Special Section on High-Energy Laser Systems and Components

High-power diode-pumped solid-state lasers

[+] Author Affiliations
Steven R. Bowman

Optical Materials & Devices Branch, US Naval Research Laboratory, Washington, DC 20375

Opt. Eng. 52(2), 021012 (Oct 10, 2012). doi:10.1117/1.OE.52.2.021012
History: Received July 2, 2012; Revised September 10, 2012; Accepted September 11, 2012
Text Size: A A A

Open Access Open Access

Abstract.  Diode-pumped solid-state laser technology is reviewed with an emphasis on laser materials and concepts important to high power systems with excellent beam quality. A first principal approach to power scaling is taken in order to summarize and compare the wide variety of laser architectures. Successful approaches are highlighted and directions for future work are discussed.

Figures in this Article

Of all our technologies, lasers have the unique ability to concentrate tremendous power onto extremely small targets. This property is called brightness and is a measure of a beam’s intensity and directionality. Since the invention of the laser in 1959, scientists have been dreaming up ways to push them to ever higher powers and brightness. Even after 50 years this continues to be a very active area of research. Each year sees significant performance improvements and new capabilities. Advances in materials, techniques, and components move through an accelerated R&D process and into a flourishing world-wide industry. Solid-state lasers are incorporated broadly into manufacturing, communication, biomedical, and defense applications. This paper will review some of the principal innovations that have led to these remarkable sources.

The path to higher laser brightness has taken two distinctly different routes; pulsed and continuous operation. This paper focuses on steady-state operation, but the pulsed regime is perhaps even more exciting. Advances in short pulse lasers have pushed peak powers upward at a very rapid pace. Here power scaling has come principally by moving toward ever shorter laser pulses. Even a modest millijoule laser pulse can achieve terawatts of peak power when compressed to a few femtoseconds. Petawatt peak powers have been demonstrated at several laboratories and plans are being developed to push to the exawatt level (1018W).1 When focused to a diffraction limited spot such a system could achieve a staggering 1025W/cm2. This would be the equivalent of briefly focusing the total power emitted from the sun onto the surface of a dime, very bright indeed.

By comparison, progress toward high average power has been much slower. Unlike the short pulse lasers, average power systems have to achieve some sort of equilibrium with their surroundings. At high powers any inefficiency in the lasing process can generate tremendous heating inside the laser medium. The resulting temperature gradients will distort the laser and have proven particularly difficult to engineer around. One solution is to flow the lasing medium, as in a gas laser. While pressure gradients and flow instabilities create some distortion problems, early progress with gas laser systems demonstrated their scaling potential with sustained megawatt powers in a single beam. Nevertheless, high average power chemical lasers proved to be difficult and expensive to maintain, since they require transport and storage of large volumes of hazardous materials. Recent efforts at average power scaling have emphasized electrically driven solid-state lasers. These lasers can readily utilize existing power networks and are much easier to field and maintain. Practical scaling of these sources has been made possible by the remarkable advances in semiconductor laser technology.

Three decades of progress in semiconductor research have yielded electrically pumped lasers that now span most of the electro-magnetic spectrum from the deep blue to the long wave infrared. Gallium arsenide (GaAs) devices are the most mature with multi-watt single-facet emitters routinely operating at over 70% electrical efficiency. With service lifetimes measured in years and production costs of only a few dollars per watt, GaAs-based laser diode arrays can now deliver nearly unlimited continuous optical power in the spectral range from 0.8 to 1.1 µm.

While laser diode arrays can be powerful, efficient, and practical; they are not particularly bright. With their limited coherence and complex mode structure, simple rack-and-stack diode arrays operate with a brightness of around 2kW/cm2-sr. (For comparison purposes unobstructed sunlight at the earth has a brightness of roughly 600W/cm2-sr.) Careful optical packaging can preserve most of the single emitter’s brightness and fiber coupled diode arrays routinely achieve a respectable 5MW/cm2-sr. Unfortunately, maintaining a laser diode’s brightness increases its cost by roughly an order of magnitude. When scaling to high average power, system cost is a primary concern.

The currently preferred approach for increasing laser brightness is to use the electrically pumped diode arrays to optically pump another solid-state laser. The optically pumped laser medium enhances energy storage, coherence, spatial mode control, and thermal management. This approach can produces much higher levels of brightness at lower costs. For example, a one kilowatt single-mode fiber laser generates nearly 100GW/cm2-sr. This is 20,000 times brighter than the laser diodes that pump it. State-of-the-art bulk solid-state lasers pumped with rack-and-stack diode arrays generate even higher levels of brightness amplifications. How these levels of brightness amplification are possible is the subject of this paper. The focus will be on optically pumped lasers which produce a single high quality beam. Important techniques for further scaling through the addition of high brightness beams will be reviewed separately.

In the half century since the invention of the ruby laser, many thousands of solid-state laser materials have been developed. Hundreds of crystals, glasses, and ceramics have been introduced and refined to laser quality. These host materials have been doped with one or more elements to absorb pump light and generate gain. The lanthanide series has proven to be particularly prolific as active ions with demonstrated lasers transitions in 12 of the 14 elements. The shielding of the 4f electrons in the rare earth ions produces numerous metastable energy levels which can store absorbed energy on the 0.1 to 1 millisecond time scale. Over 80 distinct rare earth laser transitions have been reported spanning the spectral range from 0.17 to 7.2 µm.2 Most of these laser transitions can be pumped with existing laser diode technology either directly or via sensitization through nonradiative energy transfer. This makes them potential candidates for high average power systems.

It should be noted that another important class of solid-state lasers have been develop based on transition metal ions, specifically Cr, Ti, Co, and Fe.3 These lasers generally exhibit much broader emission spectra than those produced by the rare earth lasers. This makes them ideal for tunable and ultra-short pulsed systems. Unfortunately, transition metals typically store energy for no more than a microsecond at room temperature. This means they require a pump source brighter than currently available from laser diode arrays. High average powers from transition metal lasers will therefore have to wait for the development of a practical pump source.

With so many solid-state laser materials to choose from, it is remarkable that only a handful of these lasers have ever been scaled to kilowatt average powers. Virtually all successful efforts to scale up solid-state lasers have been based on only two rare earth transitions, one in neodymium and one in ytterbium. Indeed, most recent efforts have focused on only two laser materials; neodymium doped into Y3Al5O12 (NdYAG) and ytterbium doped into amorphous silica (Ybsilica). It is instructive to consider the reasons for the selection of these materials.

Neodymium Doped Laser Materials

By far the most mature rare earth lasers are based on the Nd3+F3/24I11/24 transition shown in Fig. 1. This 1.1 µm laser scheme is referred to as a four-level system since the lower laser level is strongly quenched via electron-phonon coupling. In a four-level system the lower laser level will have negligible population and the laser emission is generally free from absorption resonances. As a result four-level lasers operate with relatively low excitation densities.

