Open Access
Add to BibSonomy
Abstract
The bonds established with Onyx’s adhesive-free bonding (AFB) technique between optical materials, such as trivalent rare-earth ion doped YAG, un-doped YAG, or sapphire, rely on Van der Waals forces. The AFB technique is being applied to fabricate true crystalline fiber waveguides for high power, high beam quality laser emission, and walk-off compensated nonlinear crystal stacks for high efficiency, high beam quality optical nonlinear conversion.
© 2017 Optical Society of America
1. Introduction
Adhesive-Free Bond (AFB) composites are well suited for laser environments because they enable separation of different laser functionalities by producing strong and optically transparent bonds with minimal loss or damage without any organic or inorganic adhesive or chemical activation agents at their interfaces.
AFB composites for solid state laser components have formed the basis for providing components of crystals and optical ceramics to facilitate power scaling and compact solid state lasers. The key process step of forming an AFB is to obtain optical contact between the components to be bonded. The interfaces are bonded mainly through Van der Waals forces after at least moderate heat treatment and out-diffusion of volatile species. AFB may be achieved by application of pressure, capillary adhesion, or by bringing two cleaned surfaces into intimate contact.
AFB composites have favorable features such as 1) Elimination of ground state absorption losses of quasi-3-level lasing ions, especially for high power Yb:YAG based rod and slab systems; 2) Reduction of parasitic oscillations by using Adhesive-Free Bond (AFB) un-doped or laser radiation absorbing crystals; 3) Reduction of thermal lensing and other thermal effects with undoped YAG or sapphire acting as heat sink; 4) Passive q-switching with e.g. Cr4+:YAG as integral component of lasing element; 5) Mechanical support of thin lasing layers, down to about 2-4 µm thickness, thereby also essentially eliminating thermal effects in this geometry to a negligible level; and 6) Light guiding and wave guiding effects by combining dissimilar compatible materials of different refractive index, e.g. YAG lasing medium with sapphire cladding, YAG with spinel, and GGG with YAG cladding.
Materials that can be Adhesive-Free Bonded (AFB) include common laser host media such as oxides (YAG, GGG, Sapphire, Spinel, sesquioxides), fluorides (YLF, LuLiF, S-FAP), vanadate (YVO4, GdVO4), and other crystals, as well as glasses; optical ceramics, semiconductor metals, such as silicon, and semiconductor compounds; chalcogenides; and nonlinear optical crystals. In selected compatible cases (for example YAG/Spinel, YAG/Sapphire and YAG/Glass), dissimilar materials may be AFB'd with high bond strength. Although there are differences between composites, the mechanical strength and thermal shock resistance of similar materials are as high as non-composite samples, and may be finished and optically coated like conventional crystals. The composites show negligible stress birefringence, and are capable of keeping the transmitted wave front unchanged. If the refractive indices of both components of a doublet are exactly the same, there will not be any observable reflection at the interface. They form a large twin crystal where only X-ray analysis is able to detect the different crystal orientation. TEM has confirmed that the interface between a YAG/ Er:YAG composite is about 2 Å. The interfaces demonstrate very high laser damage resistance, and negligible scatter or absorption loss. A recent evaluation of the suitability of AFB Yb:YAG for high energy laser applications has concluded that the mechanical strength, optical quality and laser-induced damage threshold are not adversely affected by the existence of a bonded interface and are similar to those of monolithic substrates [1, 2].
This paper is organized as follows: In Section 2, we discuss the measurement of thermal properties, including thermal conductivity and heat transfer coefficient. In Section 3, we describe aspects of two major applications of AFB: true crystalline fiber waveguides (CFWs) for laser emission, and walk-off compensated (WoC) nonlinear crystal stacks for frequency conversion.
2. Thermal property investigation for AFB composites
In the design of single-clad and double-clad crystal fiber waveguides (CFW) as laser gain media, thermal conductivities of and heat transfer coefficients between core and cladding materials are critical for dissipation of heat, which is generated in the active core area. Temperature gradients induced by corresponding heat flow could lead to thermal lens effects, thermal depolarization in isotropic gain media such as Nd:YAG, and even thermal fracture. The value of thermal conductivities of and heat transfer coefficient between core and cladding materials can help model the possible thermal effects in laser gain media. Intuitively, lower thermal conductivities and heat transfer coefficients will cause excess heat to accumulate in the core region, resulting in a steeper temperature gradient, and more severe thermal effects.
