Patent Application
20110216792
1. Field of the Invention
The present embodiments relate generally to laser systems and more particularly to wavelength beam combining systems and methods.
2. Description of the Prior Art
Wavelength beam combining (WBC) is a method for scaling the output power and brightness from diode bars and stacks.
WBC methods have been developed to combine beams along the slow dimension of each emitter as well as the fast dimension of each emitter. See for example U.S. Pat. Nos. 6,192,062, 6,208,679 and 2010/0110556 A1. However, the traditional methods described therein do not allow for greater flexibility in scaling the overall footprint of the system and flexibility in addressing large aperture optic concerns while scaling the output power and brightness to produce kilowatts, tens and hundreds of kilowatts, and even megawatts of power. Improved methods and systems to increased spectral brightness and output power methods to meet various industrial, scientific and defense applications are needed.
The following application seeks to solve the problems stated.
Lasers have numerous industrial, scientific, and defense applications. Industrial applications include metal cutting, spot welding, seam welding, drilling, fine cutting, and marking. Scientific applications include laser-guide stars for astronomy, gravitational wave detection, laser cooling and trapping, and laser-based particle accelerator. Defense applications include the laser-based weapon, the laser-induced spark, and LIDAR.
Example lasers that are applicable as described above include high average and high peak power fiber lasers and amplifiers, high average and peak power eye-safe Erbium-doped (Er-doped) fiber lasers and amplifiers, quasi-continuous wave (QCW) or pulsed or long-pulsed operation of industrial lasers, short pulsed (pulse widths of a few ns to hundreds of ns) operation of industrial lasers, and the like.
When wavelength beam combining is applied to any of the lasers described herein, including the lasers that are applicable as described above many of the relevant factors that impact laser utilization can be substantially improved. Aspects such as power output can be significantly increased, brightness can be substantially improved, cost can be dramatically reduced, thermal and fiber-related optics challenges can be readily overcome or mitigated to the point of insignificance, overall size can be reduced, and the like. Results of these and other factors may be improved by two or more orders of magnitude with wavelength beam combining.
A laser, such as a laser described in the following general description may be used in association with embodiments of the innovations described herein.
Lasers may generally be defined as devices that generate visible or invisible light through stimulated emission of light. “Laser” originally was an acronym for “Light Amplification by Stimulated Emission of Radiation”, coined in 1957 by the laser pioneer Gordon Gould, but is generally now mostly used for devices that produce light using the laser principle.
Lasers generally have properties that make them useful in a variety of applications. Laser properties may include: emitting light as a laser beam which can propagate over long lengths without much divergence and can be focused to very small spots; a very narrow bandwidth as compared to most other light sources which produce a very broad spectrum; light can be emitted continuously, or in short bursts (pulses) that may be as short as a few femto-seconds.
Lasers may come in a variety of types. Common laser types include semiconductor lasers, solid-state lasers, fiber lasers, and gas lasers.
Semiconductor lasers (mostly laser diodes) may be electrically or optically pumped and generally efficiently generate very high output powers often at the expense of poor beam quality. Semiconductor lasers may produce low power with good spatial properties for application in CD and DVD players. Yet other semiconductor lasers may be suitable for producing high pulse rate, low power pulses (e.g. for telecom applications). Special types of semiconductor lasers include quantum cascade lasers (for mid-infrared light) and surface-emitting semiconductor lasers (VCSELs and VECSELs), the latter also being suitable for pulse generation with high powers. Semiconductor lasers are further described elsewhere herein under the heading “LASER DIODE”
Solid-state lasers may be based on ion-doped crystals or glasses (e.g. doped insulator lasers) and may pumped with discharge lamps or laser diodes for generating high output power. Alternatively solid-state lasers may produce low power output with very high beam quality, spectral purity and/or stability (e.g. for measurement purposes). Some solid-state lasers may produce ultra short pulses with picosecond or femtosecond durations. Common gain media for use with solid state lasers include: Nd:YAG, Nd:YVO4, Nd:YLF, Nd:glass, Yb:YAG, Yb:glass, Ti:sapphire, Cr:YAG and Cr:LiSAF.
Fiber lasers may be based on optical glass fibers which are doped with some laser-active ions in the fiber core. Fiber lasers can achieve extremely high output powers (up to kilowatts) with high beam quality have limited wavelength-tuning operation. Narrow line width operation and the like may also be supported by fiber lasers.