Graphic Jump LocationF1 :

The lower lying energy levels of Nd:YAG are plotted in wavenumbers. Energy manifolds are labeled according to their 4f electronic quantum numbers, S, L, and J using the notation L2J+12S+1. Manifolds are also numbered according to the four-level laser scheme with the upper laser level in red and lower in green. Pump and laser transitions are illustrated with arrows and labeled with the wavelength of the peak cross section.2

The upper level of the neodymium system can be directly pump or excited via a dense set of quenched higher lying levels. This gives neodymium a broad set of pump bands which were critical to it early success in lamp pumped systems. This success continues today due to the match of neodymium’s 0.81 µm pump band to efficient laser diode emission wavelengths. The metastable F3/24 manifold consists of a single pair of closely spaced energy levels leading to narrow emission lines. A short radiative lifetime combined with a high branching ratio gives the 1.1 µm transition a large stimulated emission cross section. The resulting high gains make Nd3+ ideal for compact systems and short pulse generation. These characteristics have made the neodymium 1.1 µm transition the workhorse of the solid-state laser industry.

Now consider the problem of scaling a diode pumped neodymium lasers to high powers. To operate a four-level laser continuously and efficiently, the laser intensity in the gain medium must exceed the laser saturation intensity, ILsat.4 This is most easily seen from the simple relation between the steady-state stimulated emission rate, Wstim, and the pump rate, Wpump: Display Formula

Wstim=ILIL+ILsatWpump,whereILsathvLσLτ3.(1)
Here hvL is the laser photon energy, σL is the stimulated emission cross section, and τ3 is the upper state lifetime. Reasonably efficient stimulated emission clearly requires operation at several times ILsat. NdYAG, for example, has a laser saturation intensity of 2.8kW/cm2, so operation at 10kW/cm2 should produce good efficiency. Optical efficiency for NdYAG under these conditions should approach 50%.

Practical operation requires the laser medium also have adequate saturated gain. In order to maintain efficiency and gain the laser must be pumped with a sufficient intensity. We can estimate the required pump intensity from the steady-state gain coefficient given by Display Formula

g=αPλPλLIP(IL+ILsat).(2)
Four-level gain scales linearly with the pump intensity and absorption coefficient, αP, but saturates as laser intensity grows. Note that the emission cross section appears only through the saturation intensity. Higher emission cross section materials lower saturation intensity, but does little else to facilitate steady-state power scaling.

While the selection of the desired gain depends on the laser architecture, consider two important cases for high power systems; a power oscillator and a power amplifier. For the case of a power amplifier, a gain coefficient of roughly 0.2cm1 allows efficient single-pass extraction from a 10-cm long crystal. Assuming the 10kW/cm2 intensity at 1.064 µm and a 4cm1 diode pump absorption at 0.81 µm, yields a minimum pump intensity of 840W/cm2. For the case of a power oscillator, lower gain is appropriate but laser intensities will be significantly higher. Assuming a gain of 0.05cm1 and a circulating laser intensity of 100kW/cm2 requires a pump intensity of 1.7kW/cm2.

These relatively low pump intensities illustrate the principal advantage of power scaling the neodymium laser. Neodymium can be efficiently pumped with low brightness diode pump arrays. This keeps system costs down. The trade-off for using low brightness diode arrays is that it necessitates increasing the laser aperture. Further, the relatively low laser saturation intensity of NdYAG requires centimeter exit apertures in order to generate multi-kilowatt beams. These large apertures become an issue at high powers since the dissipation of waste heat slows as the square of laser medium thickness. We will examine the problem of thermal management in the next section.

Excess heat generation is one of the few down sides to four-level laser operation. Waste heat is generating during the rapid quenching of the pumped level and the lower laser level. Limited efficiency and the resulting heat are the price to pay for the low thresholds and high gain. The thermal loading of the gain medium is simply proportional to the absorbed power density Display Formula

Q=χhαPIP.(3)
For neodymium lasers, the fractional heat generation, χh, is largely determined by the laser quantum defect, χQD=1λP/λL. For the wavelengths discussed above NdYAG has a quantum defect of 24%. Nonradiative loss mechanisms increase the measured value of χh to 0.32 during laser operation and it rises to roughly 0.4 in the absence of lasing.5,6 The thermal loading expected for the power oscillator described above would be expected to be 2.2kW/cm3. Managing this level of thermal loading without losing laser brightness is the principal challenge of scaling to high brightness.

The discussion here has focused solely on steady-state operation. Of course, it is also possible to scale to high average powers in a pulsed, high repetition rate system. Energy storage and pulsed extraction allow for efficient operation with average pump intensities comparable to the steady-state minimums. Nevertheless, reducing the duty factor increases the peak power required from the pump source. This typically increases the number of diodes and system cost. Thermal transients adversely affect all components in the system and peak power creates additional problems for the optical train. High fluence drives the selection of the laser architecture as discussed in the next section. Further, energy storage complicates stability issues in large mode volume systems. Suppressing parasitic oscillations, isolating gain elements, and avoiding intensity spikes significantly complicate the design. For applications in which the goal is high average power, continuous wave (CW) operation helps avoid most of these problems.

As mentioned earlier, many other host materials for neodymium have been developed over the years, but none has yet demonstrated superior high power performance to NdYAG. The properties of several interesting materials are compared to NdYAG in Table 1. For example, significant effort has been invested in birefringent hosts like YAlO3 and YLiF4.2 These naturally birefringent materials do not exhibit the thermal depolarization that can occur in optically isotropic crystals like YAG. This property proved useful in early lamp pumped rod lasers. Nevertheless, at high average powers anisotropic laser hosts typically exhibit complex astigmatic wavefront distortions. As a result, interest in these laser materials has declined. Recent material efforts have focused on isotropic hosts with higher thermal conductivity, κ, such as Y2O3, and Lu2O3.7 Sesquioxide and other high κ crystals hold considerable promise for power scaling, if growth issues can be resolved.

Table Grahic Jump Location
Table 1Neodymium laser material properties at 293 K parameters as defined in the text. Peak values of spectral parameters are listed followed by the polarization direction.

At least as important as its bulk material properties are a laser host’s availability, size, and quality. This has historically been NdYAG’s greatest advantage. Commercial markets developed initially for lamp pump lasers have provided continuous support for the production of large high quality YAG. This industry has developed refinements to the growth which have been adapted for high power efforts. Large scale commercial processing has improved YAG’s chemical purity, critical to controlling nonradiative losses. Refined Czochralski growth has reduced static wavefront distortions, particularly important in large volume systems. Nevertheless, recently developed high quality ceramic YAG has challenged the laser crystal growth industry.

Ceramic YAG can now be produced with higher optical clarity and size than the best single crystal material.8 Formed by sintering of compacted nanopowders precursors, ceramic YAG allows for better control of dopant distribution than single crystal growth. The spatial distribution of the dopant can be highly uniform or graded as desired. Higher dopant concentrations are possible without the ionic segregation and strain developed during growth from a melt. Further, since ceramic YAG is macroscopically isotropic it does not exhibit the thermal depolarization seen in crystalline YAG lasers. Ceramics are generally more resistive to stress fracture than single crystals, but have somewhat reduced thermal conductivities. For that reason, recent efforts to develop ceramic sesquioxides appear to be very promising.912 Still the availability of ceramic hosts for high power scaling will likely depend on market forces in the wider laser industry.