Fabrication of CFW with Onyx’s Adhesive-Free Bonding technique requires thermal expansion coefficients of core and cladding materials to be compatible. Therefore, the core material is usually selected from trivalent rare-earth ion doped Yttrium Aluminum Garnet (RE3+:YAG), where RE3+ = Nd3+, Yb3+, Er3+, Ho3+ and Tm3+,. The inner cladding material for a double-clad structure, which has slightly lower refractive index than the core in order to enable waveguiding, is usually chosen from un-doped YAG, or laser-inactive ion doped YAG. Thermal conductivities vary for different dopants, and doping levels.
In this section, we summarize an interferometric method developed at Onyx Optics, Inc. to accurately measure the thermal conductivities and heat transfer coefficients between RE3+ doped YAG, and un-doped single crystal YAG. The reference value of thermal conductivity for un-doped YAG is taken from Ref [3], as basis. Such data enables estimates of power scaling modeling of crystal fiber waveguides.
The sample to be measured with a Zygo interferometer has been prepared by Onyx’s standard adhesive-free bond (AFB) technique for most reliable test results. Onyx makes doublets for testing intended core/inner cladding combinations, e.g. 3% Er:YAG/Undoped YAG, as shown in Fig. 1. The polished surfaces in the optical path should be of good quality with a parallelism <15 arc second so that the transmitted wavefront can be accurately measured.
Figure 2 shows the experimental setup for interferometrically measuring the thermal conductivity of AFB doublets. In such a scheme, the optical path difference (OPD) is measured by a Zygo phase-shifting interferometer at thermal equilibrium. A similar technique has been used to measure the thermal conductivity of novel magneto-optical materials [4]. In our case, the AFB’d composite is firmly clamped between a heater with fixed heat load (1.24W) and a water-cooled heat sink, which is maintained at 18° C. The temperature of the heater end is monitored by two thermocouples. In order to prevent the heat convection with ambient atmosphere, the cryostat is evacuated before operation. In the measurement, the heater contacts with the rare-earth ion doped YAG, and heat sink contacts with un-doped YAG.
Crystal temperature increase will cause thermal expansion and refractive index change [5, 6]. The overall effect is observed by the optical path change of the crystal. Transmitted wavefront data was measured and recorded with the Zygo interferometer. The measurement for each sample is done twice at 1) room temperature and 2) thermal equilibrium with external thermal load. The optical path difference data along the specified straight line is recorded for further data analysis.
Consider a rectangular AFB composite as shown in Fig. 3 that consists of un-doped and rare-earth doped YAG crystals with one end at a uniform temperature T and the opposite end at uniform temperature T + ΔT. At steady-state, the temperature gradient between these two ends will be linear and can be written as:
where, Q is the heat flow power; κ1 and κ2 are the thermal conductivities of the un-doped YAG and the rare-earth doped YAG, respectively; H is the heat transfer coefficient of the AFB interface; S is the crystal cross section and l1 and l2 are the crystal lengths of the two crystals, respectively.
Fig. 3 Schematic illustration of the un-doped YAG and rare-earth doped YAG composite and the thermal expansion (dashed lines) caused by a uniform temperature gradient.
Knowing the change of temperature at different positions along the x-axis, one can express the length and refractive index, which are the two essential parameters to calculate the optical path difference (OPD):
The Zygo Verifire interferometer measures OPD (x) = d (x) × n (x):
If one measures OPD1 (x) without heating the other end, i.e. ΔT(x) = 0, and OPD2 (x) with certain known heating load Q and temperature T0 + ΔT, one can then subtract OPD1 (x) from OPD2 (x), to find ΔOPD (x) if we neglect higher order terms of ΔT(x):
where C0 = n0•α + dn/dt = 2.29 × 10−5 K−1.From the above equations, one can see that with knowing the thermal conductivity of the un-doped YAG crystal κ1, and the temperature distribution inside the composite via OPD measurement, the thermal conductivity of the rare earth doped YAG and the heat transfer coefficient at the AFB interface can be characterized.
To determine the thermal conductivity of the rare earth doped YAG, a simple method is to compare the slope of ΔOPD(x) between (0,l1) (η1) and (l1,l2) (η2). Expressions for η1 and η2 are derived:
Therefore, a simple relation is observed:
If we use a reference thermal conductivity value of κ1 = 11.2 W/m*K [3] for un-doped YAG, one will be able to find out the thermal conductivity of its counterpart.
