1. Introduction

Low quantum defect (QD), eye-safe laser operation of resonantly pumped laser materials based on Er-doped laser crystals and ceramics has drawn significant attention lately due to its potential for designing scalable high power laser systems. Among the most recently studied laser materials, Er-doped orthovanadate single crystals, Er³⁺:YVO₄ and Er³⁺:GdVO₄, were found to be the most efficient. Under resonant pumping (at 1528 and 1539 nm) they have demonstrated nearly QD-limited laser operation with slope efficiencies as high as 83 - 85% at cryogenic temperature and 54 - 58% at room temperature [1–5]. Both of these crystals have closely related physical properties. However, there is quite conflicting literature data about their comparative thermal properties relative to the more conventional YAG [6–8], see Table 1 . Taking into account that their reported spectroscopic features [9] are also somewhat different, there is a practical question as to which of these Er-doped orthovanadates is best suited for specific laser applications in terms of performance efficiency. The answer can be given via back to back comparison of their laser performances in identical experimental conditions. The comprehensive comparison for Er³⁺-doped YVO₄ and GdVO₄ single crystals with the ⁴I_15/2 → ⁴I_11/2 (~980-nm) excitation at room temperature was recently reported in [9]. To the best of our knowledge, no direct comparison results are so far available for these crystals with the most efficient, resonant ⁴I_15/2 → ⁴I_13/2 pumping.

Table 1. Physical, Optical and Thermal Properties of Er³⁺:YVO₄, Er³⁺:GdVO₄ and Er³⁺:YAG Crystals

View Table

This paper presents the results of a comprehensive comparison of optical, thermal and spectroscopic properties and resonantly (in-band) pumped laser performances of Er³⁺:YVO₄ and Er³⁺:GdVO₄ single crystals at cryogenic and room temperatures.

2. Optical, thermal and spectroscopic properties of Er³⁺:YVO₄ and Er³⁺:GdVO₄ crystals

Table 1 presents a thorough collection of the optical and thermal properties of Er³⁺:YVO₄ and Er³⁺:GdVO₄ single crystals measured by numerous authors, including our own results. In general, one can see that there is enough of a similarity in most of those properties. However, there are also some differences. It can be seen that there is nearly a two-fold data spread when it comes to thermal conductivity measurements of the YVO₄ crystal, while thermal conductivity measurements for the GdVO₄ crystal are very consistent with each other. For both of them, thermal conductivity undergoes ~5-6-fold increase when the crystal is cooled from room temperature (RT) to liquid nitrogen temperature (LNT). Thermal expansion coefficient (CTE) measurements for both crystals obtained by different groups also show a noticeable data spread. All of them reveal considerable (more than four times) anisotropy of CTE in both Er³⁺:YVO₄ and Er³⁺:GdVO₄ single crystals (expansion along the optical c-axis versus that in the perpendicular direction). Unfortunately, to the best of our knowledge, there is either no or very few available data on CTE and temperature coefficients of refractive index, dn/dT, for both crystals at temperatures from 295 K to 77K.

3. Spectroscopy

Spectroscopic measurements were done on Er³⁺:GdVO₄ samples with Er³⁺ ion concentration of 0.5 at.% and 0.7 at.% (ion number density, N_Er = 6.05 x 10¹⁹ cm⁻³ and 8.47 x 10¹⁹ cm⁻³, respectively) and Er³⁺:YVO₄ samples with Er³⁺ ion concentration of 0.5 at.% (N_Er = 6.23 x 10¹⁹ cm⁻³). All crystals were grown by the Czochralski technique [10]. All measurements pertain to the ⁴I_13/2 ↔ ⁴I_15/2 Er³⁺ transitions in the 1450 – 1650 nm wavelength range.

Absorption spectra associated with ⁴I_13/2 → ⁴I_15/2 transitions of Er³⁺ in GdVO₄ and YVO₄ crystals were measured in the temperature range from 8 K to 300 K using a Cary 6000i spectrophotometer (resolution 0.1 nm) with an InGaAs detector operating in the fixed-slit-width mode. Figures 1 (a, b, c, and d) indicate ground-state absorption spectra of the Er³⁺:GdVO₄ and the Er³⁺:YVO₄ for σ- (a and c) and π- (b and d) polarizations for temperatures of 77 K (a and b) and 300 K (c and d). The effective cross-sections were calculated from measured absorbances using the above ion number densities. It can be seen that most of the absorption lines of the Er³⁺:GdVO₄ at 77 K in π-polarization are stronger than in σ-polarization, while for the Er³⁺:YVO₄ - the opposite pattern prevails. A direct comparison of the absorption spectra at 77 K indicates that major absorption cross-sections are larger for the Er³⁺:YVO₄ (with the exception of the very strong π-polarized band around 1503 nm in the Er³⁺:GdVO₄), while the absorption lines of Er³⁺:GdVO₄ are generally broader. At room temperature, cross-sections and line widths of both crystals are approximately equal.

