Introduction and objectives:
Most diode-pumped solid state lasers operate with Nd-doped crystals at near-infrared (IR) wavelengths around 1060 nm. There are a number of laser materials based on Er-doped crystals that can provide much longer wavelengths, in the mid-IR around 3000 nm. In this Phase II program sponsored by the Air Force Research Laboratory at Phillips Laboratory, Kirtland AFB, NM, Contract # F29601-98-C-0009, Q-Peak developed several Er-doped systems, taking advantage of recent developments in high-power diode pump lasers. The original motivation, as indicated by the program title, was development of lasers for surgical applications, but due to changes in the sponsoring agency, the effort became a general technology advancement program.
The Technical Objectives of this Phase II SBIR effort were as follows:
1. Conduct thermal modeling of high-power, diode-pumped Er lasers. The objective was to determine the optimum crystal geometry for producing average output powers of 0.5-1.5 W from Er-doped crystals
2. Design and build a cw, diode-pumped, Er:YAG laser with an output power goal of 0.5 W.
3. Design and build a cw, diode-pumped Er:YLF laser with an output power goal of >1 W.
4. Design and build a long-pulse, diode-pumped Er:YLF laser with an output energy goal of 25 mJ at a pulse rate of 50 Hz.
Papers presented:
The following paper was presented at CLEO 2000.
1.8 W CW Diode-Pumped Er:YLF Laser
Summary of results:
The following is a summary of results on the program and, as such, omits much presentation of our efforts related to the design and characterization of the different laser systems.
1. Thermal modeling
Models of end-pumped, rod-laser heating and stress are well developed. We concentrated our work on models of the temperature and thermally-induced stress distributions in side-pumped Er:YLF slabs. These were developed using a finite element analysis computer program (COSMOS/M). The purpose of this analysis was to estimate thermal loading that would result in fracture of the laser medium. It was assumed that surface tensile stresses were the most important determinants of fracture. We initially used a relatively simple model to explore the effects of constraining the surfaces of the YLF crystal and to gain confidence in our model. Subsequently we added Beer’s Law absorption of the pump-power to the model and examined subtle effects of absorption coefficient, crystal geometry, and pump distribution.
The side-pumped Er:YLF model uses the geometry shown in Fig. 1. The slab is 2.5 x 4 x 28 mm and the c-axis is along the 4-mm dimension (which is defined as the x-axis). The slab is pumped by two 10-mm diode laser bars through the 2.5 x 28 mm faces in the regions indicated by the gray lines (solid and dashed). The pump regions, 300 microns high, are centered in the y-dimension and offset from one another along the z-axis. One pump laser diode pumps from the + x-direction and the other pumps in the – x-direction.
Figure 1.The Er:YLF slab geometry. The pump regions are indicated by the gray lines (solid and dashed).
Figure 2 shows the results of a calculation assuming an absorption coefficient for the pump of 3 cm-1 and a generated heat load for each diode of 20 W. The pump light is reflected back from the opposite surface to allow double-pass absorption.
>
Figure 2. A section plot of the temperature distribution.
Figure 3 plots the resultant stress distribution.
Figure 3.The principal stress resulting from the thermal loading shown in Figure 2.
We investigated the effect of changing the thickness of the Er:YLF slab in the y-dimension by calculating the temperature and principal stress distributions for 1-mm and 5-mm slab thicknesses. In each case, the absorption coefficient was fixed at 3 cm-1, the height of the pump region fixed at 300 mm, and double-pass pumping was modeled. The x-z faces of the slab were held at a temperature of 20 °C and the slab was mechanically unconstrained. The total heat load was assumed to be 56 W.
For the temperature distribution on the faces of a 1-mm slab, the maximum temperature
occurs at the pump faces and is 54.5 °C. In contrast, the
temperature distribution on the faces of a 5-mm slab shows a maximum temperature of 186.8 °C. This is to be
expected as the thermal resistance between the heated region and the heat sink has been
increased due to the presence of additional YLF. For the 1-mm slab the peak principal
stress is 2.2 x 108 dynes/cm2 and is 8.3 x 108 dynes/cm2
for the 5-mm slab. In the case of the 1-mm
slab, the peak principal stress occurs on the faces of the slab in contact with the heat
sink (the x-z faces). This is also true for
the 5-mm slab but now additional stress has built up on the pumped surfaces of the slab.