To determine the heat transfer coefficient H, one can investigate the slight temperature bump at x = l1, i.e. the bonding interface, derived from ΔOPD(x). It is given that, at x = l1, the temperature bump Tb equals to:
Therefore, heat transfer coefficient H is calculated as:
Figure 4 shows the temperature distribution measured inside the 3% Er:YAG/YAG composite at heater power Q of 1.24 W, as an example. In the figure, we compare the temperature at different locations of the composite to the temperature at the bonding interface. 1 pixel = 125 μm. One can see that the temperature is linearly distributed in the 3% Er-doped and un-doped YAG crystals with different gradients. The 3% Er:YAG has a larger temperature gradient than the un-doped YAG, which indicates a smaller thermal conductivity. The ratio of the thermal conductivities of the 3% Er:YAG to un-doped YAG, which can be determined by the inverse ratio of the temperature gradients in the two crystals, is (−0.137)/(−0.180) = 0.761. Since a value of κ1 = 11.2 W/m°C for the un-doped YAG has been taken as our reference, the thermal conductivity of the 3% Er:YAG is then determined to be around 8.52 W/m*K.
A very slight temperature discontinuity Tb is observed at the AFB interface of the 3% Er:YAG and un-doped YAG. Since the temperature drop at the AFB interface is so small that it is close to the accuracy of our interferometer, more experiments would be required for further characterizing the thermal properties of the AFB interfaces. As an example, the measured Tb for 3% Er:YAG is too small to arrive at a reasonable estimate of heat transfer coefficient but it is estimated to be ≤ 106 W/m2*K. Actually, experimental results have demonstrated that the AFB bonding interfaces have negligible heat transfer resistance even for bonding materials that have quite different thermal conductivities [6].
3. Applications of AFB composites
Onyx Optics has produced a variety of customer-designed composite laser components since 1992. They include rods with undoped ends; disks with undoped YAG, sesqioxides or sapphire; slabs with undoped ends or doped or undoped layers. We discuss here more recent developments at Onyx Optics of AFB components that are crystalline fiber waveguides and stacks of walk-off compensated nonlinear crystals for frequency conversions.
3.1 Crystalline fiber waveguides (CFWs) for high performance laser emission
Diode pumped fiber lasers are well-recognized for their high beam quality, high gain and high efficiency. For Yb-doped glass fiber amplifiers, near 90% of quantum efficiency has been achieved [7, 8]. Compared to diode pumped solid state lasers, fiber lasers confine both pump and lasing radiation in the fiber for long distance, resulting in high output power. Over 10kW continuous-wave (CW) single-mode laser output has been reported in a Yb-doped glass fiber master oscillator power amplifier (MOPA) system [9, 10]. Nevertheless, in pulsed lasers, due to low Stimulated Brillouin Scattering (SBS) threshold and low thermal conductivity in conventional glass fibers, power scaling is indeed a challenging task. YAG crystal has one order of magnitude higher thermal conductivity (11.2 W/m-°C) than silica fiber (1.38 W/m-°C), and one to four orders of magnitude lower SBS gain coefficient (10−15 to 10−12 m/W in YAG, 10−11 m/W in silica fiber). With such preferred properties, parasitic nonlinear effects are suppressed and thus allow high output laser operation, especially for pulsed lasers. In general, rare-earth ion doped YAG crystalline fibers are considered suitable for power scaling of both CW and pulsed lasers.
Onyx Optics has developed a methodology and process of fabricating true single-clad and double-clad crystalline fiber waveguides (CFWs) from high optical quality YAG and other crystalline materials by employing Onyx’s patented Adhesive-Free-Bonding (AFB) technology [11], where the expression “true” indicates that all materials of CFWs consisted of are crystals, and refers to the ability of controlling laser properties of CFWs in analogy to glass fibers for lasers by designing optical and physical properties of core and inner cladding, and predicting the lasing properties for single mode performance and cladding pumping. In single-clad CFWs, the rare-earth (RE) ion doped core is surrounded by one material, which could be either un-doped YAG or laser-inactive-ion-doped YAG; in double-clad CFWs, the core is surrounded by an inner cladding, which could be either un-doped YAG or laser-inactive-ion-doped YAG, and an outer cladding which could either be sapphire or spinel, see Fig. 5. These waveguides, whose inner cladding material can be precisely engineered, can provide intrinsic single mode for lasing wavelength. We consider double-clad CFWs superior to single-clad CFWs, for the reasons that 1) Inner-cladding pumping can be implemented, resulting in more pump absorption due to tight confinement of pump beam by the lower-index outer cladding, and 2) outer cladding may have higher thermal conductivity (35.5 W/m-°C for sapphire) for better heat dissipation and removal. Outstanding performance of double-clad CFWs has been demonstrated. Mu. et al have reported high gain properties of a double-clad CFW, a 45 times (or 16.5 dB) signal gain in the un-coated and un-cooled CFW before the CFW starts self-lasing [12].
