Laser Performance of Yb:S-FAP In a Prismatic Side-pumping Configuration

 

Bhabana Pati and Kevin F. Wall

Q-Peak, Inc., 135 South Road, Bedford, MA  01730

pati@qpeak.com

 

K. I. Schaffers

Lawrence Livermore National Laboratory, PO Box 5508, Livermore, California 94550

 

Abstract: We have demonstrated the first, to our knowledge, side-pumped, 1047-nm, Yb:S-FAP laser. Using a prismatic pump cavity and quasi-CW diode pump lasers, we obtained an energy-per-pulse of 8 mJ with an optical-to-optical slope efficiency of 8%.

OCIS Codes: (140.3580) Lasers, solid-state (140.3480) Lasers, diode-pumped

 

1.      Introduction

 

Yb:Sr₅(PO₄)₃ (Yb:S-FAP) is a promising candidate for high energy diode-pumped laser systems [1]. In the Mercury Laser Project [2], Lawrence Livermore National Laboratory (LLNL) is undertaking a program to build a diode-pumped Yb:S-FAP laser system that will produce 100-J pulses at 10-Hz repetition rate. Until the work discussed here, all reported Yb:S-FAP lasers, to our knowledge, have employed an end-pumped design. In this paper, we present the first example of a side pumped Yb:S-FAP laser, using a prismatic pump-cavity configuration.

The prismatic configuration has demonstrated very efficient operation as compared with other side-pumped designs [3]. The design provides a simple and efficient means of both coupling the pump light from the diode lasers to the laser medium and removing heat from the laser rod, while eliminating the requirement for liquids flowing over the pump surfaces of the rod.

 

2.      The Prismatic Pump Cavity Design

 

The prismatic design employs a pump cavity consisting of a high-thermal-conductivity prism (such as MgF₂) in the form of a truncated equilateral triangle. In our case, the side length of the untruncated equilateral triangle was 12.5 mm and the length of the prism was 15 mm. The vertices of the triangle (see Fig. 1) were beveled to provide windows for the pump light. The large prism faces contacted copper heat sinks that were diamond-turned and gold-coated. At the center of the prism, a cylindrical hole provided a “slip fit” for a 4-mm diameter laser rod. We sealed an index-matching fluid, glycerin, in place between the laser rod and the MgF₂ prism. To increase the trapping efficiency of the pump light inside the prism, we attached two triangular Spectralon^® pieces with a central hole diameter of 4.2 mm to the end faces of the prism.

Since Yb:S-FAP is a quasi-three-level laser, any unpumped region in the rod will absorb the laser emission, decreasing the efficiency of the laser. In order to pump the entire rod, we kept the length of the laser rod the same as the length of the diode laser emitters; i.e., 1 cm. Two undoped YLF rods (4 mm in diameter and 5 mm in length) at both ends of the laser rod provided a means of sealing the index-matching fluid. To minimize reflection losses, we also placed index-matching fluid between the Yb:S-FAP rod and the undoped YLF rods. The outer faces of the undoped YLF rods were anti-reflection (AR) coated and the inner faces were uncoated.

The Yb:S-FAP rod had an Yb concentration of 1.1 x 10¹⁹ cm^-3 and was fabricated from material grown at LLNL. The “c” axis of the crystal was perpendicular to the rod axis and, thus, the laser emission was p-polarized. The polarization of the diode pump lasers was parallel to the junction of the diode, i.e. parallel to the axis of the laser rod, providing s absorption. We operated the quasi-CW pump diode lasers with a pulse width of 1.1 ms at a 15-Hz repetition rate. The nominal peak output power of each diode laser was 100 W and the average wavelength of three diode lasers was 899.7 nm at 25 °C.

 

3.      Experimental Results

 

Using a CCD camera, we recorded a fluorescence profile from the rod while the pump diode lasers were operated at a peak output power of 100 W (see Fig. 2). The maximum fluorescence is at the center of the rod and the variation of intensity from the center to the edge is ~10%.

 

 

Fig. 1.  A Schematic of the Prism Pump Cavity.  The MgF2 prism forms the pump cavity.  The heat sinks cool both the prism and the pump diode lasers.		Fig. 2.  Fluorescence emission when each diode laser produces 100 W peak power.  Fluorescence profiles corresponding to the cross-hairs are shown to the left and bottom of the figure.

