The interest in Dy³⁺ doped optical materials is rapidly growing recently, and more and more researches about different Dy³⁺-containing materials were reported^{[1⇓⇓⇓-5]}. Dy³⁺ doped luminescent materials in the visible region are burgeoning due to peculiar emissions which differ from other RE (rare-earth) ions doped materials. Dy³⁺ is also the only RE ion with yellow emission in addition to Tb³⁺. Because of the strong yellow emission intensity which is several times of Tb³⁺ doping at the same doping concentration and wide emission band from 565 nm to 595 nm, Dy³⁺ doped single crystals are promising in the field of all-solid-state laser^[6]. Generally, a yellow laser is used for the treatment of fundus diseases and dermatological disorders in medical treatments^[7-8]. 578 nm laser can be used for a Yb optical clock that matches the ¹S₀-³P₀ transition of the Yb atoms, and 589 nm yellow laser can be used as a sodium beacon laser^[9-10]. Dy³⁺ shows absorption bands in both blue and UV (ultraviolet) regions that match the InGaN blue LD (laser diode). The yellow laser operation using Dy³⁺ doped single crystal was reported for the first time in 2012 and gives confidence to the researchers to focus on the yellow laser based on Dy³⁺ ion-doped crystals^[11].

In recent years, all types of Dy³⁺ doped single crystals are reported, which provide a sufficient number of references for comparison of the performances among different host materials^{[12⇓⇓⇓⇓-17]}. Unfortunately, to our best knowledge, up to now yellow laser output using Dy³⁺ activated bulk crystal (not fiber laser) as a gain medium was achieved successfully only in several host materials^{[11,18⇓ -20]}. Y₃Al₅O₁₂ (YAG) is a typical laser crystal with high mechanical strength which cannot be broken down easily during laser experiments. The higher phonon energy of YAG makes the population on ⁶H_13/2 level could greater extent transfer to ground state ⁶H_15/2 by non-radiative transitions. It is relatively easy to grow YAG crystal using the Czochralski method on account of the congruent melting property. It also has a low cost due to the relatively cheaper raw materials Y₂O₃ and Al₂O₃. The spectroscopic characteristics of 3.0% Dy³⁺: YAG crystal under 384 nm excitation were investigated by Pan, et al^[21]. The spectroscopic properties of 3.0% Dy³⁺: YAG crystal grown by the micro-pulling-down method were reported by Xu^[22]. In addition, Dy³⁺ doped single crystal could be applied in the display and lighting areas on account of the strong blue and yellow emissions in visible waveband, which makes the white emitting available. Yu, et al^[23] grew the Gd³⁺/Dy³⁺ co-doped CaF₂ crystal by the Bridgeman method and the tunable white light were obtained by changing the concentration of Gd³⁺. The research from Xu, et al^[24] showed Dy³⁺/Eu³⁺ co-doped LiLuF₄ single crystal had good optical features and thermal stability. A white light emitting under 355 nm excitation in the Dy³⁺: Gd₃Sc₂Al₃O₁₂ crystal was demonstrated by Ding, et al^[25]. The 366 nm absorption band of Dy³⁺: Y₃Al₅O₁₂ (Dy: YAG) matches the excitation wavelength of commercial GaN UV-LED chips. Compared with phosphor powders containing Dy³⁺ ion, a single crystal is free of impurity phases, has good uniformity of activated ions, which does not need epoxy resin packaging.

Dy³⁺ concentration has a great influence on the fluorescence performance of Dy³⁺ doped crystals, and it is meaningful to choose the appropriate doping concentration of Dy: YAG crystal for luminescence and laser operation. The previous researches on Dy³⁺ doped YAG single crystals always concentrated on one crystal with a specific Dy³⁺ doping concentration, and there are only a few studies on the effect of doping concentration on the spectroscopic properties. In this work, the emission spectra of different doping concentration Dy: YAG crystals were contrasted at 447 nm excitation and the spectroscopic parameters were calculated and discussed. The laser properties of Dy: YAG single crystal as laser gain medium was presented.

Y_3-xDy_xAl₅O₁₂ (x = 0.015, 0.03, 0.06, 0.09, and 0.12) polycrystalline powders were synthesized by high- temperature solid reaction method. Dy₂O₃, Y₂O₃, and Al₂O₃ were weighted and mixed according to the stoichiometric ratio and pressed into tablets for sintering. After the sintering process, they were put into an Ir crucible for crystal growth. Yttrium aluminum garnet crystals with different Dy³⁺ concentrations (0.5%, 1.0%, 2.0%, 3.0%, 4.0%, atomic fraction) were grown using the Czochralski method under a high purity N₂ atmosphere. To eliminate the defects in the crystals and release the stress, such crystals were annealed in the air. An image of as-grown crystals is presented in Fig. 1.

