Solid state short pulse laser with pump pulse shaping

Abstract

High-quality laser pulses with durations from 1 to 1000 nanoseconds are provided according to several techniques that entail driving a laser diode bar or seed diode with a pulsed electrical current source. In a first example, pump light is coupled into a gain medium without substantial energy storage during the pump pulse, ensuring the signal pulse temporal characteristics are directly controlled by the pump pulse and electrical source. A second example employs a two-segment pumping scheme within a resonant cavity. The first segment inverts the gain medium below the lasing threshold, while the second segment drives the medium above threshold to initiate pulse buildup. Both techniques ensure that the final output pulse characteristics are governed by the initial pulsed electrical current, providing precise temporal control over the generated laser pulses

Description

SOLID STATE SHORT PULSE LASER WITH PUMP PULSE SHAPINGRELATED APPLICATIONS

[0001] This application claims priority benefit of U.S. Provisional Patent Application No. 63 / 746,554 filed January 17, 2025 and U.S. Provisional Patent Application No. 63 / 758,683 filed February 14, 2025. Each application is hereby incorporated by reference in its entirety.TECHNICAL FIELD

[0002] The field of the present disclosure relates to laser devices, particularly a method for creating short laser pulses. Specifically, this disclosure describes an approach that generates the short laser pulse by modulating the pump pulse that directly controls the generation of the short laser pulse, thereby reducing the complexity compared to that of a traditional Q-Switched laser system.BACKGROUND INFORMATION

[0003] Short pulse laser systems with pulse durations between 1-1000 nanosecond and energies over 10 microjoules are an important tool for industrial applications. Due to their short pulse duration, the intensity is high enough to remove material through ionization rather than thermal interaction. This allows for delivery of targeted energy on a specific area without damage to the surrounding bulk material. Industrial applications include micro-machining, marking, and ablation. Medical applications include skin treatments, surgery, and hair removal. The short pulses coupled with high intensity also enable sensing applications like spectroscopy, range finding, or directed energy applications. Finally, these lasers can also be used as pump sources themselves for gain materials with short storage lifetimes.SUMMARY OF THE DISCLOSURE

[0004] Despite their many applications, the generation of short laser pulses below 1000 nanoseconds and especially below 100 nanoseconds is challenging and requires a complex system. The traditional approach for generating high energy, short laser pulses employs a Q-Switched laser design. Besides enabling a way of storing energy in the gain medium, many applications also require precise timing control when the laser pulse is emitted, and the Q-Switch facilitates this control 1503541\70012\FG: 104682020.1by enabling triggering with a timing signal. The design uses a laser storage medium such as a laser crystal to store optical energy of the pump light over a time scale much larger than the duration of the short laser pulse. The laser storage medium employs an atomic or molecular energy level structure to absorb the pump light and stores that energy in an upper laser level. The energy extraction that leads to lasing occurs as a stimulated transition from the upper to the lower laser level. This would happen immediately, but a Q-Switch element needs to hold off buildup of the signal light in the laser cavity during the long pump pulse and switches from blocking the light to transmitting the light at the end of the pump pulse to enable extraction of the short laser pulse. The Q-Switch changes the roundtrip loss in the cavity to be higher than the roundtrip gain for the holdoff and lower than the roundtrip gain for lasing. This Q-Switch element adds significant complexity in the system and is typically implemented using a high-voltage electro-optical switch or an RF driven acousto-optical switch. Both approaches require complex drivers to generate either a waveform with thousands of volts that can be switched on a timescale of a few nanoseconds or an RF signal with bandwidths of more than 10MHz.

[0005] An alternative approach is a passive Q-Switch that initially has a high loss and prevents lasing. With sufficient intensity in the cavity that is reached at some point during the pump pulse, the passive Q-Switch saturates and significantly reduces its roundtrip loss, enabling the buildup of the laser pulse. There is no active timing control that triggers the switch from high loss to low loss and the timing is impacted by noise fluctuations during the pump pulse buildup.Therefore, the main drawback is the temporal jitter of the emitted pulse that is typically between 10 and 1000ns.

[0006] Over the last decades, the performance and quality of semiconductor diode lasers has improved greatly. Specifically, the optical peak power that can be achieved directly out of a diode laser has continuously improved.

[0007] Furthermore, diode lasers in a semiconductor bar configuration can achieve instantaneous power levels of a few hundred watts in continuous wave operation, limited by thermal management. For pulsed diode lasers that operate at a lower duty cycle the peak power can be increased beyond the thermal constraint and the limitation will be primarily optical damage on the diode bar itself. Peak powers2503541\70012\FG: 104682020.1of a few kilowatts have been demonstrated out of a single laser diode bar without optical damage at pulse durations of tens of nanosecond. Combining multiple of these diode bars enables a total peak output power of 100 kilowatts. That peak power level is equivalent to that of a typical Q-Switched laser.

[0008] As an example, a typical Q-Switch laser has a pulse energy of 1 millijoule and a pulse duration of 10 nanoseconds. This translates into a peak power of 100 kilowatts, comparable to the peak power achievable with pulsed diode lasers.

[0009] Even though a total output power of 100 kilowatts can be realized with multiple diode bars, they cannot inherently create a single beam output with good beam quality that is equivalent to the laser beam intensity from a Q-Switched laser. This is due to the inherently poor beam quality from a laser diode bar, and even worse for multiple diode bars. Typical diode bars have beam quality parameters in the slow axis much larger than M2of 10, meaning compared to a diffraction limited beam their achievable focus spot size is more than 10 times larger. A Q-switch laser typically has an M2between less than 2 and can readily achieve diffraction limited focus spots. Enhancing the beam quality of a laser diode output to an M2value of less than 2 would allow for direct generation of short pulse, high quality laser beams without the added complexity of a Q- Switched laser system.

[0010] The present disclosure describes alternatives for creating these short pulses that significantly lower the complexity of a Q-Switched laser and reduces cost, size, weight, and volume. It specifically eliminates the need for an active Q- Switch optical element and electronic driver while maintaining the ability to temporally trigger the emission of the laser pulse, further reducing the cost. It also greatly improves the overall electrical to optical efficiency, significantly reducing the requirements for the cooling system.BRIEF DESCRIPTION OF THE DRAWING FIGURES

[0011] Fig. 1 is a block diagram of one of the laser systems described in this disclosure.

[0012] Fig. 2 is a set of graphs depicting the efficiency of a traditional Q-Switched laser system and one of the systems described in this disclosure.3503541\70012\FG: 104682020.1

[0013] Fig. 3 is a set of graphs showing the temporal evolution of the components in a traditional Q-Switched laser system and in one of the systems described in this disclosure.

[0014] Fig. 4 is a raytracing diagram and intensity profile of a single semiconductor laser diode bar in two configurations.

[0015] Fig. 5 is a side and top view of a ray tracing diagram of multiple, individually collimated laser diode bars converging into a small aperture.

[0016] Fig. 6 is a closeup side view of a ray tracing diagram of the focus area of multiple laser diode bars with a waveguide containing the rays.

[0017] Fig. 7 is a side view of a ray tracing diagram of multiple, individually collimated laser diode bars and a set of micro-optics.

[0018] Fig. 8 is a spectral graph showing different light scattering processes in matter.

[0019] Fig. 9 is a diagram showing the optical power against propagation length in a gain material that amplifies the input light.

[0020] Fig. 10 is a block diagram of an optical layout of a laser system according to one embodiment.

[0021] Fig. 11 is a block diagram of an optical layout of a laser system according to another embodiment.

[0022] Fig. 12 is a timing diagram showing the optical power over time for two embodiments of the present disclosure.

[0023] Fig. 13 is a spectral diagram showing the input diode light, the bandwidth of a Raman process, and the output signal light.

[0024] Fig. 14 is a block diagram of one of the laser systems described in this disclosure.

[0025] Fig. 15 is a timing diagram showing the timing signals, gain / loss, and power over time for one of the embodiments in the present disclosure.

[0026] Fig. 16 is a timing diagram showing the timing signals, gain / loss, and power over time for one of the embodiments in the present disclosure.

[0027] Fig. 17 is a timing diagram showing the pump and signal power and the inversion in the gain medium over time for one of the embodiments in the present disclosure.4503541\70012\FG: 104682020.1

[0028] Fig. 18 is a timing diagram showing the optical power for three pump parameters for one of the embodiments in the present disclosure.

[0029] Fig. 19 is a raytracing diagram and detector response showing the absorbed pump light in one of the embodiments in the present disclosure.

[0030] Fig. 20 shows a block diagram of a compact pump module arrangement of one of the embodiments in the present disclosure.

[0031] Fig. 21 shows an arrangement of the compact pump modules around a gain medium in a side-pumped configuration.