From these results, we conclude that there is
considerable benefit in using as thin a slab as possible.
Practical limits to reducing the slab thickness are the ability to polish thin
slabs, and diffraction of the propagating laser mode.
For higher gain laser media, such as Nd:YLF, the likelihood of parasitic
oscillation may also be increased by thinning the slab.
2. CW Er:YAG laser
The first task we undertook in the development of the Er:YAG laser was to measure the absorption coefficient of the laser material using our Perkin Elmer Lambda-9 spectrophotometer. Measurements were performed on YAG crystals doped with 1% and 16.7% of Er. Measured absorption coefficients in 950-990 nm range are given in Figure 4 for 1-%-doped material. Based on this data, we chose 967 nm as the optimum pump wavelength.
Figure 4. Absorption coefficient for 1% Er:YAG.
The optical layout of the Er:YAG laser is shown in Figure 5. We used two SDL 1.5-W, external-cavity diode lasers to end-pump a monolithic Er:YAG crystal. The output beam, horizontally-polarized, from each diode laser passes through a single-stage Faraday isolator to prevent any possibility of laser diode damage due to back reflections from different optical components. As the polarization of each beam after the isolator is rotated by 45 degrees, a half-wave plate is put in each beam path to rotate the polarization to the desired orientation. Polarization of one beam is set horizontal and the other one vertical. This allows us to combine the output beams of both lasers into a single beam using a polarization beamsplitter. The beamsplitter is shown on an adjustable mount. The combined beam is focused into the diffusion-bonded Er:YAG crystal with a ~ 50-mm-focal-length lens. The laser crystal is shown held in a simple clamp, mounted on a multi-axis adjuster to allow alignment of the crystal relative to the pump beam.
Figure 5. Layout of
the Er:YAG diode-pumped laser.
In initial experiments, we were able to obtain nearly 100 mW of cw power from the Er:YAG laser. However, the laser threshold was higher and slope efficiency lower than anticipated from prior work on cw-pumped Er:YAG lasers. We traced some of the problem to unexpectedly low beam quality from the SDL lasers. Given the short, phonon-limited upper-state lifetime, complex pumping dynamics and low gain in Er:YAG, it is crucial for efficient cw operation to have as high a pump density as possible. We had chosen the external-cavity diode lasers for their claimed near-diffraction-limited beam quality. We examined the pump-laser output beams using a Spiricon M2-101 beam profiler. Both lasers were more highly multimode than specified in the horizontal, with horizontal M2 values between 3.5 and 4.5. Vertical beam quality was better with M2 values between 1.7 and 2.6, but still far from the diffraction-limit. The poor beam quality led to much larger pump spots and resultant lower pump intensities than expected. In addition, we experienced problems with coating degradation on the Er:YAG crystals, as well as possible high losses from the diffusion-bonding process. At the conclusion of the program we delivered a packaged cw Er:YAG laser capable of 50 mW of output at 2934.6 nm
3. CW Er:YLF Laser
In order to determine the optimum center wavelength for the pump diode laser bars we performed polarized high-resolution (Dl ~ £ 1 nm) absorption measurements in 960 –990 nm range on YLF crystals doped with 20% of erbium. The absorption band in this spectral range corresponds to the Er3+ ion transition 4I15/2 –> 4I11/2 from the ground level to the upper laser level. Measured absorption coefficients in 960-990 nm range for sigma- and pi- polarization are given in Figure 6.
Figure 6. Absorption coefficient for 20% Er:YLF (s- and p-polarization).
Based on these measurements we selected a nominal pump wavelength of 977 nm to provide the appropriate level of absorption in the side-pumped crystal.
