A novel dual channel Tm:YLF laser system was developed where two degenerate laser cavities were coupled by spectrally beam combining their emission and by implementing a common output coupler. Under continuous wave running conditions, each channel’s slope efficiency was greater than 45% and the maximum combined output power was 11 W. Passive Q-switching was achieved using an 80%, Cr:ZnSe saturable absorber. The output pulses had a maximum energy of 5.8 mJ and duration of 90 ns (~65 kW of peak power) at 5.7 W of absorbed pump power. Each channel showed less than 1 nm of spectral width with central wavelengths around 1880 nm and 1908 nm correspondingly. The system had adjustable spectral difference between the channels ranging from 5 to 20 nm which corresponds to 0.4 – 1.7 THz if the system is used for nonlinear difference frequency generation.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Dual wavelength laser systems have been an area of research and innovation that has attracted much attention over the years. The ability to produce multiple lasing channels from a single system brings compactness, minimizes complexity and usually makes the system more cost effective. Such systems have found applications in the areas of nonlinear conversions, environmental monitoring, and optical communications . In the recent years, dual frequency lasers have been used to produce THz waves by nonlinear difference frequency generation (DFG) . THz waves, due to their small scattering and low photon energy, can pass through many materials and leave living organisms unaffected, making it ideal for medical imaging . As a result, there is a great desire to develop a compact, durable, and tunable high power THz source .
The first dual-wavelength laser source was demonstrated in , Betha using a Nd:YAG crystal to produce two channels centered at 1.06 and 1.318 μm. Nd3+ has been a popular medium for producing dual channel systems due to its multi-peak emission spectrum and high gain properties to meet threshold conditions [4, 5]. However, other dual wavelength systems have been shown using Er3+, Ho3+, and Tm3+ doped materials . Main approaches to create dual laser systems are through introducing wavelength selectors, such as prisms , Fabry-Perot etalons , and specialized dielectric mirror coatings . The main drawback of implementing these elements is the insertion loss which will degrade laser performance. Other issues to consider are the selector’s damage threshold, and the unwanted crosstalk.
Another promising and flexible approach for building high power multiwavelength laser systems that are also narrowband, tunable, and gain independent, is spectral beam combining using volume Bragg gratings (VBGs). VBGs are ideal beam combining optical elements because they can be optimized to have 99.9% diffraction efficiencies, have less than 1% losses, and a bandwidth could be narrowed down to 10 pm [8, 9]. The first demonstration of spectral beam combining by a VBG was described in , Ciapurin et al. The use of two sequential VBGs as output couplers in a cavity of a semiconductor laser resulted in stable dual wavelength emission . The further development of dual wavelength laser systems was produced by spectral combining of two amplifiers in a common resonator by means of two VBGs tuned to different wavelengths [12, 13]. The later scheme allowed automatic control of the emitting wavelengths and provided collinearity of the emitted beams. Both approaches demonstrated high efficiency and no significant degradation of the beam quality.
In this work, we use similar beam combining approach and demonstrate a narrowband dual wavelength system with tunable capabilities based on Tm:YLF laser medium, wavelength selective volume Bragg gratings, and a common output coupler in a two-channel resonator. The performance of the system in terms of output power levels, pulse energies, and spectral properties are presented and discussed for both continuous wave (CW) and passive Q-switched (PQS) operation.
2. Dual channel laser cavity design
A schematic of the dual channel laser system is presented in Fig. 1. Each channel consist of a 3 mm in diameter, 10 mm long Tm:YLF crystal with a doping concentration of 3% and with the c-axis cut parallel to the rod axis. To reduce thermal lensing effects and obtain high output powers, the crystals were enclosed in indium foil and held inside a water cooled copper mount. Each crystal was pumped by a high brightness fiber coupled 793 nm laser diode module. The pump modules are capable of delivering 30 W of continuous optical power through a 105 μm core diameter fiber with a 0.22 numerical aperture. End pumping with aspheric lenses provided a pumped area diameter of ~320 µm.