Waveguide mode distribution in a square core has been studied extensively, and solved by finite-element analysis and other approaches [13–15]. Single-mode would be achieved when the following condition is fulfilled:
where d is the core width, λ is the lasing wavelength, and ncore and nclad are refractive index of core and inner cladding, respectively. From the above expression, we find out that, for a given core width and material, we may carefully select the inner cladding material, with proper dopant and doping concentration, to render the B-number smaller than 1.37. We therefore pick the following core/inner cladding pairs, all with intrinsic single-mode property, in Table 1 [16]:
Table 1. Examples of single crystal core/inner cladding materials and core width to achieve intrinsic single-mode output.
The values of refractive index difference Δn is calculated based on the index difference/concentration slope we have measured previously using an interferometric method [17], which has resolution of 10−6. The actual refractive index difference between core and cladding material needs to be confirmed before fabricating specific CFW designs.
In the design of a double-clad CFW, the inner cladding thickness is critical in maintaining both single-mode and high power output. If the inner cladding thickness is too thin, it may not be sufficient to hold a single mode output since the outer cladding, sapphire, has a much lower refractive index and will affect the electric field distribution and lead to multimode output. Consider the extreme case that the inner cladding vanishes. In that way, the core is surrounded directly by sapphire and the intrinsic output is certainly multimode. However, if the inner cladding is too thick, the portion of absorbed pump power will decrease if employing a cladding pumping scheme. Therefore, an optimum inner cladding thickness exists and should be determined to predict the design of a double-clad CFW, with single-mode output. It can be calculated by finding out the mode confinement factor of a double-clad CFW with varied inner cladding thickness, and compare the value with a two-material system which only consists of the core and inner cladding.
Power scalability of YAG crystalline fibers offer the prospect of improved properties over glass fibers. Double-clad crystalline fiber waveguides can employ cladding material with engineer-able refractive index difference to core material, by varying the doping concentration of laser-inactive rare-earth trivalent ions. This design flexibility would enable intrinsic single-mode beam propagation while achieving large mode area in the fiber. A wide range of doping concentrations of the core and compatible laser inactive inner cladding is available and so is the prospect of single mode output with engineered core/ cladding dimensions and combinations since YAG-based CFWs are assumed to be straight and thereby also avoid bending losses at small refractive index differences.
J. W. Dawson et al. have developed models for evaluating hard upper limits of output power from a fiber waveguide, for silica fibers [18] and some non-silica materials [19], with circular cross sections. We applied the approach to YAG CFWs, with Yb3+,Er3+,Tm3+ and Ho3+:YAG as core, either un-doped or laser-inactive-ion-doped YAG as inner cladding, and sapphire as outer cladding, and subsequently have determined the upper limit of output power for intrinsic single mode circular double-clad CFWs. We have determined that the hard upper limit of output power for double-clad CFWs relies heavily on the physical constants of material. The results presented here should be essentially independent of fabrication technique, assuming perfect core/inner cladding/outer cladding interfaces. Similar considerations hold for square core and inner cladding cross sections, which in fact can be fabricated by Onyx’s AFB technology.
Several limitations that will affect the upper output power are listed in Ref [19], such as 1) thermal limitations, including thermal fracture, melting of the core, and thermal lens effect; 2) nonlinear phenomena limitations, including stimulated Raman scattering (SRS) and stimulated Brillouin scattering (SBS); 3) laser damage, and 4) pump power limitations. For a given fiber system, any of these limitations may negate further `````power scaling. For YAG-based CFWs, it is found that thermal lens effect, SBS, and pump power brightness would impose a boundary to the hard upper limit of possible output power:
where, ηlaser is the laser efficiency, ηheat is the heat efficiency, κ is the core thermal conductivity, λ is the lasing wavelength, dn/dT is the thermal-optical coefficient, d is the core width, and L is the fiber length.
where, Γ is the mode size to core width ratio, gB(Δν) is the Brillouin gain coefficient, and G is the laser gain.
where, Ipump is the pump diode brightness, NA is the numerical aperture of the fiber, αcore is the small-signal core absorption coefficient, and A is the small signal pump absorption. By using values of the parameters listed in Table 2, and fixing the output power, one is able to obtain the contour plots of L and d that are generated and analyzed for different CFW configurations.
Table 2. [20]. Parameters and physical constants of RE:YAG CFWs.
Figures 6(a)-6(d) illustrate contour plots in lower power regime for the four CFWs’ configurations, respectively [20].
Fig. 6 Contour plot of scalability of RE:YAG fiber lasers with core diameter of circular cross section in low power regime with core material of a) 2.5% ceramic Yb:YAG, b) 4% Tm:YAG, c) 0.5% Er:YAG, and d) 2% Ho:YAG.
