1. Introduction

Er³⁺/Yb³⁺ co-doped materials have been extensively studied as gain media of 1.55 μm laser [1–5], which is eye-safe and has been applied in many fields, such as optical communication, remote sensing and lidar [6]. Among these materials, Er³⁺/Yb³⁺ co-doped phosphate glass has been proven to be the most attractive for its large energy storage capacity (8.0 ms fluorescence lifetime of the upper laser level ⁴I_13/2 of Er³⁺) and high operation efficiency (about 30~40%) [1]. However, relatively low thermal conductivity (0.8 Wm⁻¹K⁻¹) of the phosphate glass limits its maximum average output power and then prevents it from wider practical applications [1].

In general, crystal has higher thermal conductivity and mechanical performance compared with glass. Therefore, 1.55 μm laser with higher average output power can be realized in Er³⁺/Yb³⁺ co-doped crystals. At present, Er³⁺/Yb³⁺ co-doped borate crystals have been considered as one kind of the most potential gain media for 1.55 μm laser [3,5]. For example, 1.55 μm lasers with continuous-wave output power of 1 W and slope efficiency higher than 30% have been realized in the Er³⁺/Yb³⁺ co-doped RAl₃(BO₃)₄ (R = Y, Gd or Lu) crystals [7–9]. However, due to the high phonon energy (≈1400 cm⁻¹) of the borate crystals [5], the fluorescence lifetimes of the upper laser level ⁴I_13/2 of Er³⁺ in the crystals are generally shorter than 1 ms, which decrease their energy storage capacities and limits output energy of 1.55 μm laser. Compared with phosphate glass, phosphate crystals have similar phonon energy but higher thermal conductivity so that they may be a kind of promising candidate as gain medium for 1.55 μm laser. However, to our knowledge, the investigations of Er³⁺/Yb³⁺ co-doped phosphate crystals as 1.55 μm laser gain media are still rare up to now [10].

KGd(PO₃)₄ (hereafter KGP) crystal belongs to monoclinic system with space group P2₁ and has a noncentrosymmetric structure. Its cell parameters are: a = 7.255(4) Å, b = 8.356(5) Å, c = 7.934(5) Å, β = 91.68(5)°, Z = 2, V = 480.80(5) ų [11]. The KGP is incongruent melting at 874°C. With the advantages of almost isotropic thermal expansion, large band gap and high hardness close to that of quartz [11], the KGP crystals singly-doped with Yb³⁺ or Nd³⁺ have been demonstrated as the gain media for 1.0 μm laser [12,13]. Although the maximum output power (about 100 mW) is low, which is limited by the small size of the used crystal and doping level, the slope efficiency (55%) achieved in the Yb:KGP crystal is adequately high [12]. For the Nd:KGP crystal, the laser slope efficiency realized at present is 26%, which will be enhanced by optimizing the output coupler transmission and the propagation direction and the polarization of the pumping laser [13]. Therefore, the KGP crystal is a promising laser host. Furthermore, the high cutoff phonon energy (about 1200 cm⁻¹ [11]) of the KGP crystal can significantly enhance the multiphonon relaxation from the ⁴I_11/2 multiplet of Er³⁺, and then effectively restrain the back energy-transfer from Er³⁺ to Yb³⁺ and the up-conversion losses, which are favorable for efficient 1.55 μm laser operation [1]. In this work, an Er³⁺/Yb³⁺ co-doped KGP crystal was firstly grown and its polarized spectroscopic properties related to 1.55 μm laser were investigated. Spectroscopic parameters of the crystal were compared with those of Er³⁺/Yb³⁺ co-doped phosphate glass as well as some other crystals, and its potentiality as 1.55 μm laser gain medium was evaluated.