Ytterbium Laser Materials

Only one other solid-state laser can match the power and brightness now available from neodymium lasers: the ytterbium 1.0 µm laser. First reported in 1967, this laser attracted little attention until the advent of high brightness diode pumps.13,14 While conceptually similar to the neodymium system, the properties of ytterbium have led researcher down a different path to high powers. The key difference is shown in the electronic energy diagram of Fig. 2; ytterbium has only one excited state manifold. Pump and laser must share the only ΔJ=1 transition, the Yb3+F5/24F7/22. Optical pumping to population inversion is only made possible by Stark splitting of the J manifolds by the local electric field. Still, the Stark splitting is rather small and not much larger than the thermal energies at room temperature. As a result, ytterbium is commonly referred to as a quasi-three level laser.

Graphic Jump LocationF2 :

Energy structure and important transitions in Ytterbium doped silica.15

Ytterbium’s energy structure has several significant implications for high power operation. As mentioned above, the terminal laser level is close to the ground state. This means that the lower level is partially occupied through thermal excitation. In order to minimize this effect at room temperature, oxide hosts have generally been selected for ytterbium by virtue of their large ground state splitting. Properties of the two most important materials are summarized in Table 2.

Table Grahic Jump Location
Table 2Room temperature ytterbium laser material parameters as defined in the text.

Once again, YAG has played an important role in scaling bulk ytterbium lasers. Several techniques have been employed to push YbYAG up to kilowatt powers.16,1719 For YbYAG, the top of the ground state sets at 785cm1, almost 4 kT at room temperature. This reduces thermal population of the terminal laser level to 2%, but still threshold losses are significantly higher than in four-level neodymium.

So far, the more successful host for power scaling ytterbium lasers has been fused silica. Silica has even higher ground state splitting than YAG, which combined with its inhomogeneous nature, produces even broader spectral bands and lower thresholds. Most importantly, silica can be readily fabricated into low loss optical fibers, a format in which ytterbium excels. High power silica fiber lasers now exceed the 10 kW level in single transverse mode operation.20,21

The simplicity of ytterbium’s energy structure is critical to its successful performance at high power. The absence of higher lying levels avoids some of the pitfall experienced by other laser schemes. Additional energy levels can produce undesirable resonances which disrupt power flow. Upconversion, cross-relaxation, excited state absorption, or two photon absorption are all inhibited by the absence of a higher energy level. Doping levels of ytterbium can be increase without the strong concentration quenching exhibited by neodymium. Still the purity of the host material is critical. Since the ytterbium excitations can migrate through resonant energy transfer, even a trace amount of unintentional dopant with a 1 µm resonance will lead to significant nonradiative losses. Chemical purity is an important aspect to the success of YAG and silica as high power hosts.

The broad emission and absorption spectra of ytterbium affect performance in important ways. The wavelength agility of the system is impressive. Ytterbium silica is readily tuned by more than 10% of its wavelength. Not only does this make ytterbium ideal for ultra-short pulse generation, but it also allows CW laser performance to be refined. Laser designers have broad latitude in the selection of the operating wavelengths and by extension can control excitation density or quantum defect heating. A few illustrative wavelength options are summarized in Table 2.

To see how the quasi-three level nature of ytterbium affects laser performance, consider again the saturated laser gain. Start by defining some fundamental parameters. Following the notation in 22, we shall use σL and σP for the effective absorption cross sections for the pump and laser lines. The term effective means that these are the cross sections inferred directly from the absorption spectra at a specific temperature. The fraction of ions that must be excited to bleach out absorption at the laser wavelength is labeled βL. Similarly, the excitation fraction at which the pump absorption bleaches out will be βP. These fractions are easily computed assuming a Boltzmann distribution of ions in each energy manifold. With these definitions the steady-state laser gain coefficient can then be written as Display Formula

g=σLNYbβL[(βPβL)IP/IPsat-βL](IP/IPsat+IL/ILsat+1),whereILsatβLhvLσLτ2andIPsatβLhvPσPτ2.(4)
It is clear from Eq. (4) that the effective emission cross section is simply σL/βL, so the definition of laser saturation intensity is the same as in the four-level system. Unlike the four-level system, the laser gain of the three-level system also saturates with pump intensity. This is important, as ytterbium laser are usually driven very hard. We therefore define the pump saturation intensity, IPsat, in terms of the emission cross section at the pump wavelength, σP/βP. The thermal occupation of the lower laser level, βL, determines the pump intensity required to reach gain. This intensity is determined largely by the choice of wavelengths Display Formula
IPmin=βPβLβPβLhvPσPτ2.(5)

The saturation and minimum intensities vary considerably with the choice of host, wavelength, and temperature as illustrated in Table 2. Compare the pump requirements for a YbYAG power oscillator operating at 1030 nm with the earlier 1064 nm NdYAG example. Let’s assume the same saturated gain coefficient, 0.05cm1, and same circulating laser intensity, 100kW/cm2. This scenario will now require a diode pump intensity at 940 nm of 3.6kW/cm2. This is about twice as high as the NdYAG example, but still just within reach of low brightness pump arrays. An YbYAG power amplifier with 0.2cm1 gain and 3ILsat extraction would require 4.6kW/cm2, roughly 5 times higher pump intensities than NdYAG. So room temperature ytterbium lasers require somewhat higher brightness than currently available with large area pump arrays.

Heat loading in the ytterbium system is more complex than in neodymium. While the requirement for higher pump intensities increases thermal loading, the reduced laser quantum defects of the quasi-three level system largely mitigates the effect. As illustrated in Table 2, laser quantum defects vary between 0.09 and 0.02 depending on the selection of operating wavelengths. Unlike the four-level neodymium system, the laser quantum defect of ytterbium does not tell the whole story of heat generation. This will be discussed more in the last section.

Perhaps the most important property of the ytterbium system for power scaling is that the de-excitation of the lower laser level is extremely fast. Since the lower laser is only a few hundred wavenumbers above the ground state, thermal equilibriums are established on the picosecond time scale.23 By comparison, the neodymium lower level lifetime, τlow, is typically around 10 nanoseconds. As a result the steady-state stimulated emission rates can be much higher in ytterbium lasers than in neodymium lasers. In the absence of other limitations, the maximum laser intensity that can be maintained at steady-state is simply Display Formula

ImaxhνLσLτlow.(6)
For ytterbium silica this upper bound is of order 100TW/cm2 or 1MW/μm2. While no known material could survive continuous intensities of this magnitude, ytterbium lasers can clearly generate extremely high average powers from even very small laser apertures. Keeping the laser aperture small allows for much more rapid heat dissipation and suggests a waveguide approach to power scaling. By contrast, four-level NdYAG will self-terminate for intensities above 1W/μm2.

Over the years, a number of strategies for the construction high power solid-state lasers have emerged. Most of the basic concepts were developed early and have been repeatedly refined as the technology improved.24 Many of these approaches have been described in detail by several excellent reviews.4,16,25 In this section, the intent is to compare the concepts on a more fundamental level. Comparing different architectures within a common framework can give insights into investments for future systems.