The L-d plots generally state that, for a fiber with specific core diameter d0 and fiber length L0 on a certain contour line, the maximum output power will be the value marked on that line. As an example, in Fig. 6(a), with d = 400μm and L = 0.35m, the fiber can reach at most 100W output power and no more. We notice that an optimum core diameter exists for each fiber structure. In Fig. 6(a), for any given output power requirement, the smallest CFW diameter always occurs at a core diameter of d = 69.81μm. This value is actually obtained by equating the RHS of the two inequities of factor 1) thermal lens effect, and factor 3) pump power brightness. For different output power contour plots, the optimum core diameter remains the same. If the CFW diameter is larger than the corresponding minimal values, there would exist a pair of allowed core diameters, both of which will meet the output power requirement. For instance, at P = 4kW, and L = 0.68m, the allowed core diameter d can be either 50-μm or 90-μm. Large core size favors output power at certain diode pump brightness, but the enhanced thermal lens effect will suppress the output power at some point if the core diameter is too large. Therefore, larger core size does not always lead to higher output power. However, smaller refractive index difference between core and cladding will yield larger mode diameter to core diameter ratio (Γ), which would allow higher upper limit to output power, at the price of more pump power since the mode confinement is weaker. The reason why large Γ value could result in high output power is because, compared with a larger refractive index difference, smaller index difference will result in larger mode diameter, i.e. smoother/flatter transverse mode distribution in the core region, or in other words, the electric field strength at the core/cladding interfaces will attenuate less. This on average will give more output power in the core region at the lasing wavelength assuming sufficient pump power is supplied.
Currently, CFWs with total length of > 0.2 - 0.5m are difficult to fabricate with constant refractive index difference. A concept for overcoming this limitation has been identified and is actively investigated to enable scaling to high output power. Considering the present manufacturing limit for CFW length to be, e.g., 0.5 m, the actual maximum output power will be limited to ~4.4kW for 2.5% ceramic YAG/Undoped YAG double-clad CFW, as an example. The latter two CFWs (Er-doped and Ho-doped) have much higher output power upper limits due to the high Γ value of its core/cladding combination. It is worth noting that the thermal conductivity for 2.5% Yb:YAG is used in the calculation while the thermal conductivity of inner cladding did not come into consideration in the simulation. Intuitively, the larger the thermal conductivity of inner cladding and outer cladding, the better the thermal management will be and the weaker the thermal lens effect. In the simulation, the thermal conductivity value does not seem to affect the output power much since we assume sufficient residue heat removal, but different choices of inner cladding materials are available for different core dopants. A future model will take into account the heat transfer coefficient between the final cladding and whatever cooling medium is used since the beam quality and output power in the waveguide will be critically dependent on removal of heat at the coolant/outer cladding interface [21].
Actual laser experiments have also been performed at Onyx Optics with double-clad Yb:YAG CFWs [22]. The double-clad CFW used in the experiment consists of a single crystal 1at.% Yb:YAG square waveguide core with un-doped YAG inner cladding, and ceramic spinel outer cladding . It was made from a single crystal Yb:YAG slab followed by multiple precision polishing and crystal bonding processes for the final core size as well as the formation of the inner and outer claddings. The measured cross sections for the waveguide core and the inner cladding are around 40 μm × 40 μm and 102 μm × 102 μm, respectively. The CFW length is 65mm.
The laser experimental setup is schematically illustrated in Fig. 7. In the experiment, we have found that when the pump power is increased to above 2.7 W, strong laser output at 1030 nm is achieved even without the attached laser mirrors. This demonstrates that the double clad CFW has a much larger gain coefficient (gl > 5) than in any bulk lasers. The laser beam in this case is coupled out by the two uncoated ends (R ~8%) of the CFW in both forward and backward directions with identical output powers as shown in Fig. 8. The total laser power in the two directions is 8.5 W at an input pump power of 39.5 W, which corresponds to an optical-to-optical efficiency of 21.5%. The measured slope efficiency from Fig. 8 is 22.6%. The laser wavelength is confirmed at 1030 nm and the spectral line width (full width at half maximum) is 4 nm with a 1/8 meter monochromator (Spectral Product, CM110). After attaching input and output laser mirrors on the two waveguide ends, the laser threshold is reduced to below 0.54 W. The maximum output power is increased to 13.2 W at pump power of 39.5 W. Due to the low laser threshold, the laser optical to optical efficiency and the slope efficiency are almost same. They are both around 33.4%.
Fig. 7 Experimental layout of the cladding pumped CFW and laser output profile. f1 and f2, aspheric lenses; BS, Beam splitter.
Fig. 8 Laser output power as a function of pump power. The upward and downward open triangles and the solid squares are the measured forward, backward and total laser power from the uncoated CFW ends, respectively. The solid circles are the laser power after input and output mirrors attached.