[0032] Fig. 22 shows an arrangement of the compact pump modules around a gain medium in an end-pumped configuration.DETAILED DESCRIPTION OF EMBODIMENTS

[0033] Short pulse lasers with pulse durations between 1 and 1000 nanoseconds are an important tool for industrial, medical, and sensing applications due to their high peak power that allows for non-thermal processing. Other applications include pump sources for gain materials with short storage lifetimes.

[0034] A laser generates coherent light through amplification of photons via stimulated emission in a gain medium. The stimulated emission creates a coherent beam that can be focused to a diffraction-limited spot to achieve the highest intensity on target. The gain medium provides a mechanism for a pump source such as a laser diode to deposit its energy and release it at a lower energy through stimulated emission. The energy can be stored for time periods of picoseconds up to tens of milliseconds. For traditional laser systems longer storage times are preferred since they reduce the peak power required for the pump since the same pump energy can be delivered over a longer period of time. There is an efficiency penalty for long pump durations since the gain medium constantly emits a portion of the stored light via spontaneous emission that will not contribute to the stimulated emission. If the storage time and the pulse duration of the laser output would be similar, this loss would be close to zero, but the peak power of the optical pump source would need to be similar to the peak power of the laser output as well.

[0035] Generating these short pulses traditionally requires a laser incorporating a Q- Switch element. This is imposed because the time scale of the pump light is typically on the order of the storage lifetime of the laser gain material and usually 5503541\70012\FG: 104682020.1between 10 microseconds and 10 milliseconds. The Q-Switch prevents pulse buildup during the duration of the pump pulse and only allows pulse build up after the gain material is sufficiently energized. The timescale of the Q-Switch pulse is determined by the roundtrip time of the cavity and speed of the temporal switching but typically achieves between 1 and 1000 nanoseconds in pulse duration. The disclosure describes alternative implementations that achieve short laser pulses by directly driving a semiconductor diode laser to facilitate the generation of the short pulses. This approach provides a much higher system efficiency and results in a more compact system with less components.

[0036] Q-Switches are implemented using active or passive approaches. An active approach employs an electro-optical or acousto-optical switching element including the drivers necessary to generate the corresponding waveforms. The advantage of this approach is the ability to temporally trigger the Q-Switch event and thus control the emission of the laser pulse with a jitter typically below 10ns. Passive Q-Switches employ a saturable absorber that initially has a high loss but switches to a low loss when a threshold intensity is reached. These devices are much simpler since they eliminate the complexity of an active Q-Switch and the associated drivers, but they exhibit a large temporal jitter, typically above 10ns. This is due to small fluctuations during the pump pulse buildup and the resulting temporal variation on when the lasing threshold is crossed.

[0037] This disclosure describes several implementations that use a temporally shaped electrical pulse to drive one or more semiconductor diodes generating a temporally shaped optical pump pulse.

[0038] In one embodiment the pump pulse is temporally substantially equivalent to the output pulse. A gain process in which optical gain is present substantially only during the pump light pulse and without substantial energy storage in the gain medium is utilized to improve the beam quality of the output beam. Examples of selected processes that have no substantial energy storage are described below. Those processes, in contrast to traditional laser gain media where the pump light pulse inverts the gain media and stores the energy over time periods much longer than the pump light pulse, return the energy immediately and thus only provide gain during the pump light pulse.6503541\70012\FG: 104682020.1

[0039] In another embodiment the optical pump pulse contains multiple temporal segments that provide energy to the gain medium with one of the temporal segments initiating the depletion in an active gain medium.

[0040] Fig. 1 illustrates the block diagram of the first approach presented in this disclosure. A main controller (101) provides timing and power signals (102) to the diode driver (103). The diode driver generates a pulsed electrical current (104) with pulse durations between 1 and 1000 nanoseconds, a selectable pulse repetition rate, and a selectable pulse format. The electrical current pulse is sent to the laser diodes (105) which generate optical pump pulses (106) of similar pulse duration. The optical pump pulses are directed towards the gain medium (107) that is configured to generate a signal laser pulse. A beam shaping optic (113) is configured to couple the pump pulses into the gain medium (107) where the pump pulse is converted into the signal laser pulse (114). The signal laser pulse is propagated through an optical telescope (108 and 109) to provide an output pulse with an M2beam quality of less than 3 in both directions (symmetric). Heat from the components (112) is cooled through one or multiple cooling elements in the system (111).

[0041] The signal laser pulse (114) is the output pulse after conversion of the pump pulse (106) in the gain medium (107). It is spatially coherent and thus provides a close to diffraction limited beam quality. There are some energy losses in the conversion in the gain medium, resulting in a wavelength increase of the signal pulse from the pump pulse. The signal laser pulse has substantially the same temporal characteristics as the pump pulse.

[0042] The output pulse (110) is the output pulse after beam shaping of the signal laser pulse (114). The output pulse has substantially the same temporal characteristics as the signal laser pulse and therefore as the pump pulse.

[0043] The temporal characteristics of the pump pulse (106), as determined by the controller (101) and the diode driver (103) include the definition of the temporal shape of the laser pulse, for example, pulse width, pulse shape, peak power, or any power modulation that is imprinted on the temporal profile. Furthermore, the temporal characteristics also include definitions about the repetition frequency of the pump pulses, or configurations that can create one or multiple pulses within the time window of the main repetition frequency of the laser system. Multiple7503541\70012\FG: 104682020.1pulse definitions would include number of pulses, their spacing relative to each other, and definitions about each individual pulse in that subgroup such as power or temporal shape characteristics already mentioned.

[0044] The temporal characteristics of the signal laser pulse (114) reflect the temporal characteristics of the pump pulse (106). This includes the temporal shape of the laser pulse, for example, pulse width, pulse shape, peak power, or any power modulation that is imprinted on the temporal profile. It also includes the repetition frequency of the pulse train and whether one or multiple pulses are present within the time window of the main repetition frequency. It also replicates the shapes of multiple pulses if present, such as the number of pulses, their spacing relative to each other, and variations amongst each individual pulse in that subgroup such as power or temporal shape characteristics.

[0045] The temporal characteristics of the output pulse (110) reflects the temporal characteristics of the signal laser pulse (114) that in turn reflects the temporal characteristics of the pump pulse (106). This includes the temporal shape of the laser pulse, for example, pulse width, pulse shape, peak power, or any power modulation that is imprinted on the temporal profile. It also includes the repetition frequency of the pulse train and whether one or multiple pulses are present within the time window of the main repetition frequency. It also replicates the shapes of multiple pulses if present, such as the number of pulses, their spacing relative to each other, and variations amongst each individual pulse in that subgroup such as power or temporal shape characteristics.

[0046] In traditional Q-Switched laser systems the gain medium receives pump light over a long period of time. To allow the gain medium to store enough energy and avoid lasing during the long pump pulse, the cavity needs to be blocked. This is typically achieved optically using an electro-optical or acousto-optical Q-Switch. The Q-Switch crystal holds off lasing in the cavity for the duration of the pump pulse by introducing a loss and reduces that loss closet to zero at the end of the pump pulse. This allows the laser pulse to build up in the cavity and a short pulse is emitted. The Q-Switch crystal is driven with an electronic pulsed driver that generates a short high voltage pulse or an RF pulse in the case of the acousto- optical Q-Switch.8503541\70012\FG: 104682020.1

[0047] All of these elements create a considerable amount of heat (as illustrated in Fig. 2) that needs to be removed through cooling, often resulting in water-cooled designs. It also impacts the overall efficiency of the laser, limiting the electrical to optical efficiency to less than 10%. Eliminating the need for a Q-Switch, eliminating losses in the gain medium during the pump pulse, and operating the components more efficiently reduces overall complexity and improves wall plug efficiency.

[0048] Fig. 2 is a set of two flow diagrams (124 and 128) from electrical input power to waste heat and useful laser light for a traditional Q-Switched laser system (121 ) and the first system proposed in this disclosure (122). For the traditional Q- Switched laser, the input (123) is split into Q-Switch driver (10%), diode driver (70%) and control electronics (20%). The diode driver, the diodes, and the diode light that pumps the gain medium lose the majority of their input power and only about 4% results in usable laser light (126). The remainder of about 96% is converted into heat (125), needs to be cooled, and is wasted.

[0049] The approach in this disclosure does not use a Q-Switch and the losses in the gain medium are considerably less, resulting an overall efficiency of about 26% (130) of usable laser light and about 74% (129) of waste heat from the overall input (127). The efficiencies for the diode driver and diodes are similar and the majority of the improvement is in the conversion of the diode light into useable laser light. The whole system is more than 6 times efficient, resulting in significantly less input power and less cooling capacity required.