 

Fig. 1 ⁴I_15/2 → ⁴I_13/2 absorption spectrum of Er³⁺ in GdVO₄ and YVO₄ single crystals measured at 77 K (a, b) and 300 K (c, d) for and σ- and π-polarizations.

Download Full Size | PDF

The emission spectrum of the Er³⁺:GdVO₄ crystal was obtained by illuminating the 0.5 at.% doped sample with 970-nm diode laser emission and by collecting the fluorescence light with the Optical Spectrum Analyzer [4,5]. Cross-sections were calculated using the standard Fuchtbauer-Landenburg method [11]. Figures 2 show the σ-polarized (a and c) and the π-polarized (b and d) emission spectra of the Er³⁺:GdVO₄ and Er³⁺:YVO₄ measured at 77 K (a and b) and 300 K (c and d). One can see that the Er³⁺:GdVO₄ has slightly larger π-polarized emission cross-sections in the vicinity of 1600 nm. In addition, the slightly shorter π-polarized emission wavelength (~1598 nm in the Er³⁺:GdVO₄ vs ~1603 nm in the Er³⁺:YVO₄) should provide a benefit of the lower QD operation.Based on the analysis of our spectroscopic data we produced two energy level diagrams for Er³⁺:YVO₄ and Er³⁺:GdVO₄ which are presented in Fig. 3 .

 

Fig. 2 ⁴I_13/2 → ⁴I_15/2 emission spectrum of Er³⁺ in GdVO₄ and YVO₄ single crystals measured at 77 K (a, b) and 300 K (c, d) for σ- and π-polarizations.

Download Full Size | PDF

 

Fig. 3 Energy level diagram of Er³⁺ in GdVO₄ and YVO₄ single crystals at 77 K. Orange arrows represent major absorption transitions. Blue and pink arrows represent the observed laser transitions. For each multiplet, energies of Stark sub-levels are presented on the left and Boltzmann population factors (for 77 and 300 K) are presented on the right.

Download Full Size | PDF

4. Laser experiments

Laser experiments were performed with the anti-reflection (AR) coated 0.7 at.% Er³⁺:GdVO₄ and 0.5 at.% Er³⁺:YVO₄ 10 mm-long, single crystal slabs with the cross section of 3 mm x 7 mm. The Er³⁺:GdVO₄ slab had the crystallographic c-axis aligned with the lateral 7 mm direction. The Er³⁺:YVO₄ slab had the c-axis aligned with the longitudinal 10 mm direction, therefore only σ-polarized absorption and emission could be utilized in lasing. For laser experiments at cryogenic temperatures, slabs were mounted on a copper cold finger inside the boil-off liquid nitrogen cryostat with two AR coated fused silica windows (T > 99.9% at 1500-1650 nm). For RT laser experiments, the crystals were clamped between the water-cooled copper plates and conductively cooled from top and bottom to + 18° C.

Two different lasers, suitable for pumping into one of the major absorption bands of Er³⁺:GdVO₄ or Er³⁺:YVO₄ within the 1525 - 1540 nm range, were used in our experiments. One of them was a narrowband, single-mode (SM), continuous wave (CW) Er-fiber laser (IPG Photonics) at 1538.8 nm with the output bandwidth of ~0.3 nm (full width at half maximum, FWHM). This narrow bandwidth output is fully accommodated by the 1538.8 nm absorption line of both Er³⁺:GdVO₄ and Er³⁺:YVO₄ crystals. The unpolarized pump beam was focused into a laser slab by a spherical lens with the focal length of 100 mm. With the Er-fiber laser pump the excited volume inside the slab was nearly cylindrical along the entire slab length with the ~380 μm diameter (at 1/e² intensity level).