The Dy³⁺ concentration of the upper part of the crystals was measured by an inductively coupled plasma atomic emission spectrometer (ICP-AES, Ultima2, Jobin- Yvon). Powder X-ray diffraction (XRD) was characterized on an X-ray diffractometer (Miniflex-600, Rigaku) in step scan mode. The absorption spectra of the samples with a size of 10 mm×10 mm×1 mm were recorded by a UV-VIS-NIR spectrometer (Lamda-980, Perkin-Elmer). The emission spectra and decay curves were measured by a fluorescence spectrometer (FLS980, Edinburgh Instruments). The pump source was a diode laser with the central wavelength of 447 nm and maximum output power of 3.58 W. The laser cavity consisted of two mirror with high transmission (>96%) at 447 nm and high reflection (>99.7%). The 2.0%Dy:YAG crystal with the size of 5 mm×5 mm×10 mm was wrapped with the indium foil and put on the thermos electric cooler.

As seen in Fig. 1, there appeared bubbles and crack in 4.0%Dy: YAG crystal which was grown at first in this work. The appearance of bubbles and crack might stem from the impure polycrystalline powders. To avoid the impaction of impurity phases, we used higher purity of raw materials and extended the grinding time to make the mixture more homogeneous. High quality seed crystal was also used to inhibit the defects of crystals. Furthermore, some measurements were taken such as slowing down the rate of heating and keeping the temperature above the melting point 100 ℃ for several hours, to make sure the polycrystalline melt completely which eliminated gas bubbles effectively. Changing the temperature field around crucible could improve this situation efficiently too. To strengthen the insulation, the gap between zirconia cylindrical thermal insulation materials and crucible was filled with zirconia powders. At the same time, double-layer zirconia cylinder was used to prevent crystal cracking during annealing.

Absorption spectra of Dy: YAG crystals measured at room temperature are depicted in Fig. 2, which indicates that the absorption coefficient (α) is proportional to the Dy³⁺ concentration. With an increase in doping concentration, the peak wavelengths of absorption bands are almost unchanged. Peaks at 353, 366, 386, 447, 479, 752, 804, 897, 1073, 1291, and 1687 nm correspond to the transitions from the ground state ⁶H_15/2 to ⁶P_7/2+⁴I_11/2, ⁶P_5/2+⁴M_19/2, ⁴K_17/2+⁴M_21/2+⁴F_7/2+⁴I_13/2, ⁴I_15/2, ⁴F_9/2, ⁶F_3/2, ⁶F_5/2, ⁶F_7/2, ⁶H_7/2+⁶F_9/2, ⁶H_9/2+⁶F_11/2, and ⁶H_11/2 upper-states, respectively.

Following the ICP-AES results, the segregation coefficient of Dy³⁺ in each crystal is calculated as the ratio of its concentration in the crystal and melt. Obtained results are given in Table 1. The effective segregation coefficient of Dy³⁺ in each crystal is about 0.48, which is in agreement with the previous report^[21].

Next, absorption cross-section (σ_abs) at 447 nm is obtained using the formula:

Where α is absorption coefficient and N_c is the amount of Dy³⁺ ions per cm³.

The σ_abs of Dy³⁺ in YAG is ~1.6×10^-21 cm² as presented in Table 1. The theoretical and experimental line-strengths of Dy: YAG are obtained utilizing J-O theory^[26-27]:

Where Ω_{t(t=2, 4, 6)} are the J-O intensity parameters, and Ω₂ is related to the symmetry of Dy³⁺ and the chemical bond between Dy³⁺ and O^2-. Here, ║U^(t)║ is the squared reduced matrix elements of the tensorial operator, which has been calculated by Carnall in Ref. [28]. n is refractive indices which are calculated by the Sellmeier equation for YAG crystal^[29]. The values of Planck constant (h), speed of light (c), and electron charge (e) are 6.626×10^-27 erg∙s (1 erg∙s =10^-7 J∙s), 2.998×10¹⁰ cm∙s^-1, and 4.803×10^-10 esu (1 A=3×10⁹ esu), respectively. λ¯is average wavelength of absorption band for J→J′ transition, and L is length of the crystal in the light pass direction.

The value of root-mean-square (RMS) is evaluated as:

Where S_exp and S_cal are the line-strengths of the experimental and theoretical data respectively, and N is the number of absorption bands used for calculation. The three J-O intensity parameters of the studied crystals are fitted and the above parameters are all given in Table 2.

Later, J-O intensity parameters Ω_t are used to obtain the radiative transition rates from the excited state ⁴F_9/2 to lower states which comprise both electric dipole (A_ed) and magnetic dipole (A_md) transitions in some cases. A_ed and A_md values are computed using formulae:

Where A_ed and A_md are the radiative transition rates contributed by electric-dipole and magnetic-dipole transitions, respectively. Here, the emission transition matrix of Dy³⁺ ion ║U^(t)║ and S_md values are quoted from Ref. [30].