As described above, the principal challenge is to control power flow within the laser in order to maximize and preserve system brightness. Laser diode pumping of rare earth doped solids has emerged as the most mature approach of converting electrical power into near diffraction limited optical power. Cost and brightness of the diodes is critical to practical scaling. Thermal management has proven to be the most difficult task and the high power laser architectures are designed to address this issue. All these designs are based on the observation that the time required for thermal diffusion scales as the square of the laser medium’s thickness. This leads to a common approach for thermal management; shape the laser medium into a high aspect ratio solid and rapidly extract heat through the thin dimension. While simple geometries have been the most successful, implementation often leads to symmetry compromises and complexity. Nevertheless, useful insights can be found through a simple first principal analysis.

All important high power laser architectures can be grouped in four simple categories as illustrated in Fig. 3. We characterize the architectures by the direction of flow of optical power relative to the direction of heat flow. The aspect ratio of the laser medium will be labeled Λ and is assumed to be much greater than unity. Steady-state operation requires conservation of power within the laser medium. Simple relationships between the power fluxes into and out of the laser medium can be defined for each category. These relationships aid in the discussions of the pros and cons of architectures within each category.

Graphic Jump LocationF3 :

Generalized high power laser architectures with arrows indicating power flow of the pump, fluorescence, laser, and heat.

The simplified intensity relationships listed below are based on the following assumptions. Firstly, aspect ratios are high enough to permit the infinite solid approximation. This is certainly not true when considering wavefront propagation, but can be used in power flow analysis. Secondly, heat loading is assumed to vary proportionally with absorbed pump power. As seen in the last section, this is a good approximation for neodymium although ytterbium systems can be more complex. Third, fluorescence emissions are approximated as flowing in the same direction as heat. This is a reasonable simplification since spontaneous emission uniformly illuminates the surface of the solid, which is dominated by the largest faces. For most laser configurations fluorescent emission is trapped within the medium and pump chamber where it adds to the waste heat. Highly saturated gain media produce much less spontaneous emission, but the high fluorescence intensities generated require careful consideration for each specific system. Finally, intensities are assumed to be uniformly distributed across the apertures. Uniform intensity is often essential to successful high power operation in order to maintain clean wavefronts. For cases where only a portion of the solid is illuminated, the definition of Λ should be amended to the aspect ratio of the mode volume.

Face-Pumped Disk Lasers

The simplest and one of the more successful high power architectures are usually referred to as face-pumped disk laser. Various other terms of art have been applied to highlight distinctions; active mirror, Brewster disk, thin disk, or thin-zag.2628 In this architecture, virtually all power flows through the large aperture of the disk. Heat removal occurs through one or both large faces. Often one face is cooled and the other face is left clear for laser propagation. Conservation of energy leads to a simple relationship between the waste heat flux, H, and the laser, pump, and fluorescent intensities exiting the laser medium. For the configuration shown in Fig. 3(a), the relationship is simply Display Formula

IL=(1χH)IP2IF=IP2IFH.(7)
Since all power flows through a common aperture, the aspect ratio does not directly affect scaling. For this architecture, the relationships the gain/intensity relations derived in the previous section are mostly unconstrained by geometry. Power scaling can utilize low or high brightness diode arrays as discussed earlier. The gain per disk is quite limited, so multi-disk power oscillators systems are suggested. In principle, average power can be scaled almost indefinitely by making the disk wider. In practice, parasitic oscillation and amplified spontaneous emission limit transverse scaling of a single disk.

Another practical consideration for the face-pumped disk laser is achieving efficient pump absorption while keeping the disk thin enough for heat removal. Brightness of the face-pumped disk laser is limited by thermo-optic distortions. While longitudinal temperature gradients are averaged out during beam propagation, the thermally induced strain eventually distorts the optical figure of the disk. Numerous devices have been demonstrated with neodymium and ytterbium, in pulsed and CW operation. Face-pumped thin disks lasers are now commercially available with average powers up to 16 kW with a beam quality of 2 mm-mrad using YbYAG. Record average powers of 30 kW were reported with multiple face-pumped ceramic NdYAG Brewster slabs operating in a power oscillator configuration. The resonator was water-flooded to cool the slabs and the beam quality was reported to be 3.3 times diffraction limited.29

Side-Pumped Slab Lasers

Our second category of architectures has its origins in the original ruby laser. Commonly referred to as the side-pumped rod or slab, it is defined by the laser propagation through the length of the medium and the pump power enters transversely. For this case, the aspect ratio, Λ, is the ratio of the largest area face to smallest. Heat and fluorescence also flow principally through the largest surface. With this configuration power balance requires that the net intensities scale according to Display Formula

IL=(1χH)ΛIPΛIF=Λ(IPIFH).(8)
So for this case, the aspect ratio of the gain medium effectively magnifies the pump brightness relative to the laser intensity. In other words, side pumping allows diode sources to add to the laser power in parallel. Waste heat and fluorescence fluxes are reduced by the same factor.

Important early work in laser power scaling relied principally on this approach.30,31 This remains the principal configuration for lamp pumped systems and moderate power diode pumped systems. Capable of high gain and high powers side-pumped slab lasers are primarily used as pulse or power amplifiers. Intensive cooling of the largest faces allows for high thermal loading. A serious limitation of this configuration is the strong astigmatic thermo-optic distortions that develop in the long laser medium. Self-correcting zig-zag approaches were long pursued but proved difficult and expensive to implement. Wavefront aberrations developed while propagating through the slabs edge have remained problematic for these designs. Power scaling of slabs with good beam quality remains around the 500 W level.32 The simple rod geometry continues to offer appeal. Record CW powers of 4.5 kW were reported from side-pumped NdYAG rods in a power oscillator configuration. This system had an optical efficiency of 48% with a modest beam quality of 12 mm-mrad.33

Edge-Pumped Disk Lasers

The third category of possible laser configurations is best described as an edge-pumped disk or slab. As shown in Fig. 3(c), this geometry propagates the laser through the large face, while the pump is injected transversely through the long dimension. Again, waste heat and most spontaneous emission are removed through the thin dimension. The motivation of this approach is to enhance and simplified pump coupling in a thin disk laser.34,35 Designs can be compact and allow large area beams suitable for short pulsed systems. Steady-state power flow analysis for the configuration in illustrated in Fig. 3(c) yields Display Formula

IL=(1χH)IPΛ2IF=IPΛ(2IF+H).(9)
One issue with this approach is immediately clear; edge-pumped disks effectively reduce pump brightness by the aspect ratio. In other words, each diode is required to pump a larger volume of laser material than in the other configurations. The combination of reduced pump brightness and short interaction length means that this configuration is likely to have marginal saturated gain. Further, transverse pump absorption complicates the uniform power loading needed to control wavefront distortions. There have also been thermal damage issues with injection of the high intensity pumps into the thin edge of the disk. Record average powers of 309 W have been reported with a beam quality of 2.7 times diffraction limited. This laser used low quantum defect pumping of Yb:KGdWO4 to enhance gain and reduce thermal loading.36