Gas lasers may include helium-neon lasers, CO2 lasers, argon ion lasers, and the like may be based on gases which are typically excited with electrical discharges. Frequently used gases include CO2, argon, krypton, and gas mixtures such as helium-neon. In addition, excimer lasers may be based on any of ArF, KrF, XeF, and F2. Other less common laser types include: chemical and nuclear pumped lasers, free electron lasers, and X-ray lasers.
A laser diode, such as a laser diode described in the following general description may be used in association with embodiments of the innovations described herein and in the exhibits referenced herein.
A laser diode is generally based around a simple diode structure that supports the emission of photons (light). However, to improve efficiency, power, beam quality, brightness, tunability, and the like, this simple structure is generally modified to provide a variety of many practical types of laser diodes. Laser diode types include small edge-emitting varieties that generate from a few milliwatts up to roughly half a watt of output power in a beam with high beam quality. Structural types of diode lasers include double hetero-structure lasers that include a layer of low bandgap material sandwiched between two high bandgap layers; quantum well lasers that include a very thin middle layer (quantum well layer) resulting in high efficiency and quantization of the laser's energy; multiple quantum well lasers that include more than one quantum well layer improve gain characteristics; quantum wire or quantum sea (dots) lasers replace the middle layer with a wire or dots that produce higher efficiency quantum well lasers; quantum cascade lasers that enable laser action at relatively long wavelengths which can be tuned by altering the thickness of the quantum layer; separate confinement heterostructure lasers, which are the most common commercial laser diode and include another two layers above and below the quantum well layer to efficiently confine the light produced; distributed feedback lasers, which are commonly used in demanding optical communication applications and include an integrated diffraction grating that facilitates generating a stable wavelength set during manufacturing by reflecting a single wavelength back to the gain region; vertical-cavity surface-emitting laser (VCSEL), which have a different structure that other laser diodes in that light is emitted from its surface rather than from its edge; vertical-external-cavity surface-emitting-laser (VECSELs) and external-cavity diode lasers, which are tunable lasers that use mainly double heterostructures diodes and include gratings or multiple-prism grating configurations. External-cavity diode lasers are often wavelength-tunable and exhibit a small emission line width. Laser diode types also include a variety of high power diode-based lasers including: broad area lasers that are characterized by multi-mode diodes with 1×100 um oblong output facets and generally have poor beam quality but generate a few watts of power; tapered lasers that are characterized by astigmatic mode diodes with 1×100 um tapered output facets that exhibit improved beam quality and brightness when compared to broad area lasers; ridge waveguide lasers that are characterized by elliptical mode diodes with 1×4 um oval output facets; and slab-coupled optical waveguide lasers (SCOWL) that are characterized by circular mode diodes with 4×4 um and larger output facets and can generate watt-level output in a diffraction-limited beam with nearly a circular profile. There are other types of diode lasers reported in addition to those described above.
Laser diode arrays, bars and/or stacks, such as those described in the following general description may be used in association with embodiments of the innovations described herein and in the exhibits referenced herein.
Laser diodes may be packaged individually or in groups, generally in one-dimensional rows/arrays (diode bar) or two dimensional arrays (diode-bar stack). A diode array stack is generally a vertical stack of diode bars. Laser diode bars or arrays generally achieve substantially higher power, and cost effectiveness than an equivalent single broad area diode. High-power diode bars generally contain an array of broad-area emitters, generating tens of watts with relatively poor beam quality and despite the higher power, the brightness is often lower than that of a broad area laser diode. High-power diode bars can be stacked to produce high-power stacked diode bars for generation of extremely high powers of hundreds or thousands of watts. Laser diode arrays can be configured to emit a beam into free space or into a fiber. Fiber-coupled diode-laser arrays can be conveniently used as a pumping source for fiber lasers and fiber amplifiers.
A diode-laser bar is a type of semiconductor laser containing a one-dimensional array of broad-area emitters or alternatively containing sub arrays containing 10-20 narrow stripe emitters. A broad-area diode bar typically contains 19-49 emitters, each being on the order of e.g. 1×100 μm wide. The beam quality along the 1-μm dimension or fast-axis is typically diffraction-limited. The beam quality along the 100-μm dimension or slow-axis or array dimension is typically many times diffraction-limited. Typically, a diode bar for commercial applications has a laser resonator length of the order of 1 to 4 mm, is about 10 mm wide and generates tens of watts of output power. Most diode bars operate in the wavelength region from 780 to 1070 nm, with the wavelengths of 808 nm (for pumping neodymium lasers) and 940 nm (for pumping Yb:YAG) being most prominent. The wavelength range of 915-976 nm is used for pumping erbium-doped or ytterbium-doped high-power fiber lasers and amplifiers.