[0050] Fig. 3 shows a timing diagram of a typical Q-Switched laser (141). The pump pulse (143) transfers energy to the gain medium over time periods much longer than the short pulse laser output (146). The Q-Switch hold off lasing in the cavity (145) until the stored energy is build up (144) and opens the cavity to allow the laser pulse to build up (146). In contrast, the first system described in this disclosure directly generates the short pulse (147).

[0051] Modem optical pump sources that provide the energy to the gain medium are typically solid-state laser diodes, although flash lamps are still used in some systems today. The majority of laser diodes used for laser pumping emit in the near-infrared portion of the spectrum since that wavelength range overlaps with the majority of absorption features in the most commonly used gain media. For9503541\70012\FG: 104682020.1example, Nd doped host materials or Yb doped host materials are primarily pumped in the 780nm to 1000nm wavelength range. Laser diodes emitting light in this spectral range primarily use a GaAs semiconductor to efficiently convert electrical current to light.

[0052] Recent advances have enabled laser diodes to operate at significantly higher peak powers offering energies similar to that of the laser output pulse in an equivalent pulse duration. This would not have much benefit in a laser utilizing a long storage lifetime. However, if the peak power of the laser diode can provide the same energy at time scales similar to that of a Q-Switched laser, the laser pulse can be generated directly from the laser diodes. A difficulty is that the beam quality from a laser diode M2is larger than 10 in the slow axis implying that the beam can only be focused to a size that is 10 times larger than a comparable diffraction limited beam. In contrast, a Q-Switched laser system often has beam qualities of M2between 1 and 2 and can readily be focused to a small spot. One system described in this disclosure uses an intermediate gain process to convert the poor beam quality of a laser diode into a beam with good beam quality while maintaining its temporal timescales.

[0053] Laser diodes can be implemented in different form factors from single to multi emitter designs. The highest power output is achieved in a laser diode bar geometry. The basic structure of a laser diode bar is comprised of several layers that sandwiches the active light emitting region. This structure leads to a very thin aperture in one axis, called the fast axis and a wide aperture in the other axis, called the slow axis. Single mode emitters typically limit the slow axis aperture to similar dimensions as the fast axis aperture to obtain a more symmetric output beam, but these geometries are limited in the overall output power they can generate. It is more efficient for high power devices to lengthen the slow axis aperture. For typical high-power devices this slow axis aperture is 5-10mm.

[0054] Other designs such as VCSELs (Vertical Cavity Surface Emitting Lasers) overcome some of the aperture limitation and create more symmetric beams with better beam quality, but the power output they can achieve is lower than that of diode bars.

[0055] The geometry of the laser diode aperture impacts the beam quality that can be achieved. The two axis (fast axis and slow axis) have different optical10503541\70012\FG: 104682020.1properties. In fact, the beam quality in the fast axis is close to diffraction-limited with an M2close to 1 , however the slow axis exhibits poor beam quality with an M2of more than 10. To achieve a tight focus, similar to that of a Q-Switched laser, both axes must be close to diffraction limited. While this can be achieved in the fast axis, the slow axis cannot be made diffraction limited.

[0056] Fig. 4 shows a detailed view of a single laser diode bar that is used as part of the diodes depicted in Fig. 1 (105). One configuration shows the raytracing and emitted beam profile with no correction optics (161) and another configuration shows the raytracing and emitted beam profile with a fast axis collimator (162). Multiple of these diode bars, arranged in a diode stack are coupled into the gain medium (107). Without any additional optics, the laser diode bar (163) diverges (164) in both directions creating an intensity profile on a detector (165) with a Gaussian distribution in the fast axis (169) and a Super Gaussian distribution in the slow axis (170). Since the diode bar is diffraction limited in the fast axis this distribution behaves like an ideal Gaussian beam. This implies that it has excellent beam quality but the thin semiconductor aperture in that axis imparts a large divergence angle as the beam propagates. A micro-optic (166) placed immediately after the laser diode aperture (163), called a fast axis collimator (FAC), can be employed to reduce this divergence and achieve excellent collimation while maintaining close to diffraction-limited beam quality (167). When viewed on a detector (168), the fast axis distribution has a Gaussian profile but with significantly less spatial width (171). The slow axis distribution remains a Super Gaussian distribution.

[0057] The slow axis consists of multiple emitters that are distributed across the aperture. The intrinsic width of each emitter and the total number of emitters causes the beam in the slow axis to be not diffraction-limited. Micro-optics, called slow axis collimators (SACs) can mitigate this issue by individually collimating each aperture but the distributed nature of the source is still present, preventing diffraction-limited focusing in the slow axis across the full aperture.

[0058] In summary, the beam from a laser diode bar can be modified to behave like a diffraction-limited beam in the fast axis, but not in the slow axis. Thus, a beam from a laser diode bar, although providing high peak power cannot be focused in11503541\70012\FG: 104682020.1both axis like a beam generated from a Q-Switched laser. Another mechanism needs to be employed to improve the beam quality.

[0059] To achieve comparable energy levels to that of a Q-Switched laser, multiple laser diode bars need to be employed. The most common and compact arrangement is a stack of laser diode bars. Fig. 5 illustrates the raytracing of such a laser diode stack with the side view (181 ), looking at the fast axis, and the top view, looking at the slow axis (182). The light output from the diode stacks (105) is situated in the optical plane 183. In this example a beam shaping optic (113) comprising a single spherical lens (184) focuses the light into the gain medium (107) situated at the optical plane 185. The fast axis achieves a small focus, however the slow axis cannot be focused and the beam has a much larger size. The lens in this raytracing diagram is a single spherical optic. Alternatively, a different design can provide optical power in just one axis using one or multiple cylinder lenses or a combination of cylinder and spherical lenses. In addition, a specific curvature prescription (aspheric or free-form lenses) can be used to mitigate optical aberrations and provide a better focus.

[0060] One embodiment that couples the pump light output (106 and 183) into the gain medium (107 and 185) using a beam shaping optic (113 and 184) is illustrated in Fig. 5. Although the focusing in the fast axis can achieve a small spot, the beam will diverge after the focus (185). To maintain the spot size over the volume of the gain medium (107) a waveguide structure can be employed. This is illustrated in the raytracing diagram in Fig. 6, showing the propagation of rays in the fast axis, entering the waveguide. After the rays (201 ) enter the waveguide structure (202), internal reflections on the side walls, due to the material index or a coating applied to the surface, confine the rays to the volume of the waveguide (203). The waveguide would be configured to encompass the gain medium (107) to allow the pump light to interact with the gain medium over the length of the gain medium. In one embodiment the gain medium itself could act as a waveguide. In another embodiment a separate waveguide would encompass the gain medium.

[0061] The waveguide (202) consists of an optical material with a refractive index higher than its surroundings. This index difference confines rays traveling in the waveguide since they experience total internal reflection above a critical angle of12503541\70012\FG: 104682020.1incidence on the sides of the waveguide. In some embodiments the waveguide is the gain medium itself, in other embodiments the gain medium can be situated inside a separate waveguide as long as the refractive index of the gain medium and the waveguide are matched. The material surrounding the waveguide can be any material with a lower refractive index. The index of refraction difference affects the critical angle for the total internal reflection and in turn limits the angles that are generated by the pump coupling optic (184) entering the waveguide (185). Coatings can be applied to the sides of the waveguide to optimize the total internal reflection angle and attach a wider range of materials. Finally, stacks of materials with different refractive index can be utilized, as long as there is an intermediate element surrounding the waveguide with a lower refractive index than the waveguide itself. Although planar waveguides are a preferred embodiment due to their aspect ratio matching the distribution of the pump light, other geometries such as elliptical, circular, or tapered waveguides can be employed as well as long as they can match the aperture of the pump distribution. Other waveguide embodiments using microstructures, Photonic crystals, or hollow core geometries can be employed, particularly when gaseous gain media are used.

[0062] The lens (184) transforms the arrangement of diode bars into incident angles on the waveguide. The largest angle and thus the largest distance from the optical center line, needs to be shallower than the internal reflection angle. This limits the extent of the diode bars that can be used. The pitch between neighboring bars in a diode stack needs to accommodate the fast axis collimator and mechanical components and is typically larger than 1 millimeter.

[0063] An embodiment that shows how the beams from the laser diode bars can be brought closer together in the fast axis is shown in Fig. 7. A set of micro optics (221) can be employed to translate the individual emitters in the fast axis and provide a closer spacing (222) since the extent of the beam in the fast axis after collimation is typically smaller than that of the pitch.