End-Pumped Rod Lasers

Our last category of laser architectures also dates back to the earliest days of the laser.37 It includes all end-pumped rods, slabs, and fibers. Laser and pump beams again share a common path along the length. Waste heat and fluorescence flows out transversely in one or two thin dimensions. Power balance in this architecture with an aspect ratio Λ produces an intensity relationship similar to the face-pump disk, Display Formula

IL=(1χH)IPIF=IPΛ(IF+H).(10)
Diode brightness is preserved by the common aperture. In most reports pump light is guided to maintain intensity and smooth pump deposition. This helps to avoid the pump uniformity issues that have plagued the transverse pump approaches at high powers. Waste heat and fluorescence are spread out along the length. Pump coupling and high gain are assured by the long interaction length. NdYAG end-pumped slab CW lasers have been scaled to 15 kW in a power amplifier configuration. A CW beam quality of 1.3 times diffraction limit was achieved using closed-loop adaptive optics for wavefront correction.38

An important subclass of architectures uses the commonly referred to double-clad geometry, shown in Fig. 3(d). Here a central rare earth doped core is cladded inside a larger undoped dielectric layer. We define the clad-to-core diameter ratio as γ. The multi-mode pump beam is injected into the cladding region where it is guided by total internal reflection. Pump intensity smoothes as it propagates in the high numerical aperture waveguide. Partial overlap of the pump beam inside the doped core provides the laser excitation and gain. For this subclass we should revise the intensity relation to Display Formula

IL=γ2(1χH)IPΛγ2IF=γ2IPΛγ2(IF+H).(11)
Clearly the double-clad configuration allows amplification of the pump beam intensity by the ratio of the clad-to-core areas, γ2. The partial overlap of the pump with the core reduces the pump absorption coefficient by the same factor. Efficient pump coupling is maintained by increasing the length of the gain medium. In this way, high gain and high powers can be generated in the core region. Waste power is spread thinner as the geometric factor Λγ2 increases. These laser geometries have been difficult to fabricate in bulk laser until the development of ceramic laser hosts. Recent double-clad NdYAG rods have demonstrated record optical efficiencies of 66% at the 120 W level.39

When the core diameter of this architecture is reduced to the scale of a few wavelengths, this becomes a fiber laser.40 If the product of the core diameter and numerical aperture are below 0.76λL, only a single transverse mode of the laser will be guided in the fiber. Guiding of a single laser mode has proven to be an effective means of avoiding thermo-optic distortions that have plagued high power bulk laser systems. This advantage comes at the price of the very high brightness pump diodes. While active double-clad fibers cannot as yet be fabricated from crystalline laser hosts, they can be readily fabricated from fused silica.41

Amplified spontaneous emission, nonlinear thresholds, and pump brightness create practical limits to the scaling of a high power fiber laser. Consider the following extreme conditions for a conventional step-index Ybsilica fiber. Under optimal conditions ASE and nonlinearity currently limit a high power fiber’s length to about 15 m. Compositional variations in the core and clad index of refraction currently limit the minimum numerical aperture of step index fibers to no less than 0.05. This means that single-mode laser operation at 1 µm restricts core diameters to about 20 µm. Further, the requirement to deplete the pump in under 15 m limits clad-to-core ratio, γ, to be no more than 20. Current limits on the refractive index of the outer cladding set the numerical aperture of the inner clad at about 0.2. As mention earlier, state-of-the-art laser diode arrays can achieve a brightness of 5MW/cm2-sr. When this pump beam is injected into the 400 µm 0.2 NA pump clad it will generate an intensity of 3MW/cm2. For this pump intensity Eq. (11) predicts a 1.1GW/cm2 laser intensity and an output power of 3.5 kW. This is the current theoretical limit of diode pumped fiber lasers and was in fact the level at which power scaling stalled for several years.

Further progress on fiber power scaling was achieved by using ytterbium fibers as the pump source for the final amplifier. This approach increases pump brightness well beyond that available from laser diodes. Current record powers from a double-clad Ybsilica fiber now stands at 17 kW with beam quality of 2 mm-mrad.21 Increased pump brightness permits lower clad-core ratios and allows the pump fibers to operate at more optimal wavelengths. In the demonstration described above, the pumping occurred at 1018 nm and laser amplification was at 1075 nm, which corresponds to the last column in Table 2. For these wavelengths the laser quantum defect is 5.3% and the stated optical efficiency of the system was a record 94%.

Many approaches to generating high power solid-state lasers have been investigated and only a few have made their mark. While the lack of performance does not always imply a poor concept, success settles the issue. A least three of the architectures described have exceeded the 10 kW level and clearly deserve further attention. Only the edge-pumped disk category is devoid of a winner and it has been argued that this failing may be fundamental. Looking forward, there is still much room for improvement. The cost of high brightness diode leaves much room for improvement. Low cost approaches for transferring laser diode brightness are needed, but have not yet emerged. A couple of technical areas where progress can be made are listed below.

Advanced Laser Materials

One common observation is that each of these systems has pushed the available laser materials to their limit. Disk and slab architectures would benefit greatly from higher thermal conductivity materials. Fiber lasers would benefit from reduced nonlinearity and optical damage. Looking ahead, an obvious path for advancement is to work these materials properties. Earlier mention was made of the recent work on high thermal conductivity sesquioxide ceramics. New classes of materials with superior thermal properties should be explored as hosts. For fiber lasers efforts to reduce stimulated Brillouin gain have already significantly increased nonlinear thresholds.42 New materials and fabrication techniques for single mode fiber laser should be examined. Despite their complexity, crystalline and photonic crystal fibers offer a clear path to higher powers.4345 Care should be taken to insure that candidate laser materials would, if successful, be compatible with current scaling techniques and commercial markets.

Obviously omitted from this review of high power lasers are the longer wavelengths, the so-called “eye-safe” systems. This is because the longer wavelength lasers are so far behind the 1 µm systems. The state-of-the-art in eye-safe lasers with good beam quality is well less than a kilowatt. Considerable work has gone into erbium materials for 1.5 µm systems, much less so into thulium 2.0 µm and holmium 2.1 µm. With quasi-three level operation similar to ytterbium, all three laser schemes are be suitable for bulk or fiber systems.

Cryogenic Operation

One way to improve laser material parameters is to cool the system to liquid nitrogen temperatures. Recent demonstrations have proven this approach in the ytterbium laser system.18 At liquid nitrogen temperatures ytterbium behaves like a four-level laser system, except without the high intensity self-termination. While cooling sharpens ytterbium’s spectral bands, it significantly enhances the thermo-optic properties of many hosts. The opportunity to make continued advancements with existing materials is compelling. This approach has the potential for more than an order of magnitude increase in power scaling, although cryogenic operation creates its own set of practical complications.