We modified the mechanical design of our standard Gain Module using SolidWorks™. Because of the difficulty in obtaining low-loss AR coatings in the 2800-nm wavelength region, we changed the laser crystal design to use Brewster-angle end faces. Figure 7 is a view of the new design showing the Brewster-slab and the two diode lasers. The slab-top cooler is not shown since it hides the laser crystal in this view. The two vertical holes in the slab pedestal are the water flow channels for the top-cooler. The diode-lasers, mounted on their copper coolers, are offset so that they each pump a 1-cm portion of the slab. The diode-coolers and the slab pedestal are mounted on a block of PEEK, which insulates the diodes electrically and distributes the cooling water to the diode-coolers and the slab pedestal. Figure 7 includes cube-shaped HR-mirrors mounted onto angled-arms, which in turn are mounted onto small kinematic mounts. This allows the HR's to be finely adjusted. This design uses many standard Q-Peak parts, and was therefore relatively inexpensive to design and build.
>
Figure 7.Assembly drawing of the Brewster-Gain Module.
Figure 8 shows the general resonator design we used with the Er:YLF gain module. The bars, with fast-axis collimation lens included, operated in the 975-978-nm wavelength region and were capable of nearly 30-W of cw power each.
>
Figure 8. Layout of side-pumped Er:YLF laser resonator.
The input/output curves for our CW-operated Er:YLF laser with 5-cm concave mirrors and an 8.5-cm resonator length are shown in Figure 9, for two cases. In the first case we used the HR mirror and a partially reflective mirror with an outcoupling of 0.8%. In the second case we used two partially reflective mirrors with a total outcoupling of 1.6% and measured combined power of both laser beams emitted in forward and backward directions. With two-bar pumping we were able to achieve ~1.8 W of output power with two partially reflective mirrors. The beam was diffraction limited in vertical plane and multimode in horizontal plane. We also characterized the wavelength of the Er:YLF laser in these experiments. In all cases under CW pumping the laser operated at 2810 nm. We limited pump power to 10 W per bar over concern that we might fracture the Er:YLF slab. In earlier experiments, with 20-%-doped Er:YLF, we did observe slab fracture at 20 W of total pump power. However, we suspect that this material had flaws that led to premature fracture. Because of limited resources, we chose not to experimentally determine the fracture limit of the 15-%-doped material.
Figure
9. I/O curves for CW side-pumped 15% Er:YLF slab laser with 8.5-cm resonator length
and 5-cm concave mirrors.
4. Pulsed Er:YLF laser
We also arranged the resonator to produce three passes through the Er:YLF crystal, similar to our standard MPS Nd:YLF laser design. This led to improved beam quality in the horizontal direction. Without making any changes to the 3-pass resonator, we operated the laser in QCW mode. The repetition rate was set to 50 Hz and the pump pulsewidth to 10 ms, which corresponds to 50% duty cycle. The input/output curve for QCW-operated 3-pass Er:YLF laser with 10-cm concave mirrors is shown in Figure 4.28. As one can see, at maximum peak pump power of 27 W we achieved an energy of 24 mJ per pulse at 50 Hz.
>
Figure 10. I/O curves for QCW side-pumped 3-pass 15% Er:YLF
slab laser at 50 Hz.
We packaged two Er:YLF lasers in a configuration similar to our commercial Nd:YLF product. The optical head design appears below in Figure 11, and includes the Brewster-angle Gain Module, cavity optics in adjustable mounts and safety shutter for CDRH compliance.
Figure 11. Optical head design for Er:YLF lasers.
We carried out calculations to determine thermal and stress distributions in a side-pumped, slab-crystal laser designs. Both temperature rise and stress are minimized by using the thinnest possible crystals.
We were able to operate cw, diode-pumped, 2936.4-nm Er:YAG lasers with power outputs in the 50-100-mW range. The commercial (SDL) external-cavity pump lasers we used provided poorer beam quality than expected, which led to higher thresholds and lower slope efficiencies than desired. Subsequently, SDL has ceased manufacturing of the external-cavity lasers, and future work on diode-pumped, cw Er:YAG lasers will require development of alternate, high-brightness pump sources.
By using a side-pumped design, we were able to generate a record level of cw output from an Er:YLF laser, 1.8 W at 2810 nm. We also operated the system in a pulsed-pumped mode at 50 Hz, and generated 24 mJ/pulse. Higher outputs are likely possible, but determination of the maximum safe operating power requires reliable data on the fracture strength of Er:YLF crystals.