2. Crystal growth

An Er³⁺/Yb³⁺ co-doped KGP crystal was grown with self-flux KPO₃ by the top seeded solution method. Referencing to the data in Ref [11], the solution in this work was prepared with compositions (1-x-y)Gd₂O₃ + xEr₂O₃ + yYb₂O₃: K₂O: P₂O₅ = 5.6: 34.4: 60, where x = 0.032 and y = 0.10. Raw materials of Ln₂O₃ (4N) (Ln = Gd, Er, Yb), K₂CO₃ (AR) and NH₄H₂PO₄ (AR) were weighed and the mixture was melted at 900°C. A Φ40 × 40 mm³ Pt crucible was selected as container. The saturation temperature was measured to be about 782°C. A small pure KGP seed oriented along the a crystallographic direction was used to grow the crystal. The seed holder rotated at a velocity of 60 rpm. The cooling program was set at 0.02°C/h in the first 2 hours and 0.04°C/h in the next 15 hours. Furthermore, in order to investigate the fluorescence lifetime of the ⁴I_11/2 multiplet of Er³⁺, a small Er:KGP crystal was grown from 3.2 at% Er³⁺ doped solution with the similar experimental condition.

Er³⁺ and Yb³⁺ concentrations in the grown Er:Yb:KGP crystal were determined to be 1.83 at.% (7.65 × 10¹⁹ cm⁻³) and 3.98 at.% (1.66 × 10²⁰ cm⁻³), respectively, and the Er³⁺ concentration was measured to be 1.65 at.% (6.90 × 10¹⁹ cm⁻³) in the Er:KGP crystal, by the inductively coupled plasma atomic emission spectrometry (ICP-AES, Ultima2, Jobin-Yvon). Segregation coefficients of Er³⁺ and Yb³⁺ in the crystal were calculated to be 0.57 and 0.40, respectively. The Er³⁺ concentration (7.65 × 10¹⁹ cm⁻³) in the Er:Yb:KGP crystal was referred to those of the optimized Er³⁺/Yb³⁺ co-doped YAB (8.3 × 10¹⁹ cm⁻³ [7]) and YCOB (6.2 × 10¹⁹ cm⁻³ [3]) crystals, in which efficient 1.55 μm lasers have been demonstrated. The Yb³⁺ concentration in the Er:Yb:KGP crystal has been doped as high as possible while higher Yb³⁺ concentration would result in poor crystal quality for the present crystal growth conditions. The as-grown Er:Yb:KGP crystal with a size of about 16 × 21 × 10 mm³ was generally transparent and crack-free. Then, the crystal was cut and polished into a cuboid with dimensions of 4.0 × 6.0 × 6.7 mm³ for spectral experiments, as shown in Fig. 1. Each edge of the cuboid was oriented parallel to one of the three principal optical axes, named X, Y and Z (refractive index n_X<n_Y<n_Z). Some cracks appeared on the corners of the crystal were generated in the cutting process caused by the relatively high hardness (close to quartz). In this work, the growth of the Er:Yb:KGP crystal was accomplished with a Φ40 × 40 mm³ Pt crucible. The Er:Yb:KGP crystal with a larger size will be grown by using a larger size of container in the future.

 

Fig. 1 Photograph of the Er:Yb:KGP crystal sample.