A property of diode bars that are usually addressed is the output spatial beam profile. For most applications beam conditioning optics are needed. Significant efforts are therefore often required for conditioning the output of a diode bar or diode stack. Conditioning techniques include using aspherical lenses for collimating the beams while preserving the beam quality. Micro optic fast axis collimators are used to collimate the output beam along the fast-axis. Array of aspherical cylindrical lenses are often used for collimation of each laser element along the array or slow-axis. To achieve beams with approximately circular beam waist a special beam shaper for symmetrization of the beam quality of each diode bar or array can be applied. A degrading property of diode bars is the “smile”—a slight bend of the planar nature of the connected emitters. Smile errors can have detrimental effects on the ability to focus beams from diode bars. Another degrading property is collimation error of the slow and fast-axis. For example, a twisting of the fast-axis collimation lens results in an effective smile. This has detrimental effects on the ability to focus. In stack “pointing” error of each bar is the most dominant effect. Pointing error is a collimation error. This is the result of the array or bar that is offset from the fast-axis lens. An offset of 1 μm is the same as the whole array having a smile of 1 μm.
Diode bars and diode arrays overcome limitations of very broad single emitters, such as amplified spontaneous emission or parasitic lasing in the transverse direction or filament formation. Diode arrays can also be operated with a more stable mode profile, because each emitter produces its own beam. Techniques which exploit some degree of coherent coupling of neighbored emitters can result in better beam quality. Such techniques may be included in the fabrication of the diode bars while others may involve external cavities. Another benefit of diode arrays is that the array geometry makes diode bars and arrays very suitable for coherent or spectral beam combining to obtain a much higher beam quality.
In addition to raw bar or array offerings, diode arrays are available in fiber-coupled form because this often makes it much easier to utilize each emitter's output and to mount the diode bars so that cooling of the diodes occurs some distance from the place where the light is used. Usually, the light is coupled into a single multimode fiber, using either a simple fast-axis collimator and no beam conditioning in the slow-axis direction, or a more complex beam shaper to preserve the brightness better. It is also possible to launch the beamlets from the emitters into a fiber bundle (with one fiber per emitter).
Emission bandwidth of a diode bar or diode array is an important consideration for some applications. Optical feedback (e.g. from volume Bragg grating) can significantly improve wavelength tolerance and emission bandwidth. In addition, bandwidth and exact center wavelength can also be important for spectral beam combining.
A diode stack is simply an arrangement of multiple diode bars that can deliver very high output power. Also called diode laser stack, multi-bar module, or two-dimensional laser array, the most common diode stack arrangement is that of a vertical stack which is effectively a two-dimensional array of edge emitters. Such a stack can be fabricated by attaching diode bars to thin heat sinks and stacking these assemblies so as to obtain a periodic array of diode bars and heat sinks. There are also horizontal diode stacks, and two-dimensional stacks.
For the high beam quality, the diode bars generally should be as close to each other as possible. On the other hand, efficient cooling requires some minimum thickness of the heat sinks mounted between the bars. This tradeoff of diode bar spacing results in beam quality of a diode stack in the vertical direction (and subsequently its brightness) is much lower than that of a single diode bar. There are, however, several techniques for significantly mitigating this problem, e.g. by spatial interleaving of the outputs of different diode stacks, by polarization coupling, or by wavelength multiplexing. Various types of high-power beam shapers and related devices have been developed for such purposes. Diode stacks can provide extremely high output powers (e.g. hundreds or thousands of watts).
There are also horizontal diode stacks, where the diode bars are arranged side-by-side, leading to a long linear array of emitters. Such an arrangement is more easily cooled due to the naturally convective cooling that occurs between the vertically oriented diode bars, and may thus also allow for a higher output power per emitter. Generally, the number of diode bars in a horizontal stack (and thus the total output power) is more limited than in a vertical stack.
Diode bars and diode stacks can achieve very high power without significant cooling challenges by applying quasi-continuous-wave operation that includes generate pulses of a few hundred microseconds duration and a pulse repetition rate of some tens of hertz.
Technologies and embodiments of wavelength beam combining, such as those described in the following general description may be used in association with embodiments of the innovations described herein and in the exhibits referenced herein.