[0064] As described earlier, stimulated emission processes generate a coherent beam with good beam quality. Since long storage times are not required due to the short pump pulse durations, short stimulated emission processes involving different energy levels than the typical 4-level systems that operate on a13503541\70012\FG: 104682020.1timescale much closer to that of the extracted pulse duration can be considered. In fact this method, aims to employ a gain process in which the optical gain is present substantially only during the pump light pulse and without substantial energy storage in the gain medium. Without substantial energy storage in the gain medium this process is highly efficient.

[0065] Examples of these short stimulated emission gain processes are quasi 2 level or 3-level energy structures in gain media where the absorption and emission transitions are much closer together than in a 4-level system. The storage lifetime of these levels is closer to the extracted pulse duration but they can still provide good beam quality for the output beam. These materials would provide the gain medium for incoming pump light with little to no storage and immediately convert the pump light into the signal laser pulse. It is beneficial for the gain media to allow pumping with solid state diode wavelengths between 800 and 1000 nanometers.

[0066] Examples for gain media include but are not limited to host materials such as YAG, YV04, YLF, GdVO4, KGW, YAP, LiSAF, LiSGaF, LiCAF, Sapphire, Alexandrite, Silica glasses, Fluoride glasses, or Phosphate glasses with doping materials including but not limited to Neodymium (Nd), Ytterbium (Yb), Thulium (Tm), Erbium (Er), Holmium (Ho), Chromium (Cr), Cerium (Ce) or Titanium (Ti).

[0067] Another set of processes that can be exploited is shown in Fig. 8, depicting the wavelength behavior of three fundamental light scattering processes (241 ) from photons at a center wavelength (242). Rayleigh scattering (243) is an elastic process that maintains the photon energy. Brillouin scattering (244) involves phonon interaction in the material and causes a larger energy shift although it can be elastic and inelastic. Raman scattering (245) involves energy shifts due to molecular vibration or rotation and has the largest energy shift. These processes can occur in an anti-Stokes (246) regime where the light gains energy and a Stokes regime (247) where the light loses energy. These conversion processes can occur spontaneously or in a stimulated fashion. The stimulated process amplifies photons coherently and creates a beam with good beam quality.

[0068] Therefore, another choice for the short stimulated emission process is the stimulated Brillouin scattering process where a pump photon excites a phonon mode in the gain medium and then can be emitted through amplification. The14503541\70012\FG: 104682020.1timescale of this process is less than a nanosecond and occurs much faster than the expected pulse duration.

[0069] Yet, another choice for the short stimulated emission process is the stimulated Raman scattering process where a pump photon excites a vibrational or rotation mode of the gain medium and then can be emitted through amplification. The timescale of this process is less than a nanosecond and occurs much faster than the expected pulse duration.

[0070] The short stimulated emission processes described here typically require high intensity of the pump beam to provide sufficient gain for the output beam and therefore require a small aperture to couple the pump light into. This gain medium can be realized using waveguides with a small aperture in one direction and a large aperture in the opposite direction, since they closely match the expected beam profile of a focused diode stack. Waveguides can be chosen to guide the signal and pump light or only the pump light, depending on the size of the aperture.

[0071] The gain medium can be made out of a bulk material with a sufficiently high gain coefficient for the amplification process for either the energy storage process or the Brillouin or Raman process. Alternatively, the gain medium can be a gas with sufficient density to provide the required gain.

[0072] Examples for gain media using a Brillouin or Raman process include but are not limited to Silica glasses, Chalcogenide glasses, Fluoride glasses, Phosphate glasses, Lead glasses, Tellurite glasses, Silicon, Germanium, Gallium Arsenide, Bismuth Oxide, Sodium Chloride, Lithium Niobate, nonstructured and structured fibers, or nanostructured materials such as carbon nanotubes. Materials can also include the laser gain materials described earlier in a doped or undoped configuration. Alternatively, gases can be used as a gain medium to provide Brillouin or Raman gain. Examples include Nitrogen, Oxygen, Water vapor, Methane, Ammonia, Hydrogen, Carbon Dioxide, or any of the noble gases.

[0073] The amplification is illustrated in Fig. 9 where a pump beam (261) is amplifying a seed beam (262) as it travels through a gain medium. The pump beam is depleted and the seed beam is amplified. The seed beam can be supplied by an external source or spontaneously generated.15503541\70012\FG: 104682020.1

[0074] Fig. 10 shows an embodiment in a single pass configuration in a top and side view. The pump beam (281) is coupled into the gain medium (283). If a seed beam is necessary, an aperture in the coupling optics allows a seed beam to enter (282). The system can be seeded with an external beam (282) or generate the seed beam itself using spontaneous emission. The geometry of the gain medium (283) causes the amplified signal beam (284) to be larger in the top-view and smaller in the side view. Here the top-view corresponds to the slow axis of the diode array and the side view to the fast axis of the diode array. Note that in contrast to the diode pump beam the amplified laser beam (284) is coherent and is not made up from individual sources and thus allows for focusing in the slow axis (286) using a set of transform optics (285 and 287) while maintaining close to diffraction-limited beam quality. These optics are typically cylinder optics and can include free form or aspheric surface figures to compensate for aberrations. The output beam (289) is collimated and close to diffraction-limited in both axis (288). A dichroic beam splitter (290) can be used to reject the remainder of the pump light (291 ) although most of the light will naturally diverge away from the signal beam.

[0075] Fig. 11 shows another embodiment in a resonator configuration in a top and side view. The pump beam (301) is coupled into the gain medium (304). An aperture in the coupling optics allows the resonating beam to enter (302). Two end mirrors (303 and 305) provide feedback to the circulating signal beam (302). The mirror (305) is partially reflective to allow for some of the circulating signal light (302) to escape the cavity. Both mirrors (303 and 305) can include a surface figure to optimize for cavity modes and aberrations. The geometry of the gain medium (304) causes the amplified signal beam (306) to be larger in the top-view and small in the side view. Note that in contrast to the diode beam the amplified signal beam (306) is coherent and is not made up from individual sources and thus allows for focusing in that axis (308) using a set of transform optics (307 and 309) while maintaining close to diffraction-limited beam quality. These optics are typically cylinder optics and can include free form or aspheric surface figures to compensate for aberrations. The output beam (311) is collimated and close to diffraction-limited in both axis (310). A dichroic beam splitter (312) can be used to16503541\70012\FG: 104682020.1reject the remainder of the pump light (313) although most of the light will naturally diverge away from the signal beam.

[0076] A timing diagram for the pulsed output is shown in Fig. 12. In a typical Q- Switch laser one pulse (321) would be emitted at a temporal spacing (323) given by the repetition frequency of the laser system. This temporal spacing is often limited by application requirements, particularly in material processing applications, such as repositioning of the focus or controlling the energy on target. The direct generation method presented in this disclosure offers additional pulse formats. Specifically, multiple pulses (323) can be generated at the same overall pulse spacing (322) allowing for higher energy in that subgroup of pulses or lowering the requirements for the peak power of the laser diode. Since the spacing within the group of pulses can be small compared to the overall pulse spacing, this approach is still compatible with the processing requirements as they resemble a single pulse for any implementation system upstream from the laser.

[0077] Fig. 13 shows a spectral representation of the wavelengths involved in a Raman process. The laser diodes are centered around a common wavelength (341 ). It is expected that the laser diodes will have a variation of their center wavelength with different production batches and also will have a spectral emission width that is typically less than 10nm (342, 343, and 344). Each of the laser diodes would support a slightly different Raman gain spectrum (345, 346, and 347) but since the gain spectrum is broad and extends up to 100nm away from the center wavelength, even small source variations of the laser diodes can produce a stimulated gain at a common wavelength (349) for the output beam (348). This wavelength can be chosen by adjusting the wavelength of the seed beam (282) or by adjusting optics in the single pass or resonant cavity that select a specific amplification wavelength.

[0078] The wavelength of high-power laser diodes is typically centered at 808nm, 915nm, 940nm, or 975nm. A Raman-shifted signal wavelength would therefore be in the wavelength range of 1.0-1.1 micrometers, falling in the same band as existing laser systems using Nd- or Yb- doped host crystals and be able to utilize the same system optics for a drop-in replacement.17503541\70012\FG: 104682020.1

[0079] Timing systems that drive traditional Q-Switched lasers would be readily compatible with this approach, since the timing signal would drive the diodes directly and the jitter is inherently more stable. The precision timing of the pump light emission of the individual diode stacks or diode bars is not critical as long as they fall within the envelope of the overall pulse generation of the signal output.