Radiation Balanced Operation

As mentioned earlier, the quantum defect of an ytterbium laser material does not constitute the minimum level of heat generation. Fluorescence power can add or subtract from laser heating. Under the right conditions these two processes will cancel and the laser medium will generate no heat. All power cycles through the medium in the form of light. This technique has been demonstrated in a room temperature 200 W YbYAG laser with high beam quality.19,46 Optical efficiency of 26% was achieved with a heating fraction of only 0.001 using the conditions listed in the second column of Table 2. This approach also offers the prospect of continued scaling with current materials. The penalty for radiation balance is the reduction in optical efficiency. The advantage of this technique is the simplicity of eliminating active cooling of the gain medium.

Solid-state lasers have made a lot of progress in fifty years. High average power lasers have made remarkable progress in the last decade. Advances in laser diodes, fibers, and crystals have all played an important part in this progress. For scaling beyond the kilowatt power level, practical considerations such as cost and efficiency become critical. State-of-the-art systems now push the boundary of current materials and laser technology. The two most mature solid-state laser materials are described in detail. Power levels above 10 kW have been demonstrated in several laser architectures. A first principle analysis suggests future power scaling efforts should concentrate on face-pumped disks or double-clad laser architectures. Several experimental approaches offer promise for another order of magnitude increase in the single beam average power.

This work was support by the Office of Naval Research.

Optik & Photonik, “Towards exawatt laser power and sub-attosecond pulses, an interview with Gérard Mourou, project coordinator of the extreme light infrastructure (ELI),” Opt. Photonik. 5, (4 ), 7 –10 (2010). 1863-1460 CrossRef
Kaminskii  A. A., Crystalline Lasers: Physical Processes and Operating Schemes. ,  CRC Press ,  Boca Rotan, FL  (1996).
Kück  S., “Laser-related spectroscopy of ion-doped crystals for tunable solid-state lasers,” Appl. Phys. B. 72, (5 ), 515 –562 (2001). 0946-2171 CrossRef
Koechner  W., Chapter 1 in Solid-State Laser Engineering. , Sixth Revised and Updated Edition,  Springer Science and Business Inc. ,  NewYork  (2006).
Fan  T. Y., “Aperture guiding in quasi-three-level lasers,” IEEE J. Quantum Electron.. 29, , 1457 –1459 (1993). 0018-9197 CrossRef
Comaskey  B. et al., “Characterization of the heat loading of Nd-doped YAG, YOS, YLF, and GGG excited at diode pumping wavelengths,” IEEE J. Quantum Electron.. 31, (7 ), 1261 –1264 (1995). 0018-9197 CrossRef
Petermann  K. et al., “Rare-earth-doped sesquioxides,” J. Lumin.. 87, , 973 –975 (2000). 0022-2313 CrossRef
Ikesue  A. et al., “Fabrication and optical properties of high-performance polycrystalline nd∶yag ceramics for solid-state lasers,” J. Amer. Ceramic Soc.. 78, (4 ), 1033 –1040 (1995). 0002-7820 CrossRef
Lu  J. et al., “Yb3+∶Sc2O3 ceramic laser,” Appl. Phys. Lett.. 83, (6 ), 1101 –1103 (2003). 0003-6951 CrossRef
Prasadla  N. S. et al., “Development of ceramic solid-state laser host materials,” Proc. SPIE. 7193, , 71931  (2009). 0277-786X CrossRef
Takaichi  K. et al., “Lu2O3∶Yb3+ ceramics a novel gain material for high-power solid-state lasers,” Phys. Stat. Sol. (A). 202, , R1 –R3 (2005). 0031-8965 CrossRef
Sanghera  J. et al., “Ceramic laser materials,” Materials. 5, (2 ), 258 –277 (2012). 1996-1944 CrossRef
Robinson  M., Asawa  C. K., “ Stimulated emission from Nd3+ and Yb3+ in noncubic sites neodymium- and ytterbium-doped CaF2,” J. Appl. Phys.. 38, (11 ), 4495 –4501 (1967). 0021-8979 CrossRef
Hanna  D. C. et al., “Continuous-wave oscillation of a monomode ytterbium-doped fibre laser,” Electron. Lett.. 24, (17 ), 1111 –1113 (1988). 0013-5194 CrossRef
Newell  T. C. et al., “Temperature effects on the emission properties of Yb-doped optical fibers,” Opt. Commun.. 273, (1 ), 256 –259 (2007). 0030-4018 CrossRef
Krupke  W. F., “Ytterbium solid-state lasers—the first decade,” IEEE J. Sel. Topics Quantum Electron.. 6, (6 ), 1287 –1296 (2000). 1077-260X CrossRef
Giesen  A. et al., “Scalable concept for diode-pumped high-power solid-state lasers,” Appl. Phys. B. 58, (5 ), 365 –372 (1994). 0946-2171 CrossRef
Fan  T. Y. et al., “Cryogenic Yb3+-doped solid-state lasers,” IEEE J. Sel. Topics Quantum Electron.. 13, (3 ), 448 –458 (2007). 1077-260X CrossRef
Bowman  S. R. et al., “Minimizing heat generation in solid-state lasers,” IEEE J. Quantum Electron.. 46, (7 ), 1076 –1085 (2010). 0018-9197 CrossRef
Paschotta  R. et al., “Ytterbium-doped fiber amplifiers,” IEEE J. Quantum Electron.. 33, (7 ), 1049 –1056 (1997). 0018-9197 CrossRef
Ferin  A. et al., “17 kW CW laser with 50 μm delivery,” presented at  The 6th International Symposium on High-Power Fiber Lasers and Their Applications ,  IPG Photonics Corporation ,  Oxford, MA  (2012).
Bowman  S. R., Mungan  C. E., “Selecting materials for radiation balanced lasers,” OSA Trends Opt. Photon.. 34, , 446 –453 (2000). 1094-5695 
Riseberg  L. A., Weber  M. J., Progress in Optics. , Wolf  E., Ed., p. 91 ,  Plenum Press ,  New York  (1976).
Chernoch  J. P., Martin  W. S., “Multiple internal reflection face-pumped laser,” U. S. Patent No. 3633126 (1972).
Injeyan  H., Goodno  R., High Power Laser Handbook. ,  McGraw-Hill Prof Med/Tech ,  New York  (2011).
Chernoch  J. P., Koenig  H. R., “Disk laser having pumping means in direct optical combination with the disk end faces,” U. S. Patent 3423693 (1969).
Abate  J. A. et al., “Active mirror: a large-aperture medium repetition rate Nd:glass amplifier,” Appl. Opt.. 20, (2 ), 351 –361 (1981). 0003-6935 CrossRef
Giesen  A., Speiser  J., “Fifteen years of work on thin-disk lasers: results and scaling laws,” IEEE J. Sel. Topics Quantum Electron.. 13, (3 ), 598 –609 (2007). 1077-260X CrossRef
Mandal  A., Klimek  D. E., “Textron’s J-HPSSL 100 kW ThinZag® laser program,” in  Lasers and Electro-Optics (CLEO) and Quantum Electronics and Laser Science Conference (QELS) , JThH2,  Optical Society of America ,  Washington, DC  (2010).
Kane  T. J. et al., “The slab geometry laser–part II: thermal effects in a finite slab,” IEEE J. Quantum Electron.. 21, (8 ), 1195 –1210 (1985). 0018-9197 CrossRef
Albrecht  G. F. et al., “Design and characterization of a high average power slab YAG laser,” IEEE J. Quantum Electron.. 22, (11 ), 2099 –2106 (1986). 0018-9197 CrossRef
Chen  Y.-Z. et al., “A 526 W diode-pumped Nd∶YAG ceramic slab laser,” Chin. Phys. Lett.. 28, (9 ), 094208  (2011). 0256-307X CrossRef
Akiyama  Y. et al., “Efficient 10 kW diode-pumped Nd∶YAG rod laser,” Proc. SPIE. 4831, , 96 –100 (2003). 0277-786X CrossRef
Dascalu  T., “Edge-pump high power microchip Yb∶Yag laser,” Rom. Rep. Phy.. 60, (4 ), 977 –994 (2008). 1221-1451 
Dashcasan  M. J., Barati  E., Aghbolaghi  R., “Designing of an efficient multi-aperture, edge pumped thin-disk laser,” Opt. Laser Technol.. 44, (4 ), 800 –805 (2012). 0030-3992 CrossRef
Bowman  S. R., O’Connor  S., Biswal  S., “High power ytterbium disk laser,” IEEE J. Quantum Electron.. 41, (12 ), 1510 –1517 (2005). 0018-9197 CrossRef
Snitzer  E., “Proposed fiber cavities for optical masers,” J. Appl. Phys.. 32, (1 ), 36 –39 (1961). 0021-8979 CrossRef
Redmond  S. et al., “15 kW near-diffraction-limited single-frequency Nd∶YAG laser,” in  Technical Digest of the Conference on Lasers and Electro-Optics ,  Optical Society of America ,  CTUHH5  (2007).
Kracht  D. et al., “Core-doped ceramic Nd∶YAG laser,” Opt. Express. 14, (7 ), 2690 –2694 (2006). 1094-4087 CrossRef
Snitzer  E., “Cylindrical dielectric waveguide modes,” JOSA. 51, (5 ), 491 –498 (1961). 0030-3941 CrossRef(1993).
Simpson  J. R., Rare Earth Doped Fiber Lasers and Amplifiers. , Digonnet  M. J. F, Ed.,  Marcel Dekker Inc. ,  New York  (1993).
Li  M.-J. et al., “Fiber designs for higher power lasers,” Proc. SPIE. 6469, , 64690H  (2007). 0277-786X CrossRef
Limpert  J. et al., “High-power air-clad large-mode-area photonic crystal fiber laser,” Opt. Express. 11, (7 ), 818 –823 (2003). 1094-4087 CrossRef
Jeong  Y. et al., “Ytterbium-doped large-core fiber laser with 1.36 kW continuous-wave output power,” Opt. Express. 12, (25 ), 6088 –6092 (2004). 1094-4087 CrossRef
Lai  C.-C. et al., “Cr4+∶YAG double-clad crystal fiber laser,” Opt. Lett.. 33, (24 ), 2919 –2921 (2008). 0146-9592 CrossRef
Bowman  S. R., “Lasers without internal heat generation,” IEEE J. Quantum Electron.. 35, (1 ), 115 –122 (1999). 0018-9197 CrossRef