As the light emitted by a laser diode is linearly polarized, it is possible to combine the outputs of two diodes with a polarizing beam splitter, so that a beam with twice the power of a single diode but the same beam quality can be obtained (this is often referred to as polarization multiplexing). Alternatively, it is possible to spectrally combine the beams of laser diodes with slightly different wavelengths using dichroic mirrors. More systematic approaches of beam combining allow combining a larger numbers of emitters with a good output beam quality.
Beam combining is generally used for power scaling of laser sources by combining the outputs of multiple devices. The principle of beam combining can essentially be described as combining the outputs of multiple laser sources, often in the form of a laser array to obtain a single output beam. The application of a scalable beam-combining technology can produce a power-scalable laser source, even if the single lasers contributing to the combined beam are not scalable. Beam combining generally targets multiplying output power while preserving beam quality so that the brightness is increased (nearly) as much as the output power.
While there may be many different approaches for beam combining with increased brightness, all can be grouped into one of three categories: coherent, polarization, and wavelength beam combining. Coherent beam combining works with beams which are mutually coherent. In a simple example monochromatic beams with the same optical frequency can be combined. However, some schemes of coherent beam combination are much more sophisticated and therefore work with emissions occurring over multiple frequencies, with the emission spectra of all emitters being the same.
Polarization beam combining combines two linearly polarized beams with a polarizer (e.g., a thin-film polarizer). Of course, this method is not repeatable, since it generates an unpolarized output. Therefore, the method does not allow power scaling in a strict sense. Each of these three techniques can be applied to various laser sources, e.g., based on laser diodes (particularly diode bars) and fiber amplifiers, but also to high-power solid-state bulk lasers and VECSELs.
Wavelength beam combining (also called spectral beam combining or incoherent beam combining) does not require mutual coherence because it employs emitters with non-overlapping optical spectra whose beams are fed into a wavelength-sensitive beam combiner, such as a prism, a diffraction grating, a dichroic mirror, a volume Bragg grating, and the like to produce a wavelength combined beam. WBC inventions and implementations are described in detail in U.S. Pat. No. 6,192,062, U.S. Pat. No. 6,208,679 and US 2010/0110556 A1. Wavelength beam combining successfully achieves easier beam combining without any significant loss of beam quality. Wavelength beam combining is also more reliable than a single high power laser diode because the failure of one emitter simply reduces the output power accordingly.
The general principle of wavelength beam combining is to generate several laser diode beams with non-overlapping optical spectra and combine them at a wavelength-sensitive beam combiner so that subsequently all of the beams propagate in the same direction.
To combine many diode lasers and achieve good beam quality, laser diodes that are combined must each have an emission bandwidth which is only a small fraction of the gain bandwidth. Beam quality during wavelength beam combining is further affected by the angular dispersion of the beam combiner. Beam combiners with sufficiently strong dispersion and wavelength stable laser diodes go a long way toward achieving good beam quality during wavelength beam combining. Techniques for tuning laser diode wavelengths to facilitate wavelength beam combining, range from independently tuning each laser to a predetermined wavelength, to automatically adjusting each laser diode beam wavelength based on its spatial position relative to the combined beam path.
Wavelength beam combining may be used for power scaling. While a simple example of nearly unlimited power scaling would be to tile collimated beams from a large number of independently running adjacent lasers, even though the combined power increases in proportion to the number of lasers, the beam quality of the combined output decreases while the brightness will be at best only equivalent to a single laser. Typically the brightness of the system is much lower than a single element. Therefore one can see that power scaling methods which conserve the beam quality of the beam combining elements are highly desirable.
Wavelength beam combining may be applied to various types of laser diode configurations including diode bars, diode stacks, and the like. A diode bar is a one-dimensional array of broad area laser emitters that can be combined with various fiber and optical systems to produce one or more wavelength combined beams. Diode bars may include two to fifty or more laser emitters on one linear substrate. Diode stacks are essentially a two dimensional array of diode. Diode bars can be fabricated into diode stacks in vertical stacking or horizontal stacking arrangements.
The systems and methods described herein address scaling WBC methods to produce high brightness and power. Some of the embodiments described herein use a plurality of modular laser input devices with each device being comprised of a plurality of laser elements to form a scalable system. The scalability of using modular laser input devices allows for flexibility in adapting to higher or lower powers as needed, reducing the size of optics required and in some instances reducing the overall footprint of the system, which in turn creates a compact and robust system. This system is scalable to kilowatts, tens and hundreds of kilowatts, and even megawatts of power output and brightness.