[0080] A short pulse output beam with close to diffraction-limited beam quality generated through the method in this disclosure can also be used in frequency conversion processes that provide shorter wavelengths through harmonic generation, or longer wavelengths through parametric generation. These wavelengths can be employed for more efficient material processing applications or better compatibility with applications requiring specific transmission or absorption wavelengths. The output can also be utilized as a pump source for gain materials with a short lifetime.Second Part

[0081] The second part of this disclosure describes an implementation that uses a temporally shaped pump pulse with multiple segments that provides energy to the gain medium. In the first temporal segment, the pump pulse provides energy to the gain medium without crossing the lasing threshold and in the second temporal segment it provides energy to the gain medium to cross the lasing threshold on a time scale shorter than the laser pulse buildup in the cavity. The first segment of the pump pulse is similar to that of a typical diode pumped laser system. The diodes are driven by their normally rated current and emit over a time scale comparable to the spontaneous emission lifetime and deliver the majority of the stored energy into the gain medium. The spontaneous emission will limit the amount of stored energy that can be built up and the pump and cavity parameters are chosen such that the lasing threshold is not crossed with this first segment of the pump pulse.

[0082] The second segment of the pump pulse deposits its pump energy into the gain medium over a much shorter time scale. Increasing the gain above lasing threshold starts the build-up of the laser pulse in the cavity.

[0083] The dynamics of the energy deposition rate by the pump light and the energy utilization rate by the circulating laser pulse are described using rate equations that tie the rate of increase or decrease of the upper laser level population to the 18503541\70012\FG: 104682020.1power of pump and signal. The rate equation for the population of the upper laser level (N2) can be written asdN2 / dt = P(t) - A21 * N2- B21 * N2* S(t)where P(t) is the pump power, A21 the spontaneous emission rate, B21 the stimulated emission rate and S(t) the signal power of the circulating laser pulse. Due to its short duration, compared to the stimulated emission lifetime, for the second segment of the pump pulse the spontaneous emission term does not have a significant impact, and will be neglected. The temporal evolution of the pump and signal level as well as the upper laser level population N2 during the second segment of the pump pulse is shown in Fig. 17. The graphs show the pump and signal power in 421 and the upper laser level population N2 in 422. The pump power (423) starts its second segment at time T2 (426), marked by an increase of the power compared to the first pump segment. This increases the upper laser level population at a rate proportional to the pump power (425). At this time, the circulating laser pulse (424) starts to build up but does not have enough signal power to deplete the upper laser level and the upper laser level population N2 continues to increase at the rate proportional to the pump power. At a time T3 (427) close to the peak of the circulating laser power (424) the circulating laser pulse gains enough energy to start depleting the upper laser level population. At this time the second segment of the pump power ends, the pump power decreases, and provides no further increase in upper laser level population. The circulating laser pulse has enough energy to decrease the upper laser level population, therefore depleting the stored energy in the gain medium, and reducing the gain below the gain threshold. In summary, the pump power P(t) increases the upper laser level population at a rate that is proportional to its power and the circulating laser pulse decreases the upper laser level population at a rate that is proportional to the product of signal power S(t) and the population level N2 itself. This implies that the circulating laser pulse only depletes the upper laser level population if it reaches a significant power level. The gain (G) in the gain medium can be expressed asG = EXP [ sigma * N2 * I ]where sigma is the emission cross section and I the length of the gain medium. It is apparent that if there is no significant depletion of the upper state laser level,19503541\70012\FG: 104682020.1the gain is not impacted. Only if the circulating laser pulse has enough power, the stimulated emission term in the rate equation will deplete the upper laser level population and consequently the gain in the gain medium. It is apparent that the pump light from the second segment of the pump pulse needs to be deposited into the gain medium before the circulating laser pulse has built up sufficient power to deplete the upper laser level population and thus the gain. If the addition of stored energy in the second segment of the pump pulse is slower than the build-up of the laser pulse, the circulating laser pulse would deplete the stored energy before it has a chance to build up to a significant energy level and the gain in the gain medium will drop below the lasing threshold and effectively suppress further laser build-up with no laser pulse emitted. If, however, the addition of stored energy in the second segment of the pump pulse is faster than the build-up of the laser pulse, the laser pulse will not significantly deplete the stored energy during its build up stage and the gain will be above laser threshold long enough until the circulating laser pulse is built up to a significant energy level. Only then will the laser pulse deplete the stored energy and lower the gain. This will typically start just before the circulating laser pulse reaches its maximum power (427).

[0084] The time scale of the laser build up in the cavity; the time between 426 and 427 in Fig. 17, is typically on the order of tens of nanoseconds with a typical build up time between 30 nanosecond and 200 nanoseconds, although it depends on the cavity length and overall gain dynamics. Therefore, if the second segment of the pump pulse can deliver its energy in a time shorter than this build-up time, the laser pulse has sufficient time and gain to build up to full energy. Consequently, the rise time of the second segment of the pump pulse should be less than the typical build up time and shorter than 30 nanoseconds is preferred.

[0085] To deliver enough pump energy to cross the gain threshold the power of this second segment of the pump pulse needs to be much larger than that of the first segment since the emission time is much shorter. Therefore, the diodes need to operate at significantly higher power, at least three times above the power of the first segment of the pump pulse. Notably, the increase in power does not significantly increase the thermal load on the diodes since we only need to provide enough energy to cross the lasing threshold and the majority of the20503541\70012\FG: 104682020.1stored energy has already been delivered in the first segment of the pump pulse. The pump diodes are effectively operated at a thermal limitation in the first segment of the pump pulse and at a peak power limitation in the second segment of the pump pulse. The thermal limitation will put an upper limit on the achievable repetition rate of the system.

[0086] Pumping the gain medium across the gain threshold on such a short time scale directly initiates the buildup of the laser pulse, hence the timing signal that initiates the second segment of the pump pulse is tightly coupled to the occurrence of the laser pulse and can thus be used as a trigger signal with small temporal jitter. No active Q-Switch element is required which reduces cost and system complexity.

[0087] It should be noted that the response time of the laser diode bars is instantaneous for the temporal pulse shapes contemplated in this disclosure. The diodes follow the high current pump pulse of the diode driver with no impact on the rise time, therefore the diode pump pulse can have transition times on timescales of nanoseconds; fast enough to allow for the fast energy deposit required.

[0088] Fig. 14 illustrates the block diagram of the laser system presented in this disclosure. Laser diodes (361 and 362) emit pump light (364) that is coupled into a gain medium (363). In one embodiment that is described in Fig. 19, the coupling, particularly for the end-pumped configuration is a light funnel with a specific shape (363). The diodes can be arranged in a side pumped configuration (361 ) or in an end-pumped configuration (362) or a combination of both of them. A passive Q-Switch element (364) can be included in the cavity to provide a mechanism to lower the loss. A resonating cavity is formed with the high reflective end mirror (366) and a partially reflective end mirror (365). The mirrors can be curved to allow for definition of the spatial mode structure in the cavity. A timing signal generator (362) provides timing signals to a diode driver (368) that generates the high current temporal pulse (369) with multiple temporal segments to drive the laser diodes (370). If a passive Q-Switch element (364) is used (see Fig. 15) it switches from a high loss to a low loss state once the cavity begins to circulate light (367) and enables full extraction of the stored energy into the laser pulse. The output laser pulse (371) is transmitted through the partially reflective21503541\70012\FG: 104682020.1end mirror (365). For embodiments without a passive Q-Switch element (364) (see Fig. 16) multiple high power pump pulses from the diode driver produce multiple laser pulses.

[0089] One embodiment combines this with a passive Q-Switch crystal element, notably not for initiating the Q-Switch pulse itself, but rather as a mechanism to lower the loss in the cavity. The timing diagram for this configuration is shown in Fig. 15. The timing signals that drive the laser diode driver are shown in 383, the gain / loss of the cavity is shown in 382, and the power of the pump light and laser light are shown in diagram 381. At an initial time T1 (390), a timing signal (388) from the timing signal generator (372) sent to the diode driver (368) starts the first segment of the current pulse (384) to the diodes (369). The shape and amplitude of the current pulse can be determined by features of the timing signal or by parameters in the diode driver. In this example, power of a pump light pulse corresponds to the current pulse (384). The first portion of the current pulse (384) and corresponding pump light starts the buildup of the stored energy in the gain medium and correspondingly the gain in the cavity (386). At this point the cavity has a higher loss (387) than the gain due to the losses in the cavity, cavity mirrors, and the passive Q-Switch and lasing is suppressed. The buildup of the stored energy continues, but is limited by losses due to spontaneous emission from the upper laser level and therefore reaches a plateau. At a time T2 (391) another timing signal (389) starts the second segment of the current pulse (384) to the diodes (369). The shape and amplitude of the current pulse can be determined by features of the timing signal or by parameters in the diode driver. This segment of the pulse has significantly higher power delivered over a shorter period of time and causes increases in current, the corresponding pump light, and the gain in the laser cavity in a much shorter period of time compared to the first segment. This increases the stored energy in the gain medium and therefore the gain in the cavity to before the laser pulse that is building up has a chance to deplete this additional stored energy, so that at time T3 (392) the laser pulse is built-up and extracts the stored energy which will finally lower the gain in the cavity again. Due to the increased power on the saturable absorber, it switches to its low loss state and the circulating laser pulse (385) continues to build up in the cavity with an overall lower loss. This allows the circulating laser pulse to extract22503541\70012\FG: 104682020.1most of the stored energy from the gain medium. It is notable that the timing signal at T2 (391 ) can be used to control the temporal position of the laser pulse with a low jitter. Since the transition above lasing threshold is primarily determined by the second segment of the pump pulse, the passive Q-Switch element does not add any significant amount of jitter and the timing of the laser pulse is still primarily determined by the timing signal of the second segment of the pump pulse.