Grahic Jump LocationImage not available.

Steven R. Bowman received his degree in physics at the University of Maryland in College Park in 1986 and joined the Naval Research Labs in 1987. He is currently head of the Advanced Lasers Concepts Section. His areas of current research include high power lasers, new materials for solid-state laser, and nonlinear frequency conversion. He has authored 197 papers and holds 13 patents in the field of laser and electro-optics.

© 2013 Society of Photo-Optical Instrumentation Engineers

Citation

Steven R. Bowman
"High-power diode-pumped solid-state lasers", Opt. Eng. 52(2), 021012 (Oct 10, 2012). ; http://dx.doi.org/10.1117/1.OE.52.2.021012


Figures

Graphic Jump LocationF2 :

Energy structure and important transitions in Ytterbium doped silica.15

Graphic Jump LocationF1 :

The lower lying energy levels of Nd:YAG are plotted in wavenumbers. Energy manifolds are labeled according to their 4f electronic quantum numbers, S, L, and J using the notation L2J+12S+1. Manifolds are also numbered according to the four-level laser scheme with the upper laser level in red and lower in green. Pump and laser transitions are illustrated with arrows and labeled with the wavelength of the peak cross section.2

Graphic Jump LocationF3 :

Generalized high power laser architectures with arrows indicating power flow of the pump, fluorescence, laser, and heat.

Tables

Table Grahic Jump Location
Table 2Room temperature ytterbium laser material parameters as defined in the text.
Table Grahic Jump Location
Table 1Neodymium laser material properties at 293 K parameters as defined in the text. Peak values of spectral parameters are listed followed by the polarization direction.