Aspects and embodiments relate generally to the field of scaling laser sources to high-power and high-brightness using an external cavity and, more particularly, to methods and apparatus for external-cavity beam combining using both one-dimensional or two-dimensional laser sources. Aspects and embodiments further relate to high-power and/or high-brightness multi-wavelength external-cavity lasers that generate an overlapping or coaxial beam from ten to hundreds and even megawatts of output power.
In particular, aspects and embodiments are directed to a method and apparatus for combining individual laser emitters into modular units where a plurality of these modular units are all combined in a single system producing a single output profile that has been scaled in brightness and power. One advantage of the invention provided herein is a reduction in the size of optical elements required for scaling systems. Another advantage is that the overall footprint of the wavelength beam combining (WBC) system may be reduced in size.
Often times ‘simplifying’ an optical system is thought of in terms of reducing the number of optical elements present in a particular system. Increasing the number of optical elements seems to increase the complexity of the system or manufacturability of the system. However, some of the embodiments described herein increase the number of optical elements in a WBC system to achieve some of the advantages previously discussed, such as reducing the aperture of certain optical elements.
For example, a basic WBC architecture is illustrated in
Similarly,
Another WBC method shown in
In the preceding illustrations of
This becomes problematic if one of the design goals of a WBC system is to produce a compact system that combines a large number of beams or the individual beams are spatially spread out. The increased number of beams or spatial spreading between beams creates a larger one-dimensional or two-dimensional beam profile. The transform optics thus need to have a sufficiently large aperture.
When the aperture of a transform optic becomes too large, manufacturing becomes more difficult and the optics cost generally increases. Making large aperture optics with low aberrations is a difficult task. Commercially-off-the-shelve optics with acceptable quality is limited to about 5 to 6 inches in diameter, with 1 inch being the most common.
Laser sources based on common “off-the-shelf” high power laser diode arrays and stacks are based on broad-area diode laser elements. Typically, the beam quality of these elements is diffraction-limited along the fast axis and many times diffraction-limited along the slow axis of the laser elements. It is to be appreciated that although the following discussion may refer primarily to laser diodes, diode bars and diode stacks, embodiments of the invention are not limited to laser diodes and may be used with many different types of laser emitters, including fiber lasers, individually packaged diode lasers, semiconductor lasers and other types of lasers. Furthermore, as used herein, the term “array” refers to one or more laser elements placed side by side. The array dimension in
As noted, Wavelength beam combining (WBC) of laser elements is an attractive method for scaling power and brightness from a laser system. Brightness is product of N*P/(λ2*M2x*M2y)i, where N is the total number of combining elements, P is the output power of each element, and λ is the operating wavelength, M2x, M2y are the beam qualities along the two dimensions whereas N*P is the power.
Generally, all three WBC cavities consist of an array or stack of laser elements, a transform optics (cylindrical or spherical lens or mirror), a dispersive element (shown using a diffraction grating), and a partially reflecting output coupler. The transform optic(s) is placed after the laser array. The position of the transform optics depends on the source. In an ideal point source it is placed a focal length away from the source. The dispersive element is placed at where all the beams are spatially overlapped, nominally at the back focal plane of the transform optics. If the dispersive element is not placed at the nominal position then results in degradation in output beam quality. The output coupler is placed on the path of the first-order diffracted beams. As such, ideally, all output beams from the laser elements are spatially overlapped at the grating by the transform optics as shown in
For scaling to higher power and higher brightness, this basic optical setup for the three WBC cavities discussed is limiting and at times impractical. To illustrate an example using the cavity shown in
Scaling the previous illustration from a 3000-watt to a 10 kilowatt system would now require the size of the optic to be 3.3 times larger, or about 191.4 mm (˜10 inches). Procuring a 10-inch diameter optics with low aberrations is very expensive and requires time-consuming custom optical fabrication. It makes a 10 kilowatt system not competitive in the market place. Additionally, assuming the spectral bandwidth of the 3000 W and 10 kW is the same, then the focal length of a 10-inch is also going to increase the distance between the transform optic/mirror to the diffraction grating by roughly a factor of 3.