The near-field laser beam profile is measured in Fig. 9 by a pyroelectric camera (Spiricon Pyrocam III) at the image plane of a 10 × microscope objective. The inserted image of Fig. 9 is the measured 2-D beam profile. Even though the waveguide has a square core, the waveguide mode is still round with a measured roundness of about 94%. The solid curve in Fig. 9 is the normalized beam intensity along one axis of the waveguide end. It is a near-perfect Gaussian distribution with a measured beam diameter of 44 μm at the position of 1/e2 of the peak intensity. The beam quality is analyzed by the knife-edge method [23]. Figure 10 shows the measured beam radius as a function of the test position after a 200-mm focus lens. The fitted M2 from Fig. 10 is ~1.02, which demonstrates that a purely single-mode laser has been achieved. The dashed curve in Fig. 9 is the calculated mode profile of the fundamental mode by Marcatili’s approximation [13]. In the calculation, we use a waveguide core size d = 40 μm and refractive index step Δn = 1.08 × 10−4. The calculated beam diameter is 43.6 μm. One can see that the measured beam profile overlaps with the calculated fundamental mode near-perfectly, which further demonstrates that the laser output is purely single-mode.
Fig. 9 (Left) Measured (solid curve) and simulated (dashed curve) beam profiles of the doubled-clad CFW laser. The inserted image is the 2-D beam profile measured by a pyroelectric camera.
To conclude this Section, we have described the structural configuration of single-clad and double-clad CFW, analyzed designs for single-mode operation, and studied power scalability by simulation of upper limit of output power using Dawson’s model. We then demonstrated laser diode cladding pumped single-mode laser in a double-clad Yb:YAG CFW with near diffraction limited beam quality. Laser output power of 13.2 W has been obtained at pump power of 39.5 W. Our measurement and analysis indicate that the laser efficiency can be further increased by improving the coupling and absorption efficiencies of the pump beam as well as the NA and the end perpendicularity of the waveguide. The laser experiment has also demonstrated that single-mode lasing is achievable in multimode CFWs when the waveguides only contain low order modes. Potentially, the adhesive-free bonded CFWs can be used to improve the power scalability of current solid-state and fiber lasers.
3.2 Walk-off compensated nonlinear crystal stack
The intensity distribution of a laser beam that travels in anisotropic, or birefringent crystals, can drift away from the direction defined by the wave vector. This phenomenon is known as “spatial walk-off” [24]. Spatial walk-off occurs only for a beam with extraordinary polarization, propagating at an angle θ against the optical axis. Most of the nonlinear optical crystals are anisotropic such that Type I or Type II phase-matching conditions can be met. However, the extraordinary beam will spatially “walk away” from the ordinary beam, which strictly follows the wave vector direction. Therefore, the spatial overlap of extraordinary beam and ordinary beam will decrease after certain propagating distance, resulting in low conversion efficiency and unsatisfactory beam quality.
In order to overcome the restrictions of spatial walk-off existing in nonlinear conversion, Onyx Optics developed walk-off compensate nonlinear crystal stacks with AFB technique. The walk-off effect of the extraordinary beam propagating in such a stack is minimized, such that enhanced nonlinear conversion efficiency and beam quality can be anticipated.
Figure 11 shows a typical configuration of a walk-off (WoC) compensated nonlinear crystal stack. The material is KTiOPO4 (KTP).
The WoC-KTP stack consists of N single layers of KTP, where N is an even number. Each KTP layer is cut at the phase matching angle for a specific nonlinear conversion process, i.e. θ = 90°, ψ = 26° for 1064-nm (e) + 1064-nm (o) → 532-nm (e). Then, each single layer is aligned in a head-to-head or tail-to-tail configuration to form a stack, such that the extraordinary beam spatially “walks back” in adjacent layers to overlap with the ordinary beam, while keeping the sign of nonlinear coefficient d unchanged. AFB technique is employed to bond the N layers together as an entire composite. Although KTP has anisotropy in the thermal expansion coefficients, with such bonding configuration, each bonding layer experiences the same thermal expansion coefficient, such that no residue stress in the device will be present after bonding. Please note that the bonding configurations differ from one type of nonlinear crystal to another, since each type of nonlinear crystal has its own nonlinear tensor, and the sign of nonlinear coefficient is determined based on that, which needs to be analyzed individually. The advantages of WoC nonlinear crystal stacks are: 1) ease of use; 2) elimination of surfaces loss as opposed to conventional walk-off compensate approach; 3) increased effective crystal length; 4) reduced walk-off effects; and 5) insensitivity of slight misalignment.