References

Optik & Photonik, “Towards exawatt laser power and sub-attosecond pulses, an interview with Gérard Mourou, project coordinator of the extreme light infrastructure (ELI),” Opt. Photonik. 5, (4 ), 7 –10 (2010). 1863-1460 CrossRef
Kaminskii  A. A., Crystalline Lasers: Physical Processes and Operating Schemes. ,  CRC Press ,  Boca Rotan, FL  (1996).
Kück  S., “Laser-related spectroscopy of ion-doped crystals for tunable solid-state lasers,” Appl. Phys. B. 72, (5 ), 515 –562 (2001). 0946-2171 CrossRef
Koechner  W., Chapter 1 in Solid-State Laser Engineering. , Sixth Revised and Updated Edition,  Springer Science and Business Inc. ,  NewYork  (2006).
Fan  T. Y., “Aperture guiding in quasi-three-level lasers,” IEEE J. Quantum Electron.. 29, , 1457 –1459 (1993). 0018-9197 CrossRef
Comaskey  B. et al., “Characterization of the heat loading of Nd-doped YAG, YOS, YLF, and GGG excited at diode pumping wavelengths,” IEEE J. Quantum Electron.. 31, (7 ), 1261 –1264 (1995). 0018-9197 CrossRef
Petermann  K. et al., “Rare-earth-doped sesquioxides,” J. Lumin.. 87, , 973 –975 (2000). 0022-2313 CrossRef
Ikesue  A. et al., “Fabrication and optical properties of high-performance polycrystalline nd∶yag ceramics for solid-state lasers,” J. Amer. Ceramic Soc.. 78, (4 ), 1033 –1040 (1995). 0002-7820 CrossRef
Lu  J. et al., “Yb3+∶Sc2O3 ceramic laser,” Appl. Phys. Lett.. 83, (6 ), 1101 –1103 (2003). 0003-6951 CrossRef
Prasadla  N. S. et al., “Development of ceramic solid-state laser host materials,” Proc. SPIE. 7193, , 71931  (2009). 0277-786X CrossRef
Takaichi  K. et al., “Lu2O3∶Yb3+ ceramics a novel gain material for high-power solid-state lasers,” Phys. Stat. Sol. (A). 202, , R1 –R3 (2005). 0031-8965 CrossRef
Sanghera  J. et al., “Ceramic laser materials,” Materials. 5, (2 ), 258 –277 (2012). 1996-1944 CrossRef
Robinson  M., Asawa  C. K., “ Stimulated emission from Nd3+ and Yb3+ in noncubic sites neodymium- and ytterbium-doped CaF2,” J. Appl. Phys.. 38, (11 ), 4495 –4501 (1967). 0021-8979 CrossRef
Hanna  D. C. et al., “Continuous-wave oscillation of a monomode ytterbium-doped fibre laser,” Electron. Lett.. 24, (17 ), 1111 –1113 (1988). 0013-5194 CrossRef
Newell  T. C. et al., “Temperature effects on the emission properties of Yb-doped optical fibers,” Opt. Commun.. 273, (1 ), 256 –259 (2007). 0030-4018 CrossRef
Krupke  W. F., “Ytterbium solid-state lasers—the first decade,” IEEE J. Sel. Topics Quantum Electron.. 6, (6 ), 1287 –1296 (2000). 1077-260X CrossRef
Giesen  A. et al., “Scalable concept for diode-pumped high-power solid-state lasers,” Appl. Phys. B. 58, (5 ), 365 –372 (1994). 0946-2171 CrossRef
Fan  T. Y. et al., “Cryogenic Yb3+-doped solid-state lasers,” IEEE J. Sel. Topics Quantum Electron.. 13, (3 ), 448 –458 (2007). 1077-260X CrossRef
Bowman  S. R. et al., “Minimizing heat generation in solid-state lasers,” IEEE J. Quantum Electron.. 46, (7 ), 1076 –1085 (2010). 0018-9197 CrossRef
Paschotta  R. et al., “Ytterbium-doped fiber amplifiers,” IEEE J. Quantum Electron.. 33, (7 ), 1049 –1056 (1997). 0018-9197 CrossRef
Ferin  A. et al., “17 kW CW laser with 50 μm delivery,” presented at  The 6th International Symposium on High-Power Fiber Lasers and Their Applications ,  IPG Photonics Corporation ,  Oxford, MA  (2012).
Bowman  S. R., Mungan  C. E., “Selecting materials for radiation balanced lasers,” OSA Trends Opt. Photon.. 34, , 446 –453 (2000). 1094-5695 
Riseberg  L. A., Weber  M. J., Progress in Optics. , Wolf  E., Ed., p. 91 ,  Plenum Press ,  New York  (1976).
Chernoch  J. P., Martin  W. S., “Multiple internal reflection face-pumped laser,” U. S. Patent No. 3633126 (1972).
Injeyan  H., Goodno  R., High Power Laser Handbook. ,  McGraw-Hill Prof Med/Tech ,  New York  (2011).
Chernoch  J. P., Koenig  H. R., “Disk laser having pumping means in direct optical combination with the disk end faces,” U. S. Patent 3423693 (1969).
Abate  J. A. et al., “Active mirror: a large-aperture medium repetition rate Nd:glass amplifier,” Appl. Opt.. 20, (2 ), 351 –361 (1981). 0003-6935 CrossRef
Giesen  A., Speiser  J., “Fifteen years of work on thin-disk lasers: results and scaling laws,” IEEE J. Sel. Topics Quantum Electron.. 13, (3 ), 598 –609 (2007). 1077-260X CrossRef
Mandal  A., Klimek  D. E., “Textron’s J-HPSSL 100 kW ThinZag® laser program,” in  Lasers and Electro-Optics (CLEO) and Quantum Electronics and Laser Science Conference (QELS) , JThH2,  Optical Society of America ,  Washington, DC  (2010).
Kane  T. J. et al., “The slab geometry laser–part II: thermal effects in a finite slab,” IEEE J. Quantum Electron.. 21, (8 ), 1195 –1210 (1985). 0018-9197 CrossRef
Albrecht  G. F. et al., “Design and characterization of a high average power slab YAG laser,” IEEE J. Quantum Electron.. 22, (11 ), 2099 –2106 (1986). 0018-9197 CrossRef
Chen  Y.-Z. et al., “A 526 W diode-pumped Nd∶YAG ceramic slab laser,” Chin. Phys. Lett.. 28, (9 ), 094208  (2011). 0256-307X CrossRef
Akiyama  Y. et al., “Efficient 10 kW diode-pumped Nd∶YAG rod laser,” Proc. SPIE. 4831, , 96 –100 (2003). 0277-786X CrossRef
Dascalu  T., “Edge-pump high power microchip Yb∶Yag laser,” Rom. Rep. Phy.. 60, (4 ), 977 –994 (2008). 1221-1451 
Dashcasan  M. J., Barati  E., Aghbolaghi  R., “Designing of an efficient multi-aperture, edge pumped thin-disk laser,” Opt. Laser Technol.. 44, (4 ), 800 –805 (2012). 0030-3992 CrossRef
Bowman  S. R., O’Connor  S., Biswal  S., “High power ytterbium disk laser,” IEEE J. Quantum Electron.. 41, (12 ), 1510 –1517 (2005). 0018-9197 CrossRef
Snitzer  E., “Proposed fiber cavities for optical masers,” J. Appl. Phys.. 32, (1 ), 36 –39 (1961). 0021-8979 CrossRef
Redmond  S. et al., “15 kW near-diffraction-limited single-frequency Nd∶YAG laser,” in  Technical Digest of the Conference on Lasers and Electro-Optics ,  Optical Society of America ,  CTUHH5  (2007).
Kracht  D. et al., “Core-doped ceramic Nd∶YAG laser,” Opt. Express. 14, (7 ), 2690 –2694 (2006). 1094-4087 CrossRef
Snitzer  E., “Cylindrical dielectric waveguide modes,” JOSA. 51, (5 ), 491 –498 (1961). 0030-3941 CrossRef(1993).
Simpson  J. R., Rare Earth Doped Fiber Lasers and Amplifiers. , Digonnet  M. J. F, Ed.,  Marcel Dekker Inc. ,  New York  (1993).
Li  M.-J. et al., “Fiber designs for higher power lasers,” Proc. SPIE. 6469, , 64690H  (2007). 0277-786X CrossRef
Limpert  J. et al., “High-power air-clad large-mode-area photonic crystal fiber laser,” Opt. Express. 11, (7 ), 818 –823 (2003). 1094-4087 CrossRef
Jeong  Y. et al., “Ytterbium-doped large-core fiber laser with 1.36 kW continuous-wave output power,” Opt. Express. 12, (25 ), 6088 –6092 (2004). 1094-4087 CrossRef
Lai  C.-C. et al., “Cr4+∶YAG double-clad crystal fiber laser,” Opt. Lett.. 33, (24 ), 2919 –2921 (2008). 0146-9592 CrossRef
Bowman  S. R., “Lasers without internal heat generation,” IEEE J. Quantum Electron.. 35, (1 ), 115 –122 (1999). 0018-9197 CrossRef

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

PubMed Articles
Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.