[0090] In another embodiment the second segment of the pump pulse is repeated.The timing diagram for this configuration is shown in Fig. 16. The timing signal that drives the laser diode driver is shown in 403, the gain / loss of the cavity is shown in 402, and the power of the pump light and laser light are shown in diagram 401. At an initial time T1 (410), a timing signal (408) generated by the timing signal generator (372) sent to the diode driver (368) starts the first segment of the current pulse (404) to the diodes (369). The shape and amplitude of the current pulse can be determined by features of the timing signal or by parameters in the diode driver. This starts the buildup of the stored energy in the gain medium and the gain in the cavity (406). At this point the cavity has a higher loss (407) than the gain due to the losses in the cavity mirrors and lasing is suppressed. The buildup of the stored energy continues, but is limited by losses due to spontaneous emission from the upper laser level and therefore reaches a plateau. At a time T2 (411 ) another timing signal (409) starts the second segment of the current pulse (404) to the diodes (369). This segment of the pulse has significantly higher power delivered over a shorter period of time and increases the gain in the laser cavity in a much shorter period of time compared to the first segment. This increases the stored energy in the gain medium and therefore the gain in the cavity to before the laser pulse that is building up has a chance to deplete this additional stored energy, so that at time T3 (412) the laser pulse is built-up and extracts the stored energy which will finally lower the gain in the cavity again. Since the gain in the cavity is increased faster than the buildup time of the laser pulse (405) the laser has sufficient time to extract the stored energy until it falls below the loss (407) and thus primarily extracts the energy delivered with the second segment of the pump pulse. It is crucial to note that the timing signal at T2 (411 ) can be used to control the temporal position of the laser pulse23503541\70012\FG: 104682020.1with a low jitter. A repeated second segment of the pump pulse would again push the cavity above lasing and generate another laser pulse. This process can be repeated multiple time since the cavity is kept close to the lasing threshold and only the energy delivered from the second segment of the pump pulse is extracted.

[0091] The shape of the second segment of the pump pulse affects the buildup dynamics of the laser pulse in the cavity and the majority of the energy of the second segment of the pump pulse should be delivered to the gain medium before the laser pulse in the cavity builds up to energy levels that significantly deplete the stored energy. To achieve best performance and avoid throttling the laser buildup, the second segment of the laser pulse should deliver most of its energy before the laser pulse is built up. Different temporal pump pulse shapes in the second segment are illustrated in Fig. 18. Note that the integrated energy, illustrated by the hatched area underneath the pump pulse is the same in all cases. The first diagram (441) shows a long second segment pump pulse (444) in the shape of a Gaussian, super-Gaussian, or rectangular pulse profile with a duration that is longer than the buildup time of the laser pulse (447). The laser pulse (447) builds up slowly, eventually reaches saturation, and depletes the stored energy before the pump pulse was able to store all of its energy in the gain medium. The cavity gain drops below the loss and suppresses further lasing buildup. No significant amount of energy is extracted. Secondary pulses are possible, but they also will not have significant amount of energy associated with them. In this configuration the pulse duration of the second segment of the pump pulse is longer than the buildup time of the laser pulse and lasing is suppressed. The second diagram (442) shows a shorter second segment pump pulse (445) in the shape of a Gaussian, super-Gaussian, or rectangular pulse profile with a duration shorter than the buildup time of the laser pulse (448). After the second segment of the pump pulse deposits almost all of its energy into the gain medium, the laser pulse (448) builds up, reaches saturation and depletes the stored energy in the gain medium at T3”. This results in a high energy pulse, albeit with a longer pulse duration. The third diagram (443) shows a shaped second segment of the pump pulse, resembling a sawtooth waveform with most of its energy deposited into the gain medium early in the pulse close to T2. This24503541\70012\FG: 104682020.1results in a much larger initial gain and the laser pulse (449) builds up faster, achieving a shorter pulse duration than the pulse in 448.

[0092] Critically, the averaged energy deposition rate of the pump pulse over the duration of the second pump segment should exceed the average energy utilization rate of the circulating laser pulse to allow for the build up of a high energy laser pulse. As illustrated, the specific shape of the pump pulse and therefore the energy deposition rate affects the temporal parameters of the emitted laser pulse.

[0093] Adjustments to the shape, duration, and power level of the current pulse and therefore the pump pulse can also be used to compensate for changes in the system like environmental or mechanical either with an active feedback loop or through pre-calibrated correction functions.

[0094] Examples for gain media include but are not limited to host materials such as YAG, YV04, YLF, GdVO4, KGW, YAP, LiSAF, LiSGaF, LiCAF, Sapphire, Alexandrite, Silica glasses, Fluoride glasses, or Phosphate glasses with doping materials including but not limited to Neodymium (Nd), Ytterbium (Yb), Thulium (Tm), Erbium (Er), Holmium (Ho), Chromium (Cr), Cerium (Ce) or Titanium (Ti).

[0095] Timing systems that drive traditional Q-Switched lasers would be readily compatible with this approach, since the timing signal for the second segment of the pump pulse would directly generate the laser pulse. The precision timing of the pump light emission of the individual diode stacks or diode bars is not critical as long as they fall within the envelope of the overall pulse generation of the signal output.

[0096] A short pulse output beam with close to diffraction-limited beam quality generated through the method in this disclosure can also be used in frequency conversion processes that provide shorter wavelengths through harmonic generation, or longer wavelengths through parametric generation.

[0097] Single diode bars can be used for low energy output pulses, but scaling to higher energies in a single aperture output laser pulse currently requires multiple diode bars. This is typically implemented in a bar stack configuration.Nevertheless, improvements in engineering and fabrication, continue to scale the brightness and thus the peak power level of single diode bars higher and single diodes bars could be used in the future.25503541\70012\FG: 104682020.1

[0098] Coupling these diode stacks into the gain medium is challenging because the beam quality of the diode bars is poor, particularly in the slow axis and the fast axis has a high divergence angle without using fast axis collimator optics. The absorbed pump light in the crystal should be homogenous in the plane perpendicular to the optical axis to support a diffraction limited beam and low variations in the gain across the aperture. This is particularly important since the gain needs to exceed the gain threshold during the second pump segment across the whole crystal aperture, thus having a flat absorption profile reduces the possibilities of different regions of the gain medium crossing the gain threshold at different times. In the axis parallel to the optical axis the absorption should be close to areas where the crystal can be cooled to avoid strong thermal gradients.

[0099] Side pumping can achieve a homogenous absorption profile, but is typically not compatible with diode bar stacks and end pumping is challenging due to the poor beam quality that cannot readily create a beam profile that overlaps with the gain crystal. Pump funnels with diode stacks on one end and the gain medium on the other end have been employed to reduce the beam size of the pump and match it to the geometry of the gain medium, but due to conservation of brightness the pump beam diverges stronger and not all of the light can be absorbed.

[0100] The method proposed in this disclosure combines the end- and side-pumping geometry by placing the gain medium in the funnel. Furthermore, a free form funnel shape is placed around the gain medium that optimizes the beam homogeneity and the absorption along the gain medium. It is intuitive to see that the funnel arrangement provides an end-pumped geometry closer to the diodes and a side pumped geometry further away from the diodes.

[0101] The homogeneity of the absorption profile typically improves if the number of absorption passes through the gain medium is increased. This avoids the single pass Beer-Lambert exponential absorption profile that is inherent to a linear absorption function. With a linear absorption function and a single pass, the homogeneity is limited to the overall absorption, for example, if 90% of the light is absorbed, the absorbed intensity changes by a factor of 10 from entrance to exit. Employing multiple passes reduces the variations of the absorbed intensity and absorption profiles with less than 15% variation across the aperture are26503541\70012\FG: 104682020.1achievable if more than four passes are utilized, although the exact number depends on the overall geometry.