One embodiment addressing the issue of using optical elements with larger apertures is illustrated in
Cavity 200a consists of a plurality of laser elements 250 producing a one or two-dimensional profile, a transform optics for each set of laser elements 250, a dispersive element (shown transmission diffraction grating), and an output coupler. The individual laser input modules 252 of Cavity 200a are comprised of a set of plurality of laser elements 250, which forms a one or two-dimensional profile and a single transform optics 208. Additionally, as illustrated in
Continuing with the previous illustration of using laser diode bars where we assume the largest optical elements to be tolerated is 3 inches, then laser elements 250 would be comprised of a stack consisting of 30 diode bars. Thus each laser input module 252 would produce 3000 watts of total power and external cavity 200a as illustrated in
Another implementation is shown in
The addition of the telescope also reduces the spot size on the grating. For example, a typical diode bar has a f=900 um collimation optic attached. Assuming ideal collimation the collimated beam has a divergence angle of about 1 mrad. Thus, using a f=100 mm transform optics the spot size on the grating is about 100 um along the WBC dimension. Such small spot size is not desired. The spot size on the grating is small and, thus, the power density on the grating would be very high. This results in increased risk of optical damage to the grating.
With the additional telescope the spot size on the grating can be increased by the telescope magnification. In the previous example using f=10 mm f=1 mm or 10× magnification the spot size on the grating is now 1 mm instead of 100 μm. At the same time the spectral bandwidth is reduced from 39 nm to 3.9 nm per bar. Thus, as stated earlier, 10 bars can be combined to use all of the 39 nm bandwidth. One example of as shown in
As a design example, we assume the lens array is a spherical lens is a 3-inch diameter optic with a focal length of 300 mm. The stack consists of 30 bars at 2 mm pitch. We assume that the common telescope optics is cylindrical optic has a focal length of 2.8 mm. We also assume that grating has a groove density of 1760 lines per mm. The transform optic has a focal length of 200 mm. Thus, the bandwidth of the laser stack is approximately 0.78 nm. For a ten 30-bar stacks the bandwidth is approximately 10×0.78/ff, where if is the stack-to-stack fill factor. For a stack-to-stack fill factor of 0.9, the bandwidth is approximately 8.7 nm. The beam size on the grating is approximately 100 mm. If a spatial inter-leaver is used with polarization multiplexing, up to 10 (the number of stacks)×30 (the number of bars per stack)×2 (the effect of the spatial inter-leaver)×2 (due to polarization)=1200 bars can be combined with an output beam quality of a single bar with bar power of 1200×100 W=120,000 W and spectral line width of approximately 8.7 nm.
Highly reliable diode laser systems are in great demand for industrial applications. Industrial customers typically demand that the system last up to 100,000 hours, or more than 11 years. The lifetime of lasers varies by the laser type. The lifetime of actively cooled diode lasers using micro-channel coolers is approximately 10,000 hours or longer. Passively cooled diode lasers have a lifetime of approximately 20,000 hours or longer. The lifetime of sealed-tube CO2 lasers is approximately 35,000 hours. There is little maintenance. Sealed-tube CO2 lasers are limited to a few hundred Watts. High-power, kW-class CO2 laser systems typically last up to 100,000 hours or longer. However, they require maintenance every 1,000 hours and complete optical rebuild every 8,000 hours. Lamp-pumped solid state lasers have approximately the same cycle time as high power CO2 lasers. The lifetime of diode-pumped solid lasers, including fiber and bulk solid state, is much shorter than CO2 lasers. At best the lifetime is limited by the lifetime of the diode lasers. In this disclosure we show that the lifetime of our high-brightness WBC diode lasers can have a lifetime of up to 50,000 hours, 100,000 hours or greater without any need for maintenance.