 

We constructed a simple 4-cm long, linear laser resonator using a 50-cm-radius-of-curvature, high reflector (HR) and a flat output coupler (OC). In Fig. 3, we plot the laser output energy per pulse as a function of input pump energy per pulse for different transmission (T) output couplers. We obtained a maximum output energy of 8 mJ and slope efficiency of 8% for T = 9% OC. The maximum electrical to optical slope efficiency was 4%. The slope efficiency of each curve increases with increasing pump power and, hence, the slope efficiency was limited by the available pump diode laser power. At each data point on the graph, we optimized the temperature of the cooling water to obtain the maximum output power.

For a T = 9% OC with the diode lasers operating at 100 W peak power, the output laser power was monitored as a function of cooling-water temperature.  If the temperature was increased or decreased by 7 °C from the optimum value, the laser output power dropped by a factor of 12.  The wavelengths of the diode lasers shift at a rate of 0.25 nm/°C and the narrow absorption peak of Yb:S-FAP  is very sensitive to the wavelength of the pump diode lasers.

Using a CCD camera, we recorded the laser beam profile for the T = 9% OC at the maximum output power (Fig. 4). From the figure, it can be seen that the maximum laser intensity is at the center of the rod similar to the fluorescence profile in Fig. 2. Since Yb:S-FAP has a relatively small absorption coefficient for the s polarization (0.44 cm^-1), the pump intensity is higher where the pump beams overlap; i.e., at the center of the rod. Also, we find that the laser beam diameter is smaller than the rod diameter indicating that the edges of the rod have lower gain than the center of the rod.

 

Fig. 3.  Average output power as a function of average input power for different output couplers.  The maximum output power and slope efficiency were obtained for T = 9% OC.		Fig. 4.  Laser beam profile at maximum output power for the T = 9% output coupler.  The maximum laser intensity is at the center of the rod.

 

The M² of the laser beam was measured to be ~22.  Using ABCD matrices, the waist size of the TEM₀₀ mode was calculated for the above resonator.  If the laser emission filled the entire rod, we would expect the estimated M² of the laser to be approximately 80.  We obtained better M² from the laser at the expense of lower slope efficiency.

 

4.      Data Analysis

 

Calculation of slope efficiency

In order to better understand the laser results we conducted a series of measurements and calculations regarding pumping and laser efficiency. The slope efficiency, h_slope is the product of all the efficiency terms of the resonator and is given by

h_slope = h_oc h_q h_a h_c h_qd  = h_oc h_st                                                          (1)

 

where the quantum defect h_qd is the ratio of the pump wavelength to the laser wavelength and is 0.86.  The quantum efficiency, h_q,is 0.96 and h_c is the collection efficiency (the percentage of light transmitted from the diode laser into the prism) and is estimated to be 0.93. h_a is the absorption efficiency which is calculated and measured to be approximately 0.25.  h_st is the storage efficiency and calculate to be 19%.

To measure the absorption efficiency, we placed a detector and low-pass filter near the rod end face to measure the amount of 900-nm pump light that escaped through the rod. We first measured the escaping pump light with the diode cooling water temperature adjusted for maximum laser output. We then changed the diode temperature such that the pump laser wavelength shifted far away from the Yb:S-FAP pump absorption band, and measured pump light for the off-resonance condition. Comparing the two measurements, we determined that the rod, under the most optimal conditions, absorbed 25% of the pump energy that entered into the prism.

For a low loss resonator with a low T output coupler, the output coupling efficiency, h_oc, of a four level laser can be calculated from the expression below

 

h_oc = T/(T + L)                                                          (2)

 

where L is the round trip loss in the cavity.  The total round trip loss in the resonator is the sum of the reflection loss, extrinsic loss due to the impurity absorption in the laser rod, and intrinsic loss in the laser rod due to the quasi-three level nature of the laser material.  The total round trip reflection loss is calculated to be 3.2%.  Payne et. al [4] have calculated the extrinsic and the intrinsic losses for Yb:S-FAP with Yb³⁺ concentrations of 1.2 x 10¹⁹ cm^-3.  The extrinsic loss is 1.1% cm^-1 and the intrinsic loss is 3.2% cm^-1.  Including all the loss terms, the total round trip loss in the resonator is 12%.  For T = 9%, the calculated value of h_oc is 0.43.  From Eq. 1 we have calculated the storage efficiency to be 19% and estimated slope efficiency is 8%.