[0102] Placing the gain medium inside the funnel geometry and employing geometric features in the funnel geometry like hohlraums that confine the pump light and provide more absorption passes through the gain medium enable this approach. Furthermore, since the beam quality and beam profile of the diode stack is different in the horizontal and vertical directions, providing different apertures, and shapes in the funnel geometry would allow for a method to optimize the absorption profile. Homogenizing the light directed towards the crystal clamp at the end of the waveguide would provide a method for targeted distribution of the pump light absorption along the crystal axis. In addition, overlapping a rectangular pump geometry with a round absorption medium is aided by the ability to tailor the comers of the funnel and guide those portions of the light to homogenize the absorption profile.

[0103] With this approach absorption profiles with a peak-to-valley variation of less than 15 percent across the aperture of the gain medium are achievable, while at the same time distributing the absorbed light to areas along the crystal axis where heat extraction can be employed. Furthermore, most of the light can be absorbed by the gain medium, for an overall absorption efficiency close to one hundred percent.

[0104] It should be appreciated that this optimization is highly nonlinear and complex optimization strategies should be employed, including but not limited to Monte Carlo, Genetic Algorithms, Simulated Annealing, Machine Learning.

[0105] One example of an optimization result is shown in Fig. 19. The raytracing diagram is shown in 461 with the absorption profile perpendicular to the optical axis in 462 and the absorption profile parallel to the optical axis along the length of the gain medium in 463. The diodes are placed at the entrance (466) of the funnel (464) and emit their diode light into the funnel. The gain medium, in this case a round cylinder (465) is in the center of the funnel. The shape of the funnel at the side opposite of the diodes matches the profile of the gain medium to confine the rays. The position of this overlap can be adjusted to optimize cooling performance. The shape of the funnel provides reflections to the pump light that offer multiple passes through the gain medium, thereby smoothing the absorption27503541\70012\FG: 104682020.1profile as shown in 462. The absorption profile in 463 is a combination of end and side-pump absorption characteristics and distributes the absorbed energy across the crystal.

[0106] In one embodiment the funnel can be made out of machined metal with a highly reflective coating such as gold or silver. In another embodiment, a free form optical transparent material can be used that utilized total internal reflection to provide the reflections inside the tunnel.

[0107] Various cooling methods are anticipated, ranging from clamp designs that can include cooling spokes to provide a cooling path at the entrance to the tunnel, to encased designs where the gain medium would reside in a cooling liquid or another optically transparent element to provide cooling as well as transmission for the pump light.

[0108] As mentioned previously, another application of this technology is the pumping of secondary laser gain media exhibiting short storage lifetimes. A classic example is a Ti: Sapphire host material with a storage lifetime of a few microseconds. Other gain materials such as Transition Metal-Doped Crystals like CrZnSe, Fe:ZnSe, CrLiSAF, or CrLICAF, as well as Dye lasers or Excimer lasers with storage lifetimes of a few nanoseconds and can also benefit from the pumping approach. For Ti:Sapphire specifically, the currently preferred implementation is pumping using a frequency converted Q-Switched laser with pulse durations of tens of nanoseconds. This allows for absorption in the green wavelength band and a negligible spontaneous emission loss. Direct pumping without the frequency conversion is possible if the gain material absorption is matched to the emission wavelength of the diodes. Other configurations that employ additional frequency conversion such as third- or fourth-harmonic conversion are possible as well. As showcased earlier employing a traditional Q- switched laser is complex and requires sophisticated pump lasers particularly for high energy and power operation that have scaling due to their intrinsic inefficiencies. Replacing the pump source with the direct diode approach can improve efficiency and reduce system cost and ultimately allow for scaling to higher power and energy levels. This approach is not limited to Ti:Sapphire, but can also be used for other gain materials with short storage lifetimes.28503541\70012\FG: 104682020.1

[0109] Although a single aperture output laser pulse as shown in Fig. 1 is a preferred implementation for end pumping, other pump configuration exist that can implement the configuration described here. Fig 20 shows a compact arrangement implementing the multi-segment temporal pump pulse. A timing signal generator (481) provides timing signals to a diode driver (482) that generates the high current temporal pulse (483) with multiple temporal segments to drive the laser diodes (484). It should be noted that multiple diodes can share the same diode driver and multiple diode drivers can share the same timing signal generator. The output of the single diode or multiple diodes (485) is coupled via an optional beam shaper (486) into a gain medium (487). The gain medium is matched to the spatial aperture and includes a passive Q-Switch (488) as described earlier and an optional frequency conversion module (489) to facilitate frequency conversion to shorter wavelengths if needed. One embodiment is a waveguide structure described earlier (see Fig. 6) to facilitate pump confinement. The frequency conversion module would employ a nonlinear crystal in either a bulk configuration or using a quasi-phase matched configuration using a periodically poled crystal. The latter configuration improves the conversion efficiency due to improved phase matching and is well matched to the aperture of the laser diodes due to its intrinsic high aspect ratio, resembling the aperture of the diode. The elements in this configuration (487, 488, and 489) can be individual components in close proximity or fabricated as a single part with the different components bonded together. Coatings to facilitate a resonant cavity with a high-reflectivity coating (491) and a partial-reflectivity coating (492) can be applied directly to the components or be included on separate optics. The cavity formed by (491) and (492) would couple the output laser pulse directly into the frequency conversion module if used. This would generate a frequency converted output laser pulse (490) after single pass conversion that would be directed towards the gain medium that is pumped (491). Optional beam shaping optics (493) can be used to match the output laser pulse to the aperture of the short lifetime gain medium that is ultimately pumped (491). A distinct advantage of a compact embodiment would be scalability in production with automation, testing, and yielding of the smaller modules. It should be noted that the output beam quality does not need to be diffraction limited, in fact for the high-aspect ratio29503541\70012\FG: 104682020.1configuration the output beam and beam quality resembles that of a diode bar. In side-pumped applications waveguiding in the short lifetime secondary gain medium is often employed and does not require diffraction limited beam quality. Although end-pumped applications prefer better beam quality, arrangements using diode stacks or diode arrays that exhibit similar output beam profiles and beam quality have been demonstrated with appropriate beam shaping optical arrangements and can readily be adapted here.

[0110] An embodiment depicting the integration into a side-pumped geometry is shown in Fig. 21. Here the pump modules (501 ) described in Fig. 20 are arranged around a central gain module (502). Without limiting the arrangement and the number of modules, the figure shows a specific arrangement using eight modules. The pump light is coupled into the side of the gain module and lasing or amplification through the gain module occurs in the direction perpendicular to the pump coupling plane (503). Different embodiments of the gain modules are contemplated. In one embodiment, the gain module would be a rod. Pump coupling occurs in the plane perpendicular to the rod axis or parallel to the rod axis. In another embodiment the gain module is a thin disk module with pump coupling aligned to the plane of the disk. Yet another variation uses a segmented disk with features that align with the geometry of the surrounding pump modules to improve pump coupling. This thin disk module is cooled through one face of the disk and signal extraction occurs through the other face. It should be noted that for most side pumped configurations, the beam quality of the pump is not critical and often guiding through internal reflection or scrambling of the pump light is preferred to facilitate a homogenous gain distribution. It is evident from the depiction that even though an individual module might produce a low energy output laser pulse, the arrangement of multiple pump sources will ultimately increase the pump energy available to facilitate scaling. As contemplated earlier, this geometry also improves reliability of the gain module as a whole since degradation or elimination of single pump modules would only marginally affect the overall performance of the gain module. Another embodiment of this arrangement are implementations using distributed gain media. As depicted in the side view on the right hand side in Fig 21 , the rod or disk can be divided into multiple slices. Each slice in combination with the pump module becomes a sub30503541\70012\FG: 104682020.1component of the full gain module. In a typical distributed gain arrangement, cooling channels are employed between the slices (504) to provide heat removal from the gain medium. Typically, gas or fluid cooling is employed and high average power levels can be achieved. In addition, this thermal management can also facilitate cooling of the diodes.