In most cases, the output power of diode lasers has a linear dependence on time and is approximately
P(t)≈P0−βt,
where P0 is the initial power, P(t) is the laser power at time t later, and β is the degradation rate. For passively cooled diode bars the degradation rate is approximately β˜1×10−5 W per hour. Therefore, for passively cooled diode bars the end of life occurs at approximately 20,000 hours, where the end of life is defined as the time at which the output power is 80% of the initial value. To illustrate the intrinsic advantages of our system let us consider a 1,000-W system. Let us assume the cost of the diode laser is $10/W. Thus, for the 1,000-W system the cost of the diode is $10,000. We assume that our system may be required to last 100,000 hours. If we build our system with 1,000 W of diode lasers in the system we have to have to replace the diode lasers every 20,000 hours. Thus, we have to replace 4,000 W worth of diode lasers. The cost of the replacement is therefore $40,000 plus labor. All diode-pumped solid state lasers, including bulk Nd:YAG, thin disk, and fiber lasers, have to follow this model. This is fundamental of the laser resonator for diode-pumped solid state lasers. For example, in bulk Nd:YAG lasers, once the diode reaches its end of life condition, it has to replaced. It has to be replaced and the entire laser must be realigned optically. Fundamentally, this requirement may not apply to WBC systems. In our system the output beam is invariant with respect to the number of laser bars, the number of elements, or the output power of the system. For example, if we design a 1,000-W system we can install a multiplicity of 1,000 W diode laser groups in our system and just turn on one group at a time. At the end of 20,000 hours simply turn off the first group of diode lasers and turn on the second group or combination. Thus, replacement and/or realignment to the system are unnecessary. This is a fundamental property of our system. Intrinsic to our system, even though the diode lasers only last 20,000 hours we can make a laser system that lasts 100,000 hours, or any length of time desired. It is possible to reduce the number of replacement diode lasers required from the straightforward example described above. For example, when we manufacture the diode laser system, we install a total of 2,000 W worth of diode laser power. The 2,000-W laser would consist of 5 clusters of diode lasers. The first cluster has a power level of 1,200 W. The rest of the clusters are at 200 W each. For the first 20,000 hours we run just the 1200-W cluster until its end of life at the power level of 1000 W. At the end of 20,000 hours we run clusters 1 and clusters 2. At the end of 40,000 hours we run clusters 1, 2, and 3. At the end of 60,000 hours we run clusters 1, 2, 3, and 4. At the end of 80,000 hours we run all of the clusters. The output power from the system at any given time is always approximately at 1,000 W. In this example the total cost of the diode lasers is $20,000 instead of the previous example's $50,000 ($10,000 for the initial diodes plus $40,000 for the replacement diodes). More importantly there is no scheduled maintenance needed. We never have to open the sealed laser system to replace and realign the system. As part of the system, the cluster operation is programmed and controlled by a computer with a simple power threshold detector. If the diode lasers turn out to be more reliable than expected from the mean lifetime, the computer may not need to turn on additional clusters until the system power is lower than a pre-set threshold, which may further extend the operation time. The above example shows just one possible sequence of operating various clusters. A multiplicity of possible system scenarios exists. Furthermore, it is possible to combine this approach with de-rating the power from each cluster of diode lasers, which results in further, possibly drastic, improvements on the baseline lifetime expected for each cluster.
The full utilization of diode bars and stacks in many applications, such as pumping of alkali lasers and spin-exchange optical pumping, is limited by the broad output spectrum. The output spectral bandwidth of diode bars and stacks is approximately 3 to 5 nm. For some applications output spectral bandwidth of less than 0.05 nm is required. Furthermore, the output spectrum is typically not wavelength-stabilized. Thus, the center wavelength changes as a function of operating temperature. In many applications this is not desirable and can lead to catastrophic damage of the laser system. The typical change of wavelength with temperature is about 0.33 nm per degree Celsius. In some applications, like pumping alkali lasers, a shift of about 0.05 nm will cause the laser system to stop lasing.
There are three general methods of line-narrowing and wavelength-stabilized diode arrays and stacks. First method uses an internal grating as part of the laser manufacturing process, for example, the distributed feedback (DFB) laser and distributed Bragg reflector (DBR) laser. Several companies offer internally wavelength stabilized diode arrays and stacks. A disadvantage of the first method is that diode laser performance typically suffers from the addition of an internal grating. The second method uses external volume Bragg gratings (VBG). There are various companies that offer VBGs. Both methods are time consuming and expensive. Furthermore, the resulting line width is about 0.5 nm with about 0.1 nm per degree Celsius wavelength-to-temperature coefficient. Typically, there is absorption in VBGs and thus sometime active cooling is required. The third method uses an external diffraction grating. The third method has the highest dispersion and, in principle, can result in very narrow line width.
In methods 2 and 3 the resulting line width is much broader than that which the system is capable of. The broadening of the output spectrum is mainly due to imperfections in the laser emitters. Some of these imperfections apparent in laser diode bars are shown in
One attempt at addressing the imperfections presented in
An analysis (using ZEMAX) comparing the effects of smile on system 400 was produced. The model contained a three-bar diode stack. The center wavelength of each emitter had a wavelength of 980 nm. Each bar had 10 μm of smile, peak-to-valley. Each bar is collimated by an f=1 mm fast-axis collimation optic. The spherical telescope consists of two f=100 mm lenses. The grating has a groove density of 600 lines per mm. The top elements with +5 μm of smile are forced to operate at approximately 972.16 nm, and elements with −5 μm of smile are forced to operate at 987.82 nm. The elements without any smile operate at 980 nm. Therefore, the spectrum of each bar and the entire stack is now approximately 15.66 nm. Typically, single laser elements have a spectrum bandwidth of 3-5 nm. For state-of-the-art diode arrays and stacks the approximate range of smile is about 3 micron, which under this model would result in a spectral bandwidth of about 5 nm. Thus, this type of system for spectral brightness purposes may not be any better than the free-running spectrum.