Two W-shaped individual amplifiers were mutually coupled and converted to a dual wavelength laser by two VBGs and a common output coupler. The total optical path length for each channel was approximately 70 cm. Each channel’s pump beam was directed through a planar dichroic mirror (M1 and M2) for longitudinal pumping of the crystals. Two spherical mirrors (M3 and M4) were spaced two focal lengths away and aligned to the dichroic mirrors in order to increase beam diameters on any optical component. This configuration avoids overloading the antireflection coatings’ energy density damage threshold during PQS operation. Two additional spherical mirrors (M5 and M6) were spaced one focal length away and tilted to collimate and direct the crystals’ emission onto their corresponding VBGs. The role of the VBGs was to guarantee that only the specified fraction of the emission spectrum that met the Bragg condition, based on incident angle, was diffracted with high efficiency towards the planar dielectric output coupler providing narrow band feedback. As it is seen in Fig. 1, the combining of the two beams is done by the second channel’s VBG. The first channel, after diffracting from VBG1, passes through VBG2 before going to the output coupler. This channel does not undergo Bragg diffraction because VBG2’s grating period is offset to diffract near 1905.5 nm at this incident angle. Since VBG1 diffracts only 1887.5 nm, the light does not interact with the Bragg grating structure, and passes through like a normal plate of glass. If channel 1 was the same wavelength at channel 2, the beam would diffract instead of passing through VBG2 resulting in no beam combination. Thus the VBGs determined not only each channel’s wavelengths depending on their tilt angles but one of them served as a wavelength combining element. In addition, this VBG combing approach and the common output coupler provided collinearity of both beams.
The volume Bragg gratings used had a spectral width < 1 nm and a diffraction efficiency in the spectral range from 1890 to 1908 nm more than 99%. They were antireflection coated to reduce any insertion loses. After implementation into the cavity, the gratings preformed as efficiently as high reflective planar mirrors, but also behaved as spectral locking elements providing feedback for specified for each channel wavelengths only.
3. Experimental Results
3.1 Continuous wave operation
Figure 2 shows the CW output power dependence from absorbed pump power for the individual and simultaneous channel operation using 30% output coupling. For one channel, threshold for lasing was 0.6 W of absorbed pump power and the slope efficiency was 50%. Maximum output power measured was 5.7 W which is comparable to similar systems previously developed . Implementing the second lasing channel resulted in a threshold of 0.7 W, and a 46% slope efficiency. The output power during simultaneous operation of both channels resulted in a summation of each channels output power with no gain competition, achieving a maximum of 11 W. In order to ensure exact co-propagation and overlap, we maximized the dual channel output power through two apertures. One aperture was placed close to the output coupler and the other was placed at ~2 meters from the first one. Additional indication of the proper spatial overlap was the change of the repetition rate during Q-switching operation which increased when both channels were properly overlapped. This is due to both channels saturating the same area on the saturable absorber. Since both channels are aligned to the same output coupler the beams were collinear and also spatially overlapped at multiple locations. Each individual channels’ beam quality parameter M2 was measured, to be Mx2 ~3.2 and 2.2 and My2 ~2.1 and 2.4 respectively, which changed during dual operation to an Mx2 ~3.6 and My2 ~2.3. We used the relative minimization of the combined M2 value as another parameter which ensures proper overlap of the two beams. Since our main focus was to extract as much as possible pulse energies, high M2 numbers were expected because large mode volumes had to be used in order to avoid any optical damage.
Figure 3 shows the measured reflection spectra for a normal incidence angle of the VBGs used in the system. VBG1 has central wavelength for normal incidence of 1892.4 nm and spectral width of 1.7 nm (FWHM). VBG2 has central wavelength for normal incidence of 1907.8 nm and spectral width of 1.5 nm. Both VBGs showed a reflectance of exceeding 99.9%. Emission spectra of a cavity with the VBG2 and with the grating replaced by a broadband dielectric mirror are shown in Fig. 4. It is well seen that comparing to a broadband mirror, a narrowband feedback provided by the VBG, results in dramatic decrease of spectral width while keeping comparable output power.
Figure 5 presents the emission spectra, showing center wavelengths of 1887.5 and 1905.5 nm for VBG1 and VBG2 channels correspondingly. Emission wavelengths of laser channels are shifted in short wavelength side because of VBGs inclination in the cavity. Spectral width of VBG1 channel is 0.8 nm, VBG2 – 0.25 nm. Difference in spectral widths is determined by both difference of widths of the gratings and different inclination angles. Both channels were observed to be highly independent of each other, and the power level in each channel was individually tuned. This feature is demonstrated in Fig. 6 where emission spectra are measured when both channels had similar pump powers and when each channel’s pump power was separately reduced by half. As can be seen, the change in output power for each channel is completely determined by the pumping for that channel and does not depend on the pumping level of the other channel. By changing the incident angles on the VBGs the spectral spacing between the channels was tuned from 5 nm to 20 nm (0.4 to 1.7 THz), as shown in Fig. 7. For larger channel wavelength separations, a multiplexing of several VBGs in one glass element could provide easy switching between discreet THz output frequencies . For precise central wavelength positioning, fine wavelength tuning with accuracy about 0.005 nm can be achieved by controlling the temperature of each VBG due to the VBGs thermal dependence of their resonant wavelength (0.02 nm/K).
In terms of polarization properties, both channels had strongly linearly polarized outputs due to the properties of the Tm:YLF medium. For nonlinear THz frequency conversion in crystals, such as GaAs, it is important that both polarizations are aligned to the correct axis of the crystal for proper phase matching . In this work, the polarizations of the two channels were matched by rotating the individual Tm:YLF crystals along their cylindrical axis.