The fluorescence branching ratios (β) and the radiative lifetime (τ_r ) of the ⁴F_9/2 level in the investigated crystals are obtained using expressions:

Spontaneous emission transition rate (A) and fluorescence branching ratio (β) of ⁴F_9/2 to lower levels are listed in Table 3. β of yellow emission which corresponds to ⁴F_9/2→⁶H_13/2 transition is in the range of 46%-50% for Dy: YAG crystals, which indicates that the yellow emission has considerable potential value in them.

Emission spectra of all samples under 447 nm excitation wavelength are shown in Fig. 3(a). Luminescence band peaks at 484, 582, 676, and 761 nm are corresponding to transitions from ⁴F_9/2 to ⁶H_15/2, ⁶H_13/2, ⁶H_11/2, and ⁶H_9/2+ ⁶F_11/2 lower levels, respectively. The luminescence intensity and the fitted ⁴F_9/2 level lifetime with different doping concentrations at 447 nm were plotted in Fig. 3(b).

Further, quantum efficiency (η) is defined by η = τ/τ_r, where, τ_r is ⁴F_9/2 level radiative lifetime, τ is ⁴F_9/2 level lifetime. The stimulated emission cross-section (σ_em) can be obtained by Füchtbauere Ladenburg formula^[31-32] as:

Where λ, n, c, and A(J→J′) have their usual meanings. The value of σ_emτ is calculated and it is inversely proportional to the laser threshold^[33]. All related spectroscopic parameters are contrasted with other Dy³⁺ doped crystals. σ_em at 582 nm of as-grown Dy: YAG crystals are found to be in the range from 2.36×10^-21 to 2.71×10^-21 cm². The obtained results in all samples are listed in Table 4.

σ_em of 1.0%Dy: YAG is close to those of gallate and scandate crystals related value at a similar concentration while it is much smaller than those of sesquioxide and fluoride crystals^{[13-14,34⇓ -36]} respective values. From the σ_emτ product, it is noticed that Dy: YAG possesses a smaller threshold than Dy³⁺: Gd₃Ga₃Al₂O₁₂, Dy³⁺: Gd₃Ga₅O₁₂, and Dy³⁺: GdScO₃ crystals. Fluoride crystals might have a lower threshold for yellow laser operation but due to adverse factors such as thermal effect and lower mechanical strength, it is difficult to develop high-power lasers^[18-19]. η of Dy: YAG declines rapidly with Dy³⁺ doping content increment, which is a sign of non-radiative transitions existence between Dy³⁺ ions, so it is not suitable to generate lasing action with Dy: YAG crystal at higher doping levels. Among all crystals, 1.0% Dy: YAG has the largest σ_em as well as σ_emτ for 582 nm yellow emission. Among five samples, the 1.0% Dy: YAG crystal also shows an intense fluorescence intensity with the 447 nm excitation with a longer decay time of 0.823 ms, which is close to the previous work^[37]. The fluorescence intensity and σ_em of 2.0% Dy: YAG are slightly less than that of 1.0% Dy: YAG, but 2.0%Dy: YAG has a higher α. Hence, 1.0%Dy: YAG and 2.0%Dy: YAG crystals are the suitable candidates for the generation of yellow laser output.

The continuous wave (CW) yellow laser output was obtained in 2.0%Dy: YAG crystal. Fig. 4(a) shows the laser spectra of Dy: YAG crystal. The central wavelength of the laser is 582.5 nm, and the full width at half maximum is 1.2 nm. Fig. 4(b) shows the relation curve between absorption pump power and output power, and the maximum output power is 166.8 μW. The slope efficiency is calculated to be 0.32%, with the maximum output power of 0.029%. The laser operation threshold is around 535 mW, which is higher than that of Dy³⁺,Tb³⁺:LiLuF₄ (320 mW)^[38]. When the input power exceeds 3382 mW, the output power decreases significantly which is mainly due to the degradation of rare earth ions caused by laser thermal effect. Owing to the laser experiment parameters haven’t been optimized, the laser output power and efficiency are still very low, we are trying to improve the laser output power now.

The visible fluorescence features of Dy: YAG crystals with different Dy³⁺ doping concentrations were analyzed thoroughly. The optimum concentration for yellow emission under 447 nm is 1.0%, and the ⁴F_9/2 level lifetime is calculated to be 0.823 ms. 1.0%Dy: YAG has the largest σ_em at 582 nm in present work. 2.0%Dy: YAG could also be an alternative crystal for yellow lasing action on consideration of its greater yellow emission intensity and σ_em, the acceptable lifetime of ⁴F_9/2 level, and higher absorption intensity. Hence, 1.0%Dy: YAG and 2.0%Dy: YAG crystals are potential candidates for yellow laser output. Finally, the CW laser with the maximum output power of 166.8 μW yellow laser operation is obtained based on 2.0%Dy: YAG crystal.