A comparative study is made of the laser crystals 50 at. % Er:YAG and 50 at. % Er:YSGG. Both lasers are constructed in the bounce geometry with quasi continuous wave (QCW) diode pumping. In Er:YAG, pulse energies of up to ~31mJ, slope efficiency of 12.6% and a red-shift in laser wavelength are observed with a final and dominant wavelength of 2.936μm. In Er:YSGG, higher performance is achieved with pulse energies of ~55mJ, slope efficiency of 20.5% and a single transition wavelength of 2.797μm observed. The study indicates that diode pumped Er:YSGG is a superior laser source at 3μm than Er:YAG and it has greater energy storage potential for Q-switched operation.
©2011 Optical Society of America
Lasing of the trivalent erbium ion in garnet hosts at wavelengths close to 3μm has attracted considerable interest. This is primarily due to the strong absorption by water around this wavelength, making erbium doped lasers suitable for use in a range of medical and biological applications [1–3]. Q-switched laser operation at the 3μm transition is also interesting as a pump source for tunable mid-IR generation using optical parametric oscillation (OPO) or optical parametric generation (OPG) .
Figure 1 is a simplified energy level diagram for the trivalent erbium ion. Laser output near 3μm from erbium doped lasers is due to a transition between the 4I11/2 and the 4I13/2 energy levels . A peculiarity of this transition is that the upper laser level has a shorter lifetime than the lower laser level. According to classical laser theory this should make CW operation impossible and these systems should only operate in a pulsed self-terminating mode . However, owing both to the complex energy transfer processes inherent in these systems and to the splitting of the laser levels into manifolds of Stark sub-levels, efficient CW operation has been achieved .
Stark splitting of the upper and lower laser levels gives rise to manifolds of sub-levels whose populations are determined by the Boltzmann distribution. This can lead to the creation of a population inversion between two Stark levels even if macroscopic inversion does not occur. The two main upconversion processes which act in highly doped erbium lasers are labeled w11 and w22. The first of these acts to remove two excitations from the lower laser level and add one to the upper laser level. The second causes depletion of the upper laser level . The efficiency of the 3 μm laser transition is dependent on the balance between these two mechanisms . A third mechanism, labeled w05, is the cross relaxation involving an ion from one of the thermally coupled 4S3/2 or 2H11/2 levels and a ground state ion. This acts to populate both the upper and lower laser levels thus aiding both upconversion mechanisms . These processes act such that the overall effect is that ions are recycled from the lower to the upper laser level, increasing the efficiency of the laser transition and actually allowing greater than unity quantum efficiencies to be achieved .
Table 1 contains some important thermal, mechanical and spectroscopic properties of Er:YSGG and Er:YAG. In comparison with YAG, the most intensively researched of the garnet hosts, YSGG exhibits more promising characteristics for construction of an efficient laser system .
The thermal characteristics of these materials are similar, but where they differ significantly is in their laser level lifetimes. The more favorable lifetime ratio for YSGG suggests that using this over YAG as the crystal host could lead to a more efficient laser source with greater energy storage potential for Q-switching of the system.
In this work a comparative study of the laser crystals Er:YAG and Er:YSGG was undertaken. The purpose of the study was to ascertain whether diode pumped Er:YSGG was a better performance laser system than Er:YAG. For high power operation we used a diode-side pumped bounce geometry architecture in a microslab laser crystal . Prior work has been performed in this geometry with Er:YAG including by Hamilton et al. who reported 1.3W from a diode pumped 50 at. % Er:YAG laser in the bounce geometry . By utilizing three total internal reflections, Ziolek et al. obtained average output powers of up to 3.2W from a bounce geometry Er:YAG laser . Less work has been performed on Er:YSGG and none in the bounce geometry. Liu et al. have obtained CW output powers of 50.2mW from an end pumped Cr,Er:YSGG system . When operating in QCW mode, a maximum average power of 169mW and maximum laser pulse energy of 1.95mJ was obtained. Waarts et al. have obtained output powers of up to 900mW using a laser diode array to pump an Er:YSGG microlaser array made up of 30 at.% Er:YSGG monolithic microchip lasers .
2. Experimental diode pumped Er:YAG/Er:YSGG laser system
A schematic of the experimental system used is shown in Fig. 2 .