In order to provide the most practical comparison, the second pump source was a commercial spectrally narrowed (~2 nm FWHM), fast and slow axis collimated (FAC-SAC), 13-bar InP laser diode stack (BrightLock® Stacked Array, QPC Lasers). It operated in a QCW regime with the operational duty cycle of 25% (t_pulse = 25 ms, F = 10 Hz). The incident pump beam was utilized as π-polarized for Er:GdVO₄ and σ-polarized for Er:YVO_4. The spectral maximum of the diode stack was adjusted to the peak of the ~1529 nm absorption band by varying the temperature of the cooling water flow. In order to achieve nearly equal divergence of the pump beam in the vertical and horizontal directions, an additional 4^x cylindrical telescope, formed by the negative and positive lenses, was inserted after the diode bar stack. A variable attenuator, comprising of a combination of a polarizing cube and a half-wave plate, was used to vary the pump power without changing the diode current. The collimated pump beam was focused into the crystal by a spherical lens with a focal length of 75 mm through a flat dichroic mirror (HT T > 90% at 1520 −1540 nm, HR R > 99.5% at 1580-1650 nm). The excited volume inside the slab had a conical shape with the 1/e² diameter varying from ~960 μm in the center to ~1200 μm at the slab ends. The laser cavity was formed by the HR dichroic mirror and the plano-concave output coupler with a radius of curvature R_CC = 250 mm and reflectivities between 85% and 70%. The laser cavity lengths (L_CAV) were ~100 mm and ~80 mm for the cases of pumping by the Er-fiber laser and the diode bar stack, respectively. A simplified experimental setup of the cryogenically cooled Er³⁺-doped YVO₄ (GdVO₄) laser is shown in Fig. 4 .

 

Fig. 4 Simplified optical layout of the cryogenically-cooled (a) and room temperature (b) Er³⁺:YVO₄ and Er³⁺:GdVO₄ lasers.

Download Full Size | PDF

5. Comparative performances of Er³⁺:YVO₄ and Er³⁺:GdVO₄ lasers at 77 K

The CW and QCW (F = 10 Hz, t_pulse = 25 ms) performances at 77 K of the Er³⁺:GdVO₄ laser resonantly pumped into the 1538.8 nm absorption line by a CW Er-fiber laser are shown in Figs. 5 (a–d) . The fraction of the absorbed pump power varied in both Er³⁺:GdVO₄ and Er³⁺:YVO₄ crystals due to saturation effects [12] from ~0.85, right above the laser threshold, to ~0.6 - 0.7 at the maximum incident pump power.

 

Fig. 5 CW and QCW input-output characteristics of cryogenic Er³⁺(0.7%):GdVO₄ and Er³⁺(0.5%):YVO₄ lasers resonantly pumped at 1538.8 nm by an Er-doped fiber laser. The cavity length is 80 mm, plano-concave output coupler with R_CC = 100 mm for all cases, output coupler reflectivity R_OC = 0.85 (a, b), 0.8 (c, d).

Download Full Size | PDF

We observed that both cryogenically cooled Er³⁺:GdVO₄ and Er³⁺:YVO₄ lasers are capable of performing with very high slope efficiency with respect to the absorbed pump: 83% and 85% for Er³⁺:GdVO₄ and Er³⁺:YVO₄ respectively, see Fig. 5a and Fig. 6a . While Er³⁺:YVO₄ operated in σ-polarization at 1593.5 nm, with 0.7 nm bandwidth at the maximum pump power, the Er³⁺:GdVO₄ laser emitted at 1598.5 nm in π-polarization, with 0.45 nm bandwidth. As was expected, the characteristics of both lasers were very similar if they were compared in the same operating mode (CW or QCW).

 

Fig. 6 Comparison of the output power vs the absorbed pump power dependences for the Er³⁺:YVO₄ (a) and the Er³⁺:GdVO₄ (b) cryogenic lasers in CW and QCW regimes, when both laser are pumped by an Er-doped fiber laser at 1538.8 nm. QCW regime is 10 Hz, 25 ms pulse duration.

Download Full Size | PDF

In the meantime, if one compares lasers (either Er³⁺:GdVO₄ or Er³⁺:YVO₄) in the CW versus the QCW regime (with the same laser cavity parameters), it can be concluded that the effective thermal resistivity of the Er³⁺:GdVO₄ crystal is somewhat lower. Figure 6, where comparative performances of both lasers are presented, shows that while the efficiency of the Er³⁺:YVO₄ laser dropped noticeably in switching from QCW to CW operation, the efficiency of the Er³⁺:GdVO₄ laser changed only slightly. In an attempt to analyze this phenomenon, we observed that with the same amount of absorbed pump power, the temperature of the Er³⁺:YVO₄ was increasing much faster than that of the Er³⁺:GdVO₄, reaching low 90’s K at full pump power for the former versus low 80’s K for the latter. This, seemingly minor, temperature difference is very critical for the low-QD operation with the terminal laser level positioned close to the bottom of the ground state multiplet. As a result, a slight increase in the ground state absorption affects the efficiency of the Er³⁺:YVO₄ laser much more than that of the Er³⁺:GdVO₄ laser. The observed difference in the experimental data is based on the temperature excursion as well as on a discernible difference in the ground state Stark-split level positioning between the two materials, see Fig. 3.

