SPIE Digital Library Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Defense + Commercial Sensing 2017 | Call for Papers

Journal of Medical Imaging | Learn more


Print PageEmail PageView PDF

Lasers & Sources

Spin Control

Spinning-disk designs mitigate thermal problems in solid-state lasers.

From oemagazine June/July 2005
31 May 2005, SPIE Newsroom. DOI: 10.1117/2.5200506.0006

Leaders in the defense community continue to be interested in understanding the potential of high-power lasers (greater than 1 kW) for military applications. The promise of light-speed, precision engagement of soft targets, along with the multitude of other possible applications enabled by alternate wavelengths, powers, and beam formats compels technologists to consider laser-based systems in their planning. Although chemical lasers have been shown to be capable of very high output powers, compact, electrically driven solid-state lasers appear to be the enabling technology for many of the envisioned scenarios.

In spite of the tremendous potential of high-power solid-state lasers for field applications, problematic thermal effects in the gain media and general heat removal from the laser device continue to constrain their utility. Under operating conditions, nonideal or fluctuating temperature distributions in the laser material can alter the optical properties and lead to thermal lensing, beam distortion, instability, depolarization, and loss of efficiency.

The source of heat in the laser material is the absorption of intense diode-laser pump light, which is used to excite a particular transition and produce the population inversion necessary for lasing. The quantum defect (the difference in energy between a pump photon and a lasing photon), along with other material-specific, nonradiative relaxation processes, can cause significant volumetric heating. Adequate heat removal from a typical small piece of gain material generally requires the use of bulky, high-pressure flowing-liquid heat exchangers. The densely packed stacks or arrays of diode-laser pump sources commonly used to pump solid-state lasers currently operate at around 50% electrical-to-optical efficiency and require similarly aggressive cooling.

Achieving high overall laser device efficiency (laser power out divided by the electrical power applied to the diodes, chillers, controls, etc.) is critical to minimizing the volume and weight of a packaged high-power laser. Furthermore, since many applications require long-range propagation, any deviation from an ideal output beam effectively reduces the practical device efficiency by a beam-quality factor. It is therefore particularly important to minimize thermo-optic effects that degrade beam quality. Required run times, ranges, output-aperture sizes, beam quality, powers, and other system-related factors can be traded off to some extent against device size, but in most cases, the thermal management subsystem still constrains packaging and integration options. If we optimize the way waste heat is removed from both the diode-laser pump system and the laser gain media, more practical device configurations may become feasible.

Using fiber-coupled diode lasers instead of large diode-laser arrays to pump solid-state lasers, for example, allows the heat load from the diodes to be optimally distributed over much larger surface areas, possibly enabling forced-air or lower-heat-flux liquid cooling. Fiber-coupled pumps also enable system integrators to decouple the pump system from the laser head, adding more packaging flexibility. The relatively uniform circular intensity distribution exiting a multimode fiber is generally easier to match to a resonator, leading to better efficiency and beam quality. In addition, modular, replaceable fiber-coupled pump units could have output ends that are optimally positioned to pump the gain media. Although available diode-to-fiber coupling efficiencies need to be improved and the cost per watt of fiber-coupled diodes is still roughly 10 times higher than that of standard bars or stacks, this type of pump system may be needed in order to meet the system-level requirements of certain high-power, solid-state laser platforms.

Turn, Turn, Turn

Along with other important emerging technologies like high-power fiber lasers, cryogenic lasers, high-efficiency diode lasers, and thin-disk lasers, spinning-disk lasers offer the possibility of high efficiency, good beam quality, and relaxed cooling requirements.1-3 The spinning-disk solid-state laser concept was conceived in the 1980s as a subset of a more general "moving media" laser approach. This technique controls thermal loading by continually changing the position of a large piece of laser material to distribute waste heat from the pumped region over an entire slab.4 The waste heat is then removed from the large surfaces using low-heat-flux cooling methods. Rotating a round disk of laser material provides a simple way to implement the technique.

Our group at the Air Force Research Laboratory's Directed Energy Directorate (Kirtland AFB, NM) has been investigating the unique properties of spinning-disk lasers to determine if they might lead to a better way of building practical, high-power solid-state lasers for certain types of applications. A spinning-disk laser typically consists of a disk of gain media on the order of several centimeters in diameter and several millimeters thick, center-mounted and spun at a few hundred rotations per minute (see figure 1). As the disk rotates, waste heat is transported into a cooling zone defined by two water-cooled plates positioned parallel and in close proximity to the disk faces. The disk is cooled primarily by conduction across the gap region, which can be filled with high-thermal-conductivity gases like helium. Other cooling techniques, such as leveraging boiling liquid sprays or forced convection, could also be used.

Figure 1. A typical spinning-disk laser (top) includes cavity mirrors at Brewster's angle. Multiple spinning disks (green arrows, bottom) can be combined in a single cavity. The pump light is introduced from the side through the lenses visible in the photo. Pumping and lasing takes place in the exposed part of the disks. The unexposed portions of the disks are cooled as they rotate between two water-cooled plates.
(Photo Air Force Research Lab)

By setting the uncoated disks at Brewster's angle, we obtain low loss for p-polarized lasing and can introduce pump light essentially perpendicular to the disk faces. Because the disk is moving, the pumped volume (in the lab frame) does not have to be perfectly matched to the laser mode; thus we can introduce many pumps at slightly different angles without causing severe loss of laser efficiency. In fact, for materials with longer radiative lifetimes (τsp), pumping can take place off of the resonator axis, allowing spatial multiplexing of pump light along an arc leading to the lasing zone. The large, open aperture of the exposed gain media provides plenty of room to multiplex pump beams as well as very low diffractive loss for the circulating laser mode.