 

Findlay-Clay analysis

From the measured threshold pump power, P_th, and output coupler transmissions, T, we can use the analysis technique of Findlay-Clay to obtain the round trip intracavity loss and storage efficiency, h_st.  From a plot of –ln (1 – T) versus P_th the data was fit to the function

 

                                                                                            (3)

 

where the intercept of the y-axis yields –L, and the slope is proportional to h_st (see Fig. 5).  The estimated round-trip intracavity loss is 15% (compared to 12% above) and the storage efficiency is 20%.

 

Caird analysis

Another method of analyzing the output versus input power data to obtain L and h_st is attributable to Caird et al.  In this technique, the reciprocal of the slope efficiency is plotted versus the reciprocal of the output coupler transmission as shown in Fig. 6.  To calculate the slope efficiencies, we used only the 4 highest output power data points of each curve from the output power vs. input power plots (Fig. 3).  For a quasi three-level laser, a typical output power vs. input power plot has a significant curvature near threshold.  As shown in Fig. 6, the data points in the Caird curve do not follow a straight line.  This could be due to an error in calculating the slope efficiencies as the laser was operated at 2.4, 1.8, 1.7, and 1.4 times of the threshold for T = 2%, 9%, 13%, and 18% OCs respectively.  A fit to the data points is performed using the function

 

                                                                                                              (4)

 

h_st, is obtained from the y-axis intercept and the loss is the slope of the line times h_st. While applying a linear fit to the data, we neglected the data point obtained for the T = 18% OC because a large error might have been introduced in the calculation of slope efficiency as the laser was operated only 1.4 times above threshold.  From this analysis, the calculated round-trip loss is 16% and the storage efficiency is 18%.  Combining the measurements and calculations, along with the quantum defect, we predict a slope efficiency with a 9-% OC of 8%, in excellent agreement with the laser data.

 

Fig. 5.  A Findlay-Clay plot of the data obtained from the output power vs. input power plot.  The estimated round trip loss was 15% and the storage efficiency was 20%.	Fig. 6.  A Caird plot of the data obtained from the output power vs. input power plot.  The estimated round trip loss was 16% and the storage efficiency was 18%.

5.      Conclusions

 

We have demonstrated, for the first time, a side-pumped, 1047-nm, Yb: S-FAP laser.  Using quasi-CW diode lasers with peak powers of 100 W and pulse widths of 1.1 ms, we obtained a maximum slope efficiency of 8% with 8 mJ of energy per pulse, in excellent agreement with predictions based on independent measurements. Higher efficiencies are possible if the unwanted cavity losses can be reduced and pump absorption efficiency increased. A reduction in intracavity losses would be expected through reduction of unpumped regions in the laser rod and, possibly, by improvement of material quality. With regard to pump absorption, by changing the incoming pump light to predominantly p-polarization, we would expect the absorption coefficient to increase by a factor of about 2.5 [1], and the slope efficiency would improve by a similar factor. M² of the present laser was measured to be 22.

 

6.      Acknowledgements

 

This work was supported by the U. S. Navy, contract # N68335-00-0485

 

7.      References

 

1.   C. D. Marshall, L. K. Smith, R. J. Beach, M. A. Emanuel, K. I. Schaffers, J. A Skidmore, S. A. Payne, and B. H. T. Chai, “Diode-pumped Ytterbium-doped Sr₅(PO₄)₃F laser performance,” IEEE J. Quantum Electron., 32, 650 (1996).

2.  A. Bayramian, C. Bibeau, K. Schaffers, J. Lawson, C. Marshall, S. Payne, and, R. Morris, “Development of ytterbium doped Sr₅(PO₄)₃F for the mercury laser project,” OSA Trends in Optics and Photonics Vol 26, Advanced Solid-State Lasers, M. M. Fejer, H. Injeyan, and U. Kelloer (eds), (Optical Society of America, Washington, DC 1999)pp. 635-641.

3.   Y. Hirano, T. Yanagisawa, S. Ueno, T. Tajime, O. Uchino, T. Nagai, and C. Nagasawa, “All-solid-state high-power conduction-cooled Nd:YLF rod laser,” Opt. Lett. 25, 1168 (2000).

4.   C. D. Marshall, S. A. Payne, L. K. Smith, R. J. Beach, M. A. Emanuel, J. A Skidmore, H. T. Powell, W. F. Krupke, and B. H. T. Chai, “Diode-pumped Yb:Sr₅(PO₄)₃F laser performance,” in Trends in Optics and Photonics (TOPS) Volume 24, Advanced Solid-State lasers, T. Y. Fan and B. H. T. Chai (eds.) (Optical Society of America, Washington, DC, 1995) p. 333.

---