This setup utilized the bounce geometry, which operates most efficiently when using a laser crystal with a high absorption coefficient as is the case for these highly doped erbium crystals (see Table 1). In this configuration the laser mode takes a total internal reflection off the pump face of the laser crystal providing access to the region of highest inversion. In addition, spatial averaging of gain and thermal non-uniformities is provided by the bounce and gives potential for good beam quality from these systems. The Er:YAG/Er:YSGG crystals used were both 50 at. % doped with Brewster cut faces and dimensions of 20 x 5 x 2mm. Side pumping was realized via the 20 x 2mm face using two 150W diode bars with fast axis collimation operating at ~966nm in QCW mode. A vertical cylindrical lens (VCLd) with a focal length of 25mm, anti-reflection coated for the pump wavelength, was used to create a horizontal line focus on the crystal face with a height ~100μm and width ~15mm. The crystal was conduction cooled through the 20 x 5mm top and bottom faces using a copper water-cooled heat-sink. A compact cavity (l~30mm) was formed using two plane mirrors (HR back mirror and partially transmitting output coupler (OC)). The bounce angle in the system was ~6° to the pump face of the crystal.
2.1 Experimental results: Er:YAG
Figure 3 shows the output pulse energy obtained from the Er:YAG system using three different output couplers with transmissions of 1%, 2% and 8%. A pump pulse duration and frequency of 1.15ms and 14Hz was used throughout these experiments. Optimal laser performance was achieved with 2% output coupling which yielded ~31mJ pulses and a slope efficiency of 12.6%. Increasing the output coupling to 8% led to a significantly higher laser threshold as well as a reduction in slope efficiency. The spatial profile of the beam, determined using a scanning slit method, was found to be near diffraction limited in the vertical (My2~1) but multimode in the horizontal with an Mx2 value ~5.
A one-meter grating spectrometer (Hilger and Watts Monospek 1000) was used to measure the spectral content of the Er:YAG laser output. Four wavelengths were detected; λ1 = 2.699μm, λ2 = 2.803μm, λ3 = 2.830μm and λ4 = 2.936μm. Temporal traces of these signals, taken using 1% output coupling, are shown in Fig. 4 .
The laser output scanned in time from λ1 to λ2/λ3 (which lased simultaneously), before settling on λ4 which contained the majority of the laser pulse energy. This corresponds to a red-shifting of the laser wavelength and a move to transitions with more favorable Boltzmann coefficient ratios. This behavior has been reported before in erbium doped lasers . The Stark levels in the upper and lower laser level manifolds (denoted i and j, respectively) between which these transitions take place are detailed in Table 2 , along with the associated Boltzmann coefficients (αi and βj) and their ratios (αi/βj).
2.2 Experimental results: Er:YSGG
Figure 5 shows the output pulse energy obtained from the Er:YSGG system using four different output couplers with transmissions of 1%, 2%, 8% and 13%, with a pump pulse duration and frequency of 1.15ms and 14Hz used throughout, as in the Er:YAG experiments.
The pump energy necessary to reach threshold was significantly lower for Er:YSGG than Er:YAG for all levels of output coupling. For the lower levels of output coupling (1 and 2%), a higher slope efficiency was obtained with Er:YAG. However, in contrast with Er:YAG, the slope efficiency continued to increase with increasing output coupler transmission, with a maximum slope efficiency of 20.5% obtained with 13% output coupling. This yielded pulse energies of ~55mJ at maximum pumping.
The Er:YSGG laser did not display the wavelength scanning behavior observed in Er:YAG, with the wavelength of the laser output measured to be at a single transition wavelength of 2.797μm. We take this transition to be between the Stark levels i = 4 and j = 6 . As with Er:YAG the beam profile of the Er:YSGG laser output was found to be good in the vertical with an My2 value ~1.5, but multimode in the horizontal with an Mx2 value ~9. The operation on a higher order mode compared to Er:YAG was likely due to higher gain in the Er:YSGG system.
The pump pulse energy necessary to reach laser threshold (with T = 1% output coupler) for the four Er:YAG transitions λ1, λ2, λ3 and λ4 and the Er:YSGG transition which we will denote as λ1* is detailed in Table 3 . This data allowed the emission cross sections of each transition, relative to the 2.936μm Er:YAG transition, to be deduced using Eq. (1):
Taking a value of 3x10−20cm2 for the emission cross section of the 2.936μm transition as quoted in the literature  indicates an emission cross section for Er:YSGG ~1.34x10−19cm2.