[0111] Another embodiment that showcases an end-pumped geometry is depicted in Fig 22. Here multiple pump modules (521) are arranged to pump into the direction of lasing or extraction (523) of the gain module (522). An optional turning mirror (524) can be used to direct the signal through the amplifier without interfering with the pump coupling. Alternatively, a reflector orthogonal to the direction of lasing can be used to provide the reflection through the gain module. Yet another alternative is to include a coating directly on the end surface of the gain element to provide the reflection back through the gain module. Optional optical elements (525) can be utilized to shape the pump beam to match the aperture and geometry of the gain module. This can be in form of lenses, light funnels, beam shaping micro-optics, or combination of those. This geometry can also be employed with cooling approaches using distributed gain media by segmenting the gain module into multiple slabs.31503541\70012\FG: 104682020.1

Claims

CLAIMSWhat is claimed is:

1. A method of generating an output laser pulse, the method comprising: driving a diode bar with a pulsed electrical current source according to pump pulse temporal characteristics to produce a pump light pulse;coupling, using a free-space optic, the pump light pulse into a gain medium configured to convert the pump light pulse into a signal laser pulse by a gain process in which optical gain is present substantially only during the pump light pulse and without substantial energy storage in the gain medium, such that the signal light pulse has signal light temporal characteristics corresponding to the pump pulse temporal characteristics; andshaping, using beam delivery optics, the signal laser pulse to generate the output laser pulse having output pulse temporal characteristics that also correspond to the pump pulse temporal characteristics and are thereby directly controlled by the pulsed electrical current source.

2. The method of claim 1 , in which the diode bar includes integrated optics for shaping the pump light pulse.

3. The method of claim 1 , in which the gain medium is configured as a waveguide to confine the pump light pulse along at least one axis.

4. The method of claim 1 , in which the gain medium operates using a stimulated Raman scattering process to convert the pump light pulse into the signal laser pulse.

5. The method of claim 1 , in which the beam delivery optics are configured to adjust the signal laser pulse to achieve the desired beam size and divergence.

6. The method of claim 1 , in which the gain medium is configured to produce an output laser pulse with a beam quality factor (M2) of less than 3 in both axes.

7. The method of claim 1 , in which the pump pulse temporal characteristics are varied to adjust the duration, repetition rate, or format of the output laser pulse.

8. The method of claim 1 , in which the output laser pulse is further processed through harmonic generation to produce shorter wavelengths.32503541\70012\FG: 104682020.

19. The method of claim 1 , in which a cooling system is configured to dissipate heat from the diode bar and gain medium.

10. The method of claim 1, in which the gain medium is operated in a singlepass configuration to amplify the signal laser pulse.

11. The method of claim 1 , in which the gain medium is operated in a resonator configuration with feedback elements to amplify the shaped pump pulse.

12. The method of claim 1, in which the diode bar includes micro optics for shaping the pump light pulse.

13. The method of claim 1, in which the gain medium operates using a stimulated Brillouin scattering process to convert the pump light pulse into the signal laser pulse.

14. The method of claim 1, in which the gain medium operates using a stimulated amplification process in a three or four level system to convert the pump light pulse into the signal laser pulse.

15. The method of claim 1, in which the output laser pulse is further processed through parametric conversion to produce longer wavelengths.

16. The method of claim 1, in which the gain medium is operated in a singlepass configuration using a seed to amplify the signal laser pulse.

17. The method of claim 1, in which the pump light pulse duration is less than or equal to 1000 nanoseconds or is less than or equal to 100 nanoseconds.

18. A method of generating an output laser pulse, comprising:driving a laser diode driver with a timing signal to produce a temporal current signal;driving a laser diode to produce a pump light pulse according to the temporal current signal comprising:a first temporal segment causing the laser diode to generate a first pump segment of the pump light pulse at a lower power level; anda second temporal segment causing the laser diode to generate a second pump segment of the pump light pulse at a higher power level compared to the lower power level;33503541\70012\FG: 104682020.1coupling the pump light pulse into a gain medium within a resonant cavity, the resonant cavity in combination with the gain medium defining a gain threshold, such that a circulating laser pulse experiences an energy increase when gain in the gain medium exceeds the gain threshold;inverting the gain medium during the first pump segment of the pump light pulse without crossing the gain threshold and further inverting the gain medium during the second pump segment of the pump light pulse to cross the gain threshold for lasing by delivering the higher power level at an energy deposition rate exceeding an energy utilization rate of the circulating laser pulse thereby allowing the gain to build up before depletion by the circulating laser pulse; andgenerating the output laser pulse as the circulating laser pulse depletes stored energy from the gain medium until the gain drops below the gain threshold.

19. The method of claim 18, further comprising utilizing a saturable absorber to lower losses in the resonant cavity to produce the output laser pulse.

20. The method of claim 18, in which the first temporal segment is between 100 nanosecond and 100 milliseconds, and the second temporal segment is between 1 nanosecond and 1000 nanoseconds.

21. The method of claim 18, in which the energy deposition rate is an average energy deposition rate from the beginning of the second temporal segment to the end of the second temporal segment thereby allowing the gain to build up before depletion by the circulating laser pulse.

22. The method of claim 18, in which the second pump segment of the pump light pulse has a rising edge duration of less than 30 nanoseconds to rapidly increase gain and initiate lasing.

23. The method of claim 18, further comprising repeating the second pump segment after the output laser pulse is generated to produce multiple laser pulses.

24. The method of claim 18, in which the laser diode comprises a diode bar, and the second temporal segment operates the laser diode at a peak power level at least three times greater than the first temporal segment.

25. The method of claim 18, in which the pump light pulse is delivered through an optical fiber or free-space optics before reaching the gain medium.34503541\70012\FG: 104682020.

126. The method of claim 18, in which the gain medium is a solid-state laser material with a host material selected from the group consisting of YAG, YV04, YLF, GdVO4, KGW, YAP, LiSAF, LiSGaF, LiCAF, Sapphire, Alexandrite, Silica glasses, Fluoride glasses, or Phosphate glasses.

27. The method of claim 18, in which the gain medium is a solid-state laser material with a doping material selected from the group consisting of Neodymium (Nd), Ytterbium (Yb), Thulium (Tm), Erbium (Er), Holmium (Ho), Chromium (Cr), Cerium (Ce) or Titanium (Ti).

28. The method of claim 18, further comprising dynamically adjusting the power levels and timing of the first and second pump segments based on feedback from a photodetector monitoring the circulating and or emitted laser pulse.

29. The method of claim 18, in which the resonant cavity comprises a partially reflective output coupler, and the circulating laser pulse builds up with a Q-switched profile.

30. The method of claim 18, in which the gain medium is end-pumped, side-pumped, or pumped through a hybrid geometry to optimize energy deposition and mode quality.

31. The method of claim 18, in which the second pump segment of the pump light pulse is shaped with a Gaussian, super-Gaussian, rectangular pulse, or sawtooth profile to optimize energy extraction efficiency and or adjust the pulse duration of the emitted laser pulse.

32. The method of claim 18, in which the output laser pulse is further processed through harmonic generation to produce shorter wavelengths.

33. The method of claim 18, in which the output laser pulse is further processed through parametric conversion to produce longer wavelengths.

34. A method of pumping a secondary gain medium, comprising generating an output laser pulse according to the method of claim 18 and delivering the output laser pulse to a secondary gain medium having a lifetime shorter than 1 millisecond.

35. The method of claim 34, in which the secondary gain medium comprises a gain material selected from the group consisting of Ti:Sapphire, CrZnSe, Fe:ZnSe, Cr: LiSAF, or Cr: LICAF, dye laser media, and excimer laser media.35503541\70012\FG: 104682020.

136. The method of claim 34, in which the secondary gain medium has an upper-state lifetime shorter than 10 microseconds.

37. The method of claim 34, in which the output laser pulse is delivered to the secondary gain medium using a side-pumped or end-pumped geometry.

38. The method of claim 34, in which the secondary gain medium is configured in a rod, thin-disk, or distributed gain-media architecture.

39. The method of claim 34, in which the output laser pulse is frequency converted, beam shaped, or temporally shaped according to any one of claims 19-33 prior to delivery to the secondary gain medium.

40. A laser system configured to perform the method of any one of claims 18-33, in which the laser system is implemented as an integrated laser module comprising the gain medium, resonant cavity, laser diode, and associated optics within a common housing.

41. A laser system configured to perform the method of any one of claims 1-33.

42. A pump light coupling system comprising:a gain medium positioned within a tapered enclosure, the tapered enclosure configured to regulate thermal conditions and confine pump light in a volume that at least partly encompasses the gain medium;a laser diode bar configured to generate the pump light pulse and direct it toward the gain medium; andan internal reflective surface within the tapered enclosure, the reflective surface configured to redirect and optimize pump light absorption within the gain medium, in which the pump light enters the tapered enclosure and undergoes multiple reflections to provide multiple absorption passes through the gain medium to improve spatial uniformity to better than 15% peak-to-valley of absorbed light.

43. The pump light coupling system of claim 42, further comprising an integrated cooling system within the tapered enclosure, in which the gain medium is thermally regulated by either a liquid cooling medium surrounding the gain medium or a thermally conductive structure that extracts heat from the gain medium.36503541\70012\FG: 104682020.1

Images