We also compared performances of Er³⁺:YVO₄ and Er³⁺:GdVO₄ lasers under the resonant pumping by a spectrally narrowed laser diode bar stack at 1529.3 nm, see Fig. 7 . In this pumping arrangement, both lasers operated with similar slope efficiencies.

 

Fig. 7 QCW laser output power vs. absorbed pump power dependences for Er³⁺(0.7%):GdVO₄ and Er³⁺(0.5%):YVO₄ cryogenic lasers, resonantly pumped by a laser diode bar stack.

Download Full Size | PDF

It should be noted that slope efficiencies of both orthovanadate lasers, pumped by a laser diode bar stack, are lower than in the case of fiber laser pumping despite the fact that the absorbed fraction of the pump power in both cases is nearly the same. The reason for this difference is due to a much poorer spatial overlap of the pumped volume in the crystal with the laser cavity mode when pumping by the laser diode bar stack.

6. Comparative performances of Er³⁺:YVO₄ and Er³⁺:GdVO₄ lasers at 300K

RT experiments with both lasers were carried out with the same pump sources. The diode bar stack was tuned into 1529 nm absorption bands (see Figs. 1, c and d). Again, Er³⁺:GdVO₄ was pumped in a π-polarized configuration, while Er³⁺:YVO₄ crystal was pumped in the σ-polarized one.

Without a cryostat, RT experiments allowed a very short laser cavity (L_CAV = 40 mm), but required a different curvature of the output coupler (R_CC = 250 mm) to achieve an optimal spatial pump-laser mode overlap with the TEM₀₀ mode diameter of ~470 μm. Figure 8a depicts the output power of the Er³⁺:GdVO₄ and Er³⁺:YVO₄ lasers pumped by a laser diode bar stack versus the absorbed pump power for the 10% output coupling. As in the previous case, both lasers showed approximately equal slope efficiencies, although fractions of the absorbed pump were quite different: 50-40% for Er³⁺:GdVO₄ and 30-25% for Er³⁺:YVO₄ depending on the incident pump power. Such a fraction is noticeably lower at RT than that measured at LNT.

 

Fig. 8 CW and QCW laser output power vs. the absorbed pump power dependences for Er³⁺(0.7%):GdVO₄ and Er³⁺(0.5%):YVO₄ lasers operating at RT. Output coupler reflectivity R_OC = 0.9 (a), R_OC = 0.95 (b). The black and red lines are linear regressions of the experimental data points.

Download Full Size | PDF

Figure 8b shows the result of the experiments with pumping by a narrow linewidth (~0.3 nm FWHM) Er-fiber laser operating in the CW regime. Compared with the diode bar stack, the fiber laser provides a much better spatial pump-laser mode overlap. This mode matching is critical for the efficient Er-doped laser operation at RT. The pump beam was focused into the crystal by a spherical lens with focal length of 100 mm. The short laser cavity (L_CAV = 40 mm) with plano-concave output coupler (R_CC = 100 mm) defined the TEM₀₀ mode waste diameter of ~340 μm. The best results were obtained with the output coupler reflectivity of ~95%. Mention should be made that no noticeable heating of the crystals was measured up to ~20 W of the incident pump power. As can be seen, both slope efficiencies measured versus the absorbed pump were nearly the same for both crystals.

7. Conclusions

We compared thermal and spectroscopic properties and resonantly pumped laser performances of Er³⁺:YVO₄ and Er³⁺:GdVO₄ single crystals, at room and cryogenic temperatures. Both lasers demonstrated a low-QD operation with slope efficiencies well in excess of 80%. The absorption cross-sections in σ-polarization are slightly stronger for Er^3+:YVO₄, while Er:GdVO₄ absorbs stronger in π-polarization. The maximum value of the emission cross-section at LNT (1598 nm, π-polarization) is higher for Er^3+:GdVO₄. Despite these small differences in absorption and emission cross-sections, both crystals demonstrated very similar laser performances at RT and LNT. We observed that the output characteristics of cryogenic Er^3+:GdVO₄ laser are less sensitive to the thermal loading, which is an indirect indication of higher thermal conductivity (along the c-axis) of the Er^3+:GdVO₄ material, though it is not as obvious from the data available in the literature.