Download Full Size | PDF

3. Spectroscopic properties

Room temperature (RT) polarized absorption spectra in a range from 280 to 1700 nm recorded by an UV-VIS-NIR spectrophotometer (Lambda-950, Perkin-Elmer) are shown in Fig. 2(a). The spectra display strong polarized dependence due to the anisotropy of the crystal. For clarity, the polarized absorption spectra from 900 to 1050 nm consisted of the transitions ²F_7/2→²F_5/2 of Yb³⁺ and ⁴I_15/2→⁴I_11/2 of Er³⁺, which are used as the pumping band of 1.55 μm laser, are shown in Fig. 2(b). For most of the host materials, the absorption cross-section at 977 nm of Er³⁺ is far lower (about an order of magnitude) than that of Yb³⁺ [14–16]. Therefore, the absorption at 977 nm is mainly originated from the contribution of Yb³⁺ in the Er³⁺/Yb³⁺ co-doped materials, in which Er³⁺ concentration is lower than Yb³⁺ concentration. Then, for convenience, the absorption cross-section of the Er:Yb:KGP crystal at 977 nm was estimated roughly by dividing the value of absorption coefficient by Yb³⁺ concentration. The peak absorption cross-sections are calculated to be 0.72 × 10⁻²⁰, 1.06 × 10⁻²⁰ and 0.82 × 10⁻²⁰ cm² for E//X, E//Y and E//Z polarizations, respectively, and all at wavelength of 977 nm. The values are similar to those of the Yb³⁺ singly-doped KGP crystal demonstrated in Ref [12], for three polarizations, respectively, which also means that the Er³⁺ absorption at 977 nm in the Er:Yb:KGP crystal is weak. The maximum value is similar to that of the Er³⁺/Yb³⁺ co-doped phosphate glass (≈1 × 10⁻²⁰ cm² at 976 nm) [1]. The full width at half maximum (FWHM) of the absorption band around 977 nm for E//Y polarization of the crystal is 19 nm, which is larger than those of the Er³⁺/Yb³⁺ co-doped phosphate glass (10 nm) and YAB crystal (17 nm) [1,5]. It indicates that the Er:Yb:KGP crystal is more suitable to be pumped by diode laser when the emission bandwidth and wavelength shifting with temperature of diode laser are taken into account. Furthermore, based on the Judd-Ofelt theory [17,18], spontaneous emission probabilities for various transitions of Er³⁺ can be calculated from the absorption spectra. Here for brevity, only the values for the ⁴I_13/2 →⁴I_15/2 transition related to 1.55 μm laser are presented and the detailed calculating process can be found in [19]. For E//X, E//Y, and E//Z polarizations, the values are 87.32, 107.97, and 89.85 s⁻¹, respectively. Then, the radiative lifetime τ_r of the ⁴I_13/2 multiplet of Er³⁺ in the KGP crystal is estimated to be 10.52 ms.

 

Fig. 2 (a) RT polarized absorption spectra of the Er:Yb:KGP crystal in a range from 280 to 1700nm. (b) RT polarized absorption cross-section spectra of the Er:Yb:KGP crystal from 900 to 1050 nm.

Download Full Size | PDF

RT polarized fluorescence spectra from 1440 to 1640 nm under excitation at 976 nm were recorded using a spectrometer (FLS980, Edinburgh). Polarized emission cross-sections were calculated from the recorded fluorescence spectra by the Fuchtbauer-Ladenburg (FL) formula [20], and are shown in Fig. 3(a). The peak emission cross-sections are 0.54 × 10⁻²⁰, 0.60 × 10⁻²⁰ and 0.56 × 10⁻²⁰ cm² for E//X, E//Y and E//Z polarizations, respectively, and all at wavelength of 1537 nm. The maximum emission cross-section of the Er:Yb:KGP crystal is closed to those of the Er:Yb:phosphate glass (0.8 × 10⁻²⁰ cm² at 1535 nm) [1] and Er:Yb:YAG crystal (0.6 × 10⁻²⁰ cm² at 1646 nm) [2]. Because the emission cross-section spectrum corresponding to the ⁴I_13/2→⁴I_15/2 transition calculated by the F-L formula may be distorted by the radiation trapping effect [21], the reciprocity method (RM) [22] based on the measured absorption spectrum free from the radiation trapping effect was also adopted to calculate the emission spectrum, in which the zero line energy, E_zpl, corresponds usually to the maximum of absorption [23]. It can be seen from Fig. 3(b) that the spectra calculated by the above two methods are similar in the peak positions and shape, which means the radiation trapping effect in the Er:Yb:KGP crystal for the used Er³⁺ concentration is weak [24]. Then, the gain cross-section σ_g for E//Y polarization versus different population inversion ratio β, i.e. the ratio of the number of Er³⁺ in the upper laser level ⁴I_13/2 to the total number of Er³⁺ can be calculated by

 

Fig. 3 (a) RT polarized emission cross-section spectra of the ⁴I_13/2→⁴I_15/2 transition of Er³⁺ in the Er:Yb:KGP crystal calculated by the F-L formula. (b) Comparison of emission cross-section spectra of the ⁴I_13/2→⁴I_15/2 transition of Er³⁺ in the Er:Yb:KGP crystal for E//Y polarization calculated by the RM and the F-L formula, respectively.

