⁺ was formed by Mg co-doping. According to Ref. 27, the Ce⁴⁺ 5d₁ state is found to be situated at a higher level than that of Ce³⁺, and consequently comes close to the edge of the conduction band. Therefore, due to the lower band gap energy with higher Ga concentration,²⁵⁾ the Ce⁴⁺ 5d state is much closer to the edge of the conduction band than for the samples with lower Ga content. That is why there are different consequences of the Mg co-doping for different Ga concentrations.

Measured RL emission wavelengths were used in order to correct the light yield.

The pulse height spectra of Mg co-doped and non co-doped Ce:LGGAG and Ce:GAGG reference samples by 662 keV γ-ray excitation of the ¹³⁷Cs source were collected at room temperature using the PMT. The results of the Ga2; Mg co-doped and non co-doped samples are shown in Fig. 7(a). As already mentioned in the experimental section, the LY was estimated by comparing each photo-peak channel with the Ce:GAGG standard reference sample considering the QE of the PMT at the emission wavelength of the GAGG standard and each LGGAG sample. The dependence of the LY on the material composition is shown in Fig. 7(b). The Mg co-doping decreases the light yield significantly for the samples with high Ga concentration (Ga3), while the light yield for Ga1 increases with Mg co-doping. These effects are most probably caused by changes in the stability of the tetravalent Ce and separation of its 5d state from the conduction band (see also Sect. 3.2).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Scintillation decay curves were measured using a digital oscilloscope (TDS3052B) under excitation by ¹³⁷Cs radioisotope (662 keV). Figure 8(a) shows the scintillation decay curves of Ga2; Mg co-doped and non co-doped crystals. In order to obtain the decay time values, the collected scintillation decay curves were fitted by the sum of two exponentials I(t) = I₀ + ΣA_i exp(−t/τ_i) $(i = 1,2)$, where τ₁ is the fast decay time component (first decay time) and τ₂ is the slow component (second decay time). As already mentioned in the introduction, the slow components of the decay time occur due to shallow electron traps related to anti-site (Lu_Al) defects and oxygen vacancies in LuAG-based garnet scintillators.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 8(b) shows the dependence of the fast component of the decay time on the Ga concentration. Enough signal for the fitting could not be obtained in the case of the Ga:0, 1, and 4, non co-doped and Ga:0 and 4, Mg co-doped samples. The first decay time is the shortest for Ga3 as the Ga-admixture decreases the Fermi level energy while decreasing the stability of the Ce trivalent state. Then the alternative scintillation mechanism involving Ce⁴⁺ will prevail; see for example Ref. 11. Furthermore, the first decay time showed the tendency to be accelerated by Mg co-doping, which can be interpreted by the same positive role of Ce⁴⁺ center in the scintillation mechanism as described for the silicate crystal (Ce:LYSO)¹⁷⁾ and the garnet crystals (Ce:LGAG,^10,11) Ce:GAGG^8,12)) co-doped with alkali-earth ions. This mechanism of Ce⁴⁺ center is explained step by step.²⁹⁾

The obtained values of the light yield, both components of the decay time, and the theoretical density estimated from the value^5,12) are summarized in Table I. The intensities of decay components are indicated in percent with each value of decay time, which are calculated as A₁τ₁/(A₁τ₁ + A₂τ₂) × 100% (A₁, A₂: constants) for the first component in Table I. In addition to the improvement of the first decay time, the intensities of the first components were increased by Mg co-doping even if the seconddecay time value was degraded (Ga3). This is most probably due to other effects of Mg^{2+ 30,31)} such as 1) the occupants of Ga deficiency sites and 2) obstruction of the formation of shallow electron traps like oxygen vacancies introduced as charge compensators for Ga deficiencies.