3.2 Passive Q switching
Pulsed lasers, due to their high peak power, are more applicable in nonlinear optical conversion processes. In relation to that, an investigation into Q-switching of the dual channel system was done. PQS with a saturable absorber was chosen because of cost effectiveness and implementation ease. In order to select the appropriate saturable absorber, multiple PQS models from , Degnan , Xiao et al, and , Paschotta were used to calculate the estimated saturable absorber transmittance for kW peak power levels. The material Cr:ZnSe was chosen due to its high absorption cross section in the 1880-1908 nm range resulted in successful application in Tm:YLF systems [20–22]. Based on calculations, a saturable absorber with 60% to 95% transmission is able to produce sufficient pulse peak powers in the kW range, and therefore an 80% transmission was chosen for generating high peak power without damaging the cavity with excessive intracavity energy.
PQS operation was successfully achieved after implementation of the 80% transmittance Cr:ZnSe saturable absorber and a 30% transmittance output coupler. The average output power, shown in Fig. 8, shows a threshold for lasing at 1.8 W of absorbed power and increases linearly with a 31% slope efficiency. Pulse energy, peak power, and repetition rate dependences on pump power are compiled in Fig. 9. These values were measured over an average of 10,000 pulses for each level of pump power using a commercial energy meter from Gentec-EO. At 5.7 W of absorbed pump power the average pulse energy was 4.9 mJ, with a standard deviation of 0.4 mJ and a repetition rate of 225 Hz. The maximum measured pulse energy acquired over that span of 10,000 pulses was 5.8 mJ. The peak-to-peak stability was measured to be 28% with an RMS stability of 2.8%. These are the highest achieved energy levels, to the best of our knowledge, in PQS dual-wavelength Tm:YLF system. The key factor limiting further power scaling is the damage thresholds of the AR optical coatings if the same laser mode areas are used. Optical damage has been observed on the crystals and on the Cr:ZnSe saturable absorber at maximum pump power, therefore further power scaling will require increasing of the mode area.
Figure 10 shows the measured temporal pulse train and temporal pulse profile at 5.7 W of absorbed pump power. The FWHM pulse duration measured consistently near 90 ns at different pump levels, corresponding to a maximum peak power of 65 kW for a 5.8 mJ pulse. This consistency is due to the pulse durations, in PQS systems, being highly dependent on round trip time in the cavity and saturable absorber losses. This consistency in pulse duration is comparable in past published PQS Tm:YLF systems in , Korenfeld et al , Canbaz et al, and , Faro et al. No secondary pulses or parasitic emission were observed. Due the VBGs spectral selectivity being dependent on angular alignment orientation, there was no observed change in each channels spectral bandwidth when transitioning from CW to PQS operation.
A great advantage of the dual cavity PQS system presented here is the automatic time synchronization and perfect time overlap of the two pulses generated. The opening of the saturable absorber positioned in the common part of the degenerate resonator guarantees that the two pulses leave the cavity perfectly synchronized in time. This is essential for any non-linear processes where the two wavelengths contributing to the process need to overlap in space as well as in time. As seen from Fig. 10, we can only observe one common temporal shape at any time. For further verification of temporal overlap, the output pulses were spectral split and showed a 3-4 ns discrepancy in their durations. While observing each beam separately, their individual pulse energies were compared after manipulating each channel’s pump energies. In order for an equal amount of output energy in each channel, a channel 1 to 2 pump energy ratio of 0.9 had to be used. This slight extra amount of pump energy needed for channel 2 is due to the slightly lower efficiency performance than channel 1, which was also observable for the continuous wave output power in Fig. 2.
A successful development of a high power dual wavelength laser system was accomplished by spectral beam combining in a two-channel cavity using two volume Bragg gratings for narrow band feedbacks and a common wide band output coupler. Under continuous wave running conditions, each channel’s slope efficiency was greater than 45% and the maximum combined output power achieved was 11 Watts. The two laser channels showed no cross talk and their power levels were individually controlled. The spectral spacing between the channels was tuned from 5 to 20 nm (0.4 to 1.7 THz) by angular tuning of the VBGs. Wider tuning range can be achieved by introducing additional single or multiplexed VBGs. Passive Q-switching was achieved by using a 80% transmittance Cr:ZnSe saturable absorber and demonstrated output pulses with total energy greater than 4.5 mJ and pulse duration of 90 ns (~65 kW of peak power). With the demonstrated power levels and spectral spacing, this system could work well as a nonlinear THz pump source.
This work was partially supported by the Army Research Office contract W911NF-15-P-0011 and the Directed Energy Professional Society for the Graduate Fellowship.