Download Full Size | PDF

 

Fig. 4 Gain curves of the ⁴I_13/2→⁴I_15/2 transition of Er³⁺ in the Er:Yb:KGP crystal for E//Y polarization with different β.

Download Full Size | PDF

By using a microsecond flash lamp (μF900, Edinburgh) as the exciting source, RT fluorescence decay curves of some multiplets of Er³⁺ related to the 1.55 μm laser were recorded by the spectrometer (FLS980, Edinburgh). When the exciting and emission wavelengths were 976 and 1537 nm, respectively, the fluorescence decay curve of the ⁴I_13/2 multiplet in the Er:Yb:KGP bulk crystal was recorded and is shown in Fig. 5(a) and the fitted fluorescence lifetime is about 6.91 ms. To avoid the influence of radiation trapping effect on the measurement of lifetime [21], powder sample of the crystal immersed in refractive index-matching fluid monochlorobenzene (refractive index n = 1.52) was also used [21,25], the fitted fluorescence lifetime is about 6.32 ms and also shown in Fig. 5(a). Additionally, the fluorescence decay curve of ⁴I_13/2 multiplet of the Er:KGP crystal powder sample was measured at 1574 nm with exciting at 1503 nm, as shown in Fig. 5(b). The fitted fluorescence lifetime was 6.25 ms and close to that of the Er:Yb:KGP crystal powder sample. The small discrepancy of measured lifetimes between the crystal and powder samples means that the radiation trapping effect is not serious in the Er:Yb:KGP crystal, which may be originated from low Er³⁺ concentration in the crystal [26]. Combining with the radiative lifetime (10.52 ms) calculated above, the fluorescence quantum efficiency of the ⁴I_13/2 multiplet in the Er:Yb:KGP crystal is estimated to be 60%, and is much higher than those of the Er³⁺/Yb³⁺ co-doped borate crystals, such as YAB (8.3%) [5] and YCOB (21%) [3], which is caused by the lower phonon energy of KGP (≈1200 cm⁻¹) than that of borate crystals (≈1400 cm⁻¹). However, the fluorescence quantum efficiency of the ⁴I_13/2 multiplet in the Er:Yb:KGP crystal is still lower than that of the Er:Yb:phosphate glass (80%) [27], and a main reason may be originated from low crystal optical quality, such as impurities and defects existed in the as-grown crystal, which are nonradiative traps of excitation energy and thus result in decrement of the fluorescence lifetime .

 

Fig. 5 (a) RT fluorescence decay curves of the ⁴I_13/2 multiplet of the Er:Yb:KGP bulk and powder samples. (b) RT fluorescence decay curve of the ⁴I_13/2 multiplet of the Er:KGP crystal powder sample. (c) RT fluorescence decay curve of the ⁴I_11/2 multiplet of the Er:KGP crystal powder sample

Download Full Size | PDF

Taking into account the overlap of emission bands between the transitions of ²F_5/2→²F_7/2 for Yb³⁺ and ⁴I_11/2→⁴I_15/2 for Er³⁺, the grown Er³⁺ singly-doped KGP crystal was used for measuring the fluorescence lifetime of the ⁴I_11/2 multiplet of Er³⁺. Excited at 519 nm by an OPO pulse laser (Vibrant 355II, OPOTEK) with pulse duration of 5 ns, the fluorescence decay curve at 980nm of the Er:KGP crystal powder sample was recorded by a spectrometer (FLS980, Edinburgh) and is shown in Fig. 5(c). The fluorescence lifetime of the ⁴I_11/2 multiplet of Er³⁺ is fitted to be about 1.54μs, which is similar to that of the Er³⁺/Yb³⁺ co-doped phosphate glass (2~3 μs) [28] and much shorter than those of the Er³⁺/Yb³⁺ co-doped YAG (100 μs) [2] and YVO₄ (28 μs) crystals [29]. It is well known that a shorter fluorescence lifetime of the ⁴I_11/2 multiplet will make more excited ions populate in the upper laser level ⁴I_13/2 by decreasing up-conversion loss and reverse energy transfer from Er³⁺ to Yb³⁺, and then benefit the operation of 1.55 μm laser [5].