A comparison of the performance of two erbium doped laser crystals, Er:YAG and Er:YSGG, both with 50 at. % doping, has been undertaken. Identical QCW diode pumped, bounce geometry laser set-ups were utilized for both crystals. The wavelength of the Er:YAG laser was found to red-shift during laser output and move to transitions with more favorable Boltzmann coefficients, with the final and dominant wavelength determined to be 2.936μm. The wavelength of the Er:YSGG laser output was measured to be 2.797μm with no scanning behavior observed.
The most efficient laser performance from Er:YAG was obtained when using an output coupler with 2% transmission. This yielded a slope efficiency of 12.6% and maximum pulse energies of ~31mJ. Er:YSGG displayed more efficient performance, with a maximum slope efficiency of 20.5% and maximum pulse energies of ~55mJ obtained using 13% output coupling. The emission cross section of Er:YSGG was deduced to be ~1.34x10−19cm2. At ~4.5x greater than for the dominant transition of Er:YAG (3x10−20cm2 ), this explains the superior performance of the Er:YSGG laser. In addition, the longer upper state lifetime and lower rate of upconversion from the upper laser level (W22 = 7.3x10−22m3s−1 for Er:YSGG vs. W22 = 3.7x10−21m3s−1 for Er:YAG) suggest that Er:YSGG should exhibit better energy storage potential for Q-switched operation.
References and links
1. B. J. Dinerman and P. F. Moulton, “3-μm cw laser operations in erbium-doped YSGG, GGG, and YAG,” Opt. Lett. 19(15), 1143–1145 (1994). [PubMed]
3. H. Jelinkova, T. Dostálová, K. Hamal, O. Krejsa, J. Kubelka, and S. Procházka, “Er:YAG laser in dentistry,” Laser Phys. 8, 176–181 (1998).
4. K. L. Vodopyanov and C. C. Phillips, “Mid-IR nonlinear spectroscopy of low-dimensional semiconductor structures using an OPG,” Proc. SPIE 2801, 11–18 (1996). [CrossRef]
5. S. A. Pollack and D. B. Chang, “Upconversion-pumped population kinetics for 4I13/2 and 4I11/2 laser states of Er3+ ion in several host crystals,” Opt. Quantum Electron. 22(S1), S75–S93 (1990). [CrossRef]
6. S. Georgescu, “Mathematical modeling of 3-μm erbium lasers,” in Proceedings of The First French-Romanian Colloquium of Numerical Physics (Geometry Balkan Press, 2000), pp. 71–103.
7. M. Tikerpae, S. D. Jackson, and T. A. King, “Theoretical comparison of Er3+-doped crystal lasers,” J. Mod. Opt. 45(6), 1269–1284 (1998).
8. D. N. Nikogosyan, Properties of Optical and Laser-Related Materials: a Handbook (Wiley, 1997).
10. C. E. Hamilton, R. J. Beach, S. B. Sutton, L. H. Furu, and W. F. Krupke, “1-W average power levels and tunability from a diode-pumped 294-µm Er:YAG oscillator,” Opt. Lett. 19(20), 1627–1629 (1994). [CrossRef] [PubMed]
11. C. Ziolek, H. Ernst, G. F. Will, H. Lubatschowski, H. Welling, and W. Ertmer, “High-repetition-rate, high-average-power, diode-pumped 2.94-µm Er:YAG laser,” Opt. Lett. 26(9), 599–601 (2001). [CrossRef] [PubMed]
12. J. S. Liu, J. J. Liu, and Y. Tang, “Performance of a diode end-pumped Cr,Er:YSGG laser at 2.79μm,” Laser Phys. 18(10), 1124–1127 (2008). [CrossRef]
13. R. Waarts, D. Nam, S. Sanders, J. Harrison, and B. J. Dinerman, “Two-dimensional Er:YSGG microlaser array pumped with a monolithic two-dimensional laser diode array,” Opt. Lett. 19(21), 1738–1740 (1994). [CrossRef] [PubMed]
14. P. F. Moulton, J. G. Manni, and G. A. Rines, “Spectroscopic and laser characteristics of Er,Cr:YSGG,” IEEE J. Quantum Electron. 24(6), 960–973 (1988). [CrossRef]