As a practical example of a WoC NLO stack, J. Zondy et al. [25], K. Hara et al. [26] and D. J. Armstrong et al. [27] reported theoretical and experimental results of efficient second-harmonic generation in WoC nonlinear crystal stacks. At Onyx, we built a 2-μm optical parametric oscillator (OPO) employing the composite [28]. Due to the large spatial walk-off in 1.064-μm pumped 2-μm KTP optical parametric oscillations (OPO’s), large-scale and high energy lasers from several millijoules to hundreds of millijoules are usually required as pump sources. In order to reduce the pump requirement in the OPO operations, Onyx employed the KTP composite mentioned above. The experimental setup is shown below in Fig. 12:
In order to demonstrate those concepts, two 16-layer bonded WOC samples were prepared for the QNCPM and QPM OPO experiments. The thickness of the individual KTP crystals used in the two samples is 2 mm. The corresponding walk-off angle is 47.5 mrad.
The OPO threshold is measured to be 254 μJ per pulse, corresponding to 44.6MW/cm2. The overall efficiency reached ~9.3% with a slope efficiency of 27%. It is worth for us to mention that the conversion efficiencies measured in our experiment are quite similar as the result measured in a 20-mm long single KTP OPO in Ref [29], in which the OPO threshold is reported about 320 mJ/pulse (~151 MW/cm2) and a maximum pump energy of 636 mJ/pulse was used. It is demonstrated that with a WoC KTP stack, the pump threshold is greatly reduced.
The beam profile of 2-μm output is shown in Fig. 13:
The beam profile is a Gaussian-like distribution. No distortion caused by spatial walk-off is observed.
In conclusion, high optical quality WOC KTP composites have been developed by using AFB technology, and employed in an OPO system to render low threshold, high beam quality and enhanced angular acceptance. The newly developed technology can also be used in other nonlinear materials and devices, such as LBO, BBO, or ZGP, to improve the beam quality and conversion efficiency for intended wavelengths.
5. Conclusion
In this paper, we have reviewed the thermal properties and power scaling capability of AFB composites. Two major recent applications of AFB composites, crystalline fiber waveguides (CFWs) and walk-off compensated nonlinear stacks, have been introduced. AFB is a promising technology for fabricating novel optical components for high power, high beam quality laser emissions at various wavelengths.
Funding
Wright-Patterson Army Force Base Research Laboratory.
References and links
1. M. De Vido, D. Meissner, S. Meissner, K. Ertel, P. J. Phillips, P. D. Mason, S. Banerjee, T. J. Butcher, J. M. Smith, C. Edwards, C. Hernandez-Gomez, and J. L. Collier, “Characterization of adhesive-free bonded crystalline Yb:YAG for high energy laser applications,” Opt. Mater. Express 7(2), 425 (2017). [CrossRef]
3. R. L. Aggarwal, D. J. Ripin, J. R. Ochoa, and T. Y. Fan, “Measurement of thermal–optics properties of Y3Al5O12, Lu3Al5O12, YAlO3, LiYF4, LiLuF4, BaY2F8, KGd(WO4)2, and KY(WO4)2 laser crystals in the 80-300K temperature range,” J. Appl. Phys. 98(10), 103514 (2005). [CrossRef]
4. I. Kuznetsov, I. Mukhin, D. Silin, and O. Palashov, “Thermal conductivity measurements using phase-shifting interferometry,” Opt. Mater. Express 4(10), 2204 (2014). [CrossRef]
5. D. Li, P. Hong, M. Vedula, and H. E. Meissner, “Thermal conductivity investigation of adhesive-free bond laser components,” Proc. SPIE 10100, Optical Components and Materials, 1010011 (2017).
6. X. Mu, H.-C. Lee, and H. E. Meissner, “Investigation of Interface Heat Transfer in Adhesive Free Bond Laser Composites,” Proc. DEPS SSDLTR (Broomfield, 2010).
7. L. Goldberg, J. P. Koplow, and D. A. V. Kliner, “Highly efficient 4-W Yb-doped fiber amplifier pumped by a broad-stripe laser diode,” Opt. Lett. 24(10), 673–675 (1999). [CrossRef] [PubMed]
8. P. Wang, L. J. Cooper, J. K. Sahu, and W. A. Clarkson, “Efficient single-mode operation of a cladding-pumped ytterbium-doped helical-core fiber laser,” Opt. Lett. 31(2), 226–228 (2006). [CrossRef] [PubMed]
9. IPG product/application press releases 2009. http://www.ipgphotonics.com/Collateral/Documents/English-US/PR_Final_10kW_SM_laser.pdf.