The energy transfer efficiency from Yb³⁺ to Er³⁺ was estimated by ${\eta }_{ET}\text{=}1-\tau f/\tau 0$, where τ₀ and τ_f represent the fluorescence lifetimes of the ²F_5/2 multiplet of Yb³⁺ in Yb³⁺ singly-doped and Er³⁺/Yb³⁺ co-doped KGP crystals, respectively [3]. Powder samples in the refractive index-matching fluid monochlorobenzene were also used to measure the fluorescence decay curves of the ²F_5/2 multiplet of Yb³⁺, which are not shown here for the brevity. The fluorescence lifetimes of Yb³⁺ in the Er:Yb:KGP and Yb:KGP (4.43 at.%) crystals were fitted to be 0.50 ms and 1.31 ms, respectively. Therefore, the energy transfer efficiency η_ET is about 62%. The spectroscopic parameters of the Er³⁺/Yb³⁺ co-doped phosphate glass and crystals related to the 1.55 μm laser are listed in the Table 1. Compared with that of the Er:Yb:phosphate glass, the lower transfer efficiency of the Er:Yb:KGP crystal may be caused by the lower Yb³⁺ concentration (about 1.66 × 10²⁰ cm⁻³ in the crystal whereas 10 × 10²⁰ cm⁻³ in the glass) [1]. It has been demonstrated that the Yb³⁺concentration can be raised to 25~30 at.% in the KGP crystal through optimizing the crystallization region [11]. Therefore, the Yb³⁺ concentration in the Er:Yb:KGP needs to be further optimized for increasing the energy transfer efficiency.

Table 1. Comparison of some spectroscopic parameters of the Er³⁺ and Yb³⁺ co-doped crystals and glass related to the 1.55 μm laser.

View Table

4. Conclusion

A KGP crystal doped with 1.83 at.% Er³⁺ and 3.98 at.% Yb³⁺ was firstly grown by the top seeded solution method. Compared with the Er³⁺/Yb³⁺ co-doped YAG and YVO₄ crystals, in which low laser operation efficiencies (lower than 7%) have been demonstrated, the Er:Yb:KGP crystal is impressive with its much shorter fluorescence lifetime of the ⁴I_11/2 multiplet (1.54 μs). Therefore, most of the excited Er³⁺ ions in the ⁴I_11/2 multiplet can be populated in the upper laser level ⁴I_13/2 by nonradiative transition, thus higher operation efficiency of 1.55 μm laser can be expected in the Er:Yb:KGP crystal. In the other hand, the Er:Yb:KGP crystal has a longer fluorescence lifetime of ⁴I_13/2 multiplet (6.32 ms) than those of the Er³⁺/Yb³⁺ co-doped borate crystals, in which efficient continuous-wave 1.55 μm lasers have been demonstrated. Hence, the Er:Yb:KGP crystal has larger energy storage capacity and would be a good candidate for realizing high energy 1.55 μm Q-switched pulse laser. Furthermore, it can also be seen that most of the spectroscopic properties of the Er:Yb:KGP crystal are comparable to those of the Er:Yb:phosphate glass, which is the only commercial 1.55 μm laser material at present, except the fluorescence quantum efficiency and energy transfer efficiency, which are mainly caused by the poor crystal optical quality and low Yb³⁺ concentration, respectively. Therefore, when the crystal growth technique is further optimized, the optical quality and doping concentration of the KGP crystal can be improved and then the spectroscopic properties of the Er:Yb:KGP crystal may be similar to those of the Er:Yb:phosphate glass. Combined with higher thermal performance of the crystalline material, it can be concluded that the Er:Yb:KGP crystal is a promising gain medium of 1.55 μm laser.