The cylindrical lens array 533 is placed at the sum of the focal plane of the cylindrical lens array and the focal length of the collimation optic 506. The diode stack 550 is also at the focal plane of the first optic 501c of the array-telescope. The output coupler is at the focal plane of the second optic of the array-telescope 510d. The separation is the sum of their focal lengths. Along the stack dimension, as shown in the
A ZEMAX model of this configuration was also analyzed. The parameters are the same as in the previous model with the inclusion of a cylindrical lens array 533. The cylindrical lens array has a focal length of 100 mm. All the elements operate at exactly the same wavelength. In principle, the output spectrum can be single frequency. This has many applications where a wavelength operating at MHz and kHz range requires high spectral brightness. This type of a system is also robust in that it becomes athermal or not susceptible to change in the lasing wavelength with increases or decreases in temperature.
Table 1 compares this optical cavity with competing techniques. The cavity setup in
The prior external-cavity 1-D wavelength beam combining (WBC) architecture, where WBC is performed along the array dimension, is shown in
An analysis modeling
Low smile and pointing errors of the collimated diode elements are two of the key characteristics that are highly desired for robust and efficient wavelength beam combining. Smile, or physical bending of the diode laser array during packaging, and pointing error caused by misalignment of collimating micro-lens degrade the output beam quality and reduce the beam combining efficiency. External-cavity operation is highly dependent on the amount of smile and collimation error. The prior WBC cavity as shown in
The main drawback of the prior WBC cavity (
A second optical model (using ZEMAX) based on
Modeling
Prior external-cavity 1-D WBC architecture of 2-D laser elements as related art is shown in
Optically modeling
The main drawback of this WBC cavity is that the output beam quality will degrade proportionally to the amount imperfections in diode bars and stack (smile and collimation errors). These errors can degrade the output beam quality by as much as a factor of 10. While diode bars and stacks with low smile and collimation errors are available (typically leading to a 2× to 3× degradation in output beam quality) they tend to be more expensive. In this application, an embodiment of a new WBC architecture is described where the output beam quality along the beam combining dimension is nearly diffraction-limited and independent of the amount of smile and collimation error.
The full utilization of diode bars and stacks in many applications, such as pumping of solid state lasers and direct use in material processing, is limited by the poor output beam. Furthermore, the output beam quality of diode arrays and stacks is much worse than what is actually possible, with respect to the beam quality of each individual emitter. The degradation is mainly due to packaging and collimation errors. These errors are shown in
Modeling a configuration of the concept and nature of
Table 1 compares our optical cavity with competitors. The only disadvantage in our setup is that it is larger. Our setup can fully compensate for smile, pointing, and twisting errors. Our setup is universal. The setup, in principle, works for all lasers. There is no need to measure the imperfections of the diode laser bars and stacks and fabricate a custom correction optic. Our approach works for perfect arrays and stacks to grossly imperfect arrays and stacks. Since our setup requires a diffraction grating, the output spectrum will be narrow line-width, tunable, and athermal. These characteristics are highly desirable in many applications.
The above description is merely illustrative. Having thus described several aspects of at least one embodiment of this invention including the preferred embodiments, it is to be appreciated that various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications, and improvements are intended to be part of this disclosure, and are intended to be within the spirit and scope of the invention. Accordingly, the foregoing description and drawings are by way of example only.
This application claims the benefit of the following provisional applications, each of which are hereby incorporated by reference in its entirety: U.S. Ser. No. 61/310,777 filed Mar. 5, 2010; U.S. Ser. No. 61/310,781 filed Mar. 5, 2010, and U.S. Ser. No. 61/417,394 filed Nov. 26, 2010.
| Number | Date | Country | |
|---|---|---|---|
| 61310777 | Mar 2010 | US | |
| 61310781 | Mar 2010 | US | |
| 61417394 | Nov 2010 | US |