The authors would like to acknowledge OptiGrate Corp. for fruitful collaboration.
References and links
1. B. M. Walsh, “Dual wavelength lasers,” Laser Phys. 20(3), 622–634 (2010). [CrossRef]
2. P. Siegel, “Terahertz technology,” IEEE Trans. Microw. Theory Tech. 50(3), 910–928 (2002). [CrossRef]
3. C. G. Bethea, “Megawatt power at 1.318 μm in Nd3:YAG and simultaneous oscillation at both 1.06 and 1.318,” IEEE J. Quantum Electron. 9(2), 254 (1973). [CrossRef]
4. B. Xu, Y. Wang, H. Cheng, Z. Cai, and R. Moncorge, “Single- and multi-wavelength laser operation of a diode-pumped Nd:GGG single crystal around 1.33 μm,” Opt. Commun. 345, 111–115 (2015). [CrossRef]
5. B. Xu, Y. Wang, H. Cheng, Z. Lin, and R. Moncorge, “Diode-pumped CW laser operation of a c-cut Nd:YAlO3 crystal on low-gain emission lines around 1.1 μm,” IEEE Photonics J. 7, 1503407 (2015). [CrossRef]
6. M. B. Danailov and I. Y. Milev, “Simultaneous multiwavelength operation of Nd:YAG laser,” Appl. Phys. Lett. 61(7), 746–748 (1992). [CrossRef]
8. L. B. Glebov, “Volume Holographic Elements in a Photo-Thermo-Refractive Glass,” J. Holography Speckle 5, 1–8 (2008).
9. L. Glebov, “High-performance solid-state and fiber lasers controlled by volume Bragg gratings,” Rev. Laser Eng. 41, 684–690 (2013).
10. I. V. Ciapurin, L. B. Glebov, L. N. Glebova, V. I. Smirnov, and E. V. Rotari, “Incoherent combining of 100-W Yb-fiber laser beams by PTR Bragg grating,” Proc. SPIE 4974, 209–219 (2003). [CrossRef]
11. S. A. Zolotovskaya, N. Daghestani, G. B. Venus, L. B. Glebov, V. I. Smirnov, and E. U. Rafailov, “Stable dual-wavelength operation of InGaAs diode lasers with volume Bragg gratings,” Appl. Phys. Lett. 91(17), 171113 (2007). [CrossRef]
12. F. Wang, D. Shen, D. Fan, and Q. Lu, “Widely tunable dual-wavelength operation of a high-power Tm:fiber laser using volume Bragg gratings,” Opt. Lett. 35(14), 2388–2390 (2010). [CrossRef] [PubMed]
13. O. Andrusyak, V. Smirnov, G. Venus, V. Rotar, and L. Glebov, “Spectral combining and coherent coupling of lasers by volume Bragg gratings,” IEEE J. Sel. Top. Quantum Electron. 15(2), 344–353 (2009). [CrossRef]
14. S. So, J. Mackenzie, D. Shepheard, W. Clarkson, J. Betterton, and E. Gorton, “A power-scaling strategy for longitudinally diode- pumped Tm:YLF laser,” Appl. Phys. B 84(3), 389–393 (2006). [CrossRef]
15. D. Ott, I. Divliansky, B. Anderson, G. Venus, and L. Glebov, “Scaling the spectral beam combining channels in a multiplexed volume Bragg grating,” Opt. Express 21(24), 29620–29627 (2013). [CrossRef] [PubMed]
16. K. Vodopyanov, “Optical THz-wave generation with periodically-inverted GaAs,” Laser Photonics Rev. 2(1–2), 11–25 (2008). [CrossRef]
17. J. J. Degnan, “Optimzation of passively Q-switched lasers,” IEEE J. Quantum Electron. 31(11), 1890–1901 (1995). [CrossRef]
18. G. Xiao and M. Bass, “A generalized model for passively Q-switched lasers including excited state absorption in the saturable absorber,” IEEE J. Quantum Electron. 33(1), 41–44 (1997). [CrossRef]
19. R. Paschotta, “Field guide to laser pulse generation, dynamics of passive Q switching,” SPIE Field Guides 14, 18–20 (2008).
20. A. Korenfeld, D. Sebbag, U. Ben-Ami, E. Shalom, G. Marcus, and S. Noach, “High pulse energy passive Q-switching of a diode pumped Tm:YLF laser by Cr:ZnSe,” Laser Phys. Lett. 12(4), 045804 (2015). [CrossRef]
22. R. Faoro, M. Kadankov, D. Parisi, S. Veronesi, M. Tonelli, V. Petrov, U. Griebner, M. Segura, and X. Mateos, “Passively Q-switched Tm:YLF laser,” Opt. Lett. 37(9), 1517–1519 (2012). [CrossRef] [PubMed]