10. Y. Kalisky and O. Kalisky, “The status of high-power lasers and their applications in the battlefield,” Opt. Eng. 49(9), 091003 (2010). [CrossRef]
11. X. Mu, H. E. Meissner, H.-C. Lee, and M. Dubinskii, “True Crystalline Fibers: Double-Clad LMA Design Concept of Tm:YAG-Core Fiber and Mode Simulation,” Proc. SPIE 8237, 82373M (2012). [CrossRef]
12. X. Mu, S. K. Meissner, H. E. Meissner, and A. W. Yu, “Double clad YAG crystalline fiber waveguides for diode pumped high power lasing,” in Solid State Lasers XXIII: Technology and Devices, W. A. Clarkson; R. K. Shori, Editors, Proc. SPIE 8959, 895906 (SPIE, Bellingham, WA 2014).
13. K. Okamoto, Fundamentals of Optical Waveguides, 2nd Ed. (Academic Press, 2006).
14. K. S. Chiang, “Finite-Element Analysis of Optical Fibers with Iterative Treatment of the Infinite 2-D Space,” Opt. Quantum Electron. 17(6), 381–391 (1985). [CrossRef]
15. K. S. Chiang, “Dual effective-index method for the analysis of rectangular dielectric waveguides,” Appl. Opt. 25(13), 2169 (1986). [CrossRef] [PubMed]
16. D. Li, P. Hong, S. K. Meissner, and H. E. Meissner, “Design of intrinsically single-mode double clad crystalline fiber waveguides for high power lasers,” Proc. SPIE 9744, Optical Components and Materials, 97441H (2016).
17. H.-C. Lee, X. Mu, and H. E. Meissner, “Interferometric measurement of refractive index difference applied to composite waveguide lasers,” Proc. CLEO2011, AMB4, (2011). [CrossRef]
18. J. W. Dawson, M. J. Messerly, R. J. Beach, M. Y. Shverdin, E. A. Stappaerts, A. K. Sridharan, P. H. Pax, J. E. Heebner, C. W. Siders, and C. P. J. Barty, “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Opt. Express 16(17), 13240–13266 (2008). [CrossRef] [PubMed]
19. J. W. Dawson, M. J. Messerly, J. E. Heebner, P. A. Pax, A. K. Sridharan, A. L. Bullington, R. J. Beach, C. W. Siders, C. P. Barty, and M. Dubinskii, “Power scaling analysis of fiber lasers and amplifiers based on nonsilica materials,” Proc. SPIE 7686, 768611 (2010). [CrossRef]
20. D. Li, P. Hong, S. K. Meissner, and H. E. Meissner, “Power scaling estimate of crystalline fiber waveguides with rare earth doped YAG cores,” Proc. SPIE 9744, Optical Components and Materials, 97441I (2016).
21. D. C. Brown and H. J. Hoffman, “Thermal, stress, and thermo-optic effects in high average power double-clad silica fiber lasers,” IEEE J. Quantum Electron. 37(2), 207–217 (2001). [CrossRef]
22. X. Mu, S. Meissner, H. Meissner, and A. W. Yu, “High efficiency Yb:YAG crystalline fiber-waveguide lasers,” Opt. Lett. 39(21), 6331–6334 (2014). [CrossRef] [PubMed]
23. J. M. Khosrofian and B. A. Garetz, “Measurement of a Gaussian laser beam diameter through the direct inversion of knife-edge data,” Appl. Opt. 22(21), 3406 (1983). [CrossRef] [PubMed]
24. https://www.rp-photonics.com/spatial_walk_off.html?s=ak.
25. J. J. Zondy, D. Kolker, C. Bonnin, and D. Lupinski, “Second-harmonic generation with monolithic walk-off-compensating periodica structures. II. Experiments,” J. Opt. Soc. Am. B 20(8), 1695 (2003). [CrossRef]
26. K. Hara, S. Matsumoto, T. Onda, W. Nagashima, and I. Shoji, “Efficient Ultraviolet Second-Harmonic Generation from a Walk-Off-Compensating β-BaB2O4 Device with a New Structure Fabricated by Room-Temperature Bonding,” Appl. Phys. Express 5(5), 052201 (2012). [CrossRef]
27. D. J. Armstrong, W. J. Alford, T. D. Raymond, A. V. Smith, and M. S. Bowers, “Parametric amplification and oscillation with walkoff-compensating crystals,” J. Opt. Soc. Am. B 14(2), 460–474 (1997). [CrossRef]
28. X. Mu, H. Meissner, and H.-C. Lee, “Optical parametric oscillations of 2 microm in multiple-layer bonded walk-off compensated KTP stacks,” Opt. Lett. 35(3), 387–389 (2010). [CrossRef] [PubMed]
29. S. Haidar, K. Miyamoto, and H. Ito, “Generation of tunable mid-IR (5.5-9.3 μm) from a 2-μm pumped ZnGeP2 optical parametric oscillator,” Opt. Commun. 241(1-3), 173–178 (2004). [CrossRef]