We have performed experiments on several different laser materials emitting around 1 µm, including neodymium-doped yttrium aluminum garnet (Nd:YAG), Nd:glass, ytterbium-doped YAG (Yb:YAG), and Nd:YAG ceramic. Most of the experiments were based on disks of about 60 mm in diameter and a few millimeters thick. We used the low-thermal-conductivity Nd:glass disk to test our thermal profile model because substantial temperature rises could easily be produced and measured.5 In the higher-thermal-conductivity YAG materials, the temperature rise in the pumped region was only a few degrees centigrade at absorbed power densities of several kilowatts per cubic centimeter.

In an early experiment, we obtained 30 W of linearly polarized, single-transverse-mode output power from an Nd:YAG spinning-disk laser that operated at 61% slope efficiency. The optimum disk rotation speed was approximately 400 rpm. An Nd:YAG ceramic disk provided similar performance. The relatively recent advances in the fabrication of ceramic laser gain media should soon lead to the availability of large uniform disks, which will be ideal for power-scaled spinning-disk lasers.6 The smaller quantum defect Yb:YAG material displays two to three times less heating than Nd:YAG and may be capable of producing more laser power for a given volume of disk material. Many other laser materials should also perform favorably in spinning-disk configurations.

Rotation Tradeoffs

To optimize the performance of a spinning-disk laser, we need to consider how the motion impacts the heat transport and the laser dynamics. Using a simple 1-D rate-equation analysis, we can show that for lasing well above threshold, the laser output power P( ƒ ) in the case in which the pump and laser beams are overlapped is approximately

where Roc is the output coupler reflectivity, Isat is the saturation intensity, R is the rate of absorption, r is the radial position on the disk, ƒ is the disk-rotation frequency, and d is the mode-matched spot size. Nth represents the threshold inversion, which may be rotation-frequency dependent due, for example, to the higher steady-state temperatures in the lasing region at slow rotation frequencies.

In equation 1, we assume the gain is uniformly saturated by the laser mode, and the linear rotation loss from the exit of the lasing region is as an exponential decay. The first two terms represent the familiar static laser output power equation, while the third term accounts for the observed falloff in power for a spinning-disk laser operating at higher rotation frequencies (see figure 2). We can derive similar formulas for cases in which the pump beam is offset from the laser beam. Closer to threshold, one must consider the details of the steady-state saturated gain versus position in the lasing region.

Figure 2. Measured and calculated laser output power (squares) and depolarization loss (triangles) versus disk spin frequency in a ~ 10 W Nd:YAG laser show a pronounced falloff above a certain frequency. A 95% reflective output coupler was used, so the maximum internal loss off the four Brewster-angle surfaces was less than 0.25%, and much lower at higher spin frequencies.

If the threshold inversion

Nth( ƒ ) = - 1n[Roc(1 - L( ƒ ))] / 2lσ [2]

includes a frequency-dependent loss contribution L( ƒ ) of the form

L( ƒ ) = a + (b/ƒ ) [3]

we obtain a reasonable fit to our laser-power and loss-versus-spin-frequency data. In equation 2, l is the cavity length, and s is the stimulated emission cross section. The observed loss, measured by monitoring the reflections off the disk surfaces, is likely caused by slight thermal-stress-induced depolarization of the laser mode by the gain media.

Clearly, spinning-disk lasers should be operated well above threshold in order to minimize the loss in population inversion due to rotation. In addition, parameters like gain material, pump spot size, pump spot position, disk size, cooler configurations, optimum rotation frequencies, and so on need to be adjusted for optimum extraction efficiency and beam quality. Further power scaling can be accomplished by combining several spinning-disk laser heads into a single resonator or amplifier chain.

The spinning-disk laser technique offers many of the important thermal control and efficiency advantages needed for future portable, high-power solid-state laser devices. These advantages come at the expense of introducing mechanical motion into the laser cavity, however. We currently use both air-bearing spindles and electric motors to rotate the disks. Still, we observe some power fluctuation, beam jitter, and, in some cases, power dropouts, all of which are periodic at the disk rotation frequency. The power fluctuations and dropouts are caused by nonuniformities or voids in the doping profiles, variations in the crystal structure, or imperfect disk surfaces. Our current position on beam jitter is that a large portion can be engineered out using more careful disk fabrication, better mechanical rotators, and jitter-minimizing laser cavity or amplifier configurations. Any remaining jitter can then be removed using relatively straightforward beam-control techniques. oe


D. Ripin et al, Optics Letters 29, p. 2154 (2004).

C. Stewen et al., IEEE J. Selected Topics in Quant. Electron. 6, p. 650 (2000).

S. Massey et al., Diode Pumped Nd:YAG and Nd:Glass Spinning Disk Lasers, JOSA B (accepted for publication in May 2005).

S. Basu and R. Byer, Appl. Opt. 29, p. 1765 (1990).

A. Paxton et al., Proc. SPIE 5333, p. 12 (2004).

J. Lu et al., Appl. Phys. Lett. 77, p. 3707 (2000).

Harold Miller, Alan Paxton, Steven Massey
Harold Miller and Alan Paxton are research group leaders and Steven Massey is an expermentalist in the Directed Energy Directorate of the Air Force Research Laboratory, Kirkland AFB, NM.