High-energy Q-switched Tm3+-doped polarization maintaining silica fiber laser Christina C. C. Willis*a, Lawrence Shaha, Matthieu Baudeleta, Pankaj Kadwania, Timothy S. McCombb, R. Andrew Simsa, Vikas Sudeshc, Martin Richardsona a -Townes Laser Institute, CREOL College of Optics and Photonics, University of Central Florida, 4000 Central Florida Boulevard, Orlando, Florida 32816, USA b -now at Northrop Grumman, 1 Space Park Blvd., Redondo Beach, California 90278 c -now at Quantum Tech Inc., 108 Commerce St., Suite 102, Lake Mary, Florida 32746 ABSTRACT We report the performance of an actively Q-switched Tm fiber laser system. The laser was stabilized to sub-nanometer spectral width using each of two feedback elements: a blazed reflection grating and a volume Bragg grating. Maximum sub-nanometer pulse energy using the grating was 325 µJ pulses at 1992 nm with a 125 ns duration at a 20 kHz repetition rate. Maximum sub-nanometer pulse energy using the volume Bragg grating was 225 µJ pulses at 2052 nm with a 200 ns duration also at 20 kHz. We also report the laser’s performance as an ablation source for LIBS experiments on copper. Keywords: fiber laser, thulium, polarization maintaining, Q-switch, Laser-induced breakdown spectroscopy
1. INTRODUCTION This paper reports the performance of an actively Q-switched, Tm3+-doped, polarization-maintaining (PM), silica, large mode area (LMA) fiber laser (Fig. 1). Thulium fiber lasers are attractive for applications that require moderate pulse energy and high average power in the “eye-safe” wavelength regime (λ > 1.4 µm). These applications include LIDAR, atmospheric propagation, laser induced breakdown spectroscopy (LIBS), free space communications and directed energy [1-7]. Thulium-doped fiber lasers possess the characteristics necessary for such applications: high efficiency, broad spectral emission, near diffraction-limited beam quality, and a compact and robust design [5, 8-10]. Thulium fiber lasers have been shown to produce pulse energies of 270 μJ with 41 ns duration [6] without polarization control. However, some of these applications require not only precise wavelength control and tunability, but also fine polarization control.
Figure 1: Table-top view of Q-switched laser system
Thulium fiber lasers are also very efficient, both in terms of pump and slope efficiencies. Pumping can be performed efficiently with high-power 790 nm diodes [8]; these diodes are both technologically mature and readily available commercially. And despite thulium’s large quantum defect, thulium-based lasers tend to have high slope efficiencies due to a well-known cross relaxation process. The cross-relaxation process allows two signal photons at ~ 2 µm to be radiated for a single absorbed 790 nm pump photon [11]. With sufficient doping, cross-relaxation occurs in *
[email protected]; phone 407.823.6832; fax 407.823.6880; http://lpl.creol.ucf.edu
silica fiber and it gives thulium a maximum theoretical slope efficiency of 80% [12]. Experimentally the highest slope efficiency achieved by a thulium fiber laser within our research group was 65% [13]. The actively Q-switched thulium fiber laser presented here can generate 325 µJ pulses at a 20 kHz repetition rate with a 125 ns duration and a spectral width of 10 dB. During multi-wavelength operation (using the HR mirror) the pulse energy can be increased to 360 µJ and the quality of the pulse energy and polarization are maintained at increasing repetition rates (up to 100 kHz), reaching >28 W average power. The laser was used to perform preliminary laser induced breakdown spectroscopy (LIBS) experiments on copper samples. Application of a 2 μm laser as an ablation source in LIBS experiments promises to yield cleaner measurements free of certain signal noise, such as continuum emission.
2. FIBER LASER 2.1
Experimental Setup
The heart of the fiber laser is a 4 m piece of Tm3+-doped PM fiber. The fiber, from Nufern Inc., has a 25 µm/0.10 NA core and a 400
µm cladding. The active fiber is wrapped around an 11 cm diameter aluminum mandrel water-cooled at 140 C. Because thulium behaves as a quasi three-level laser system, re-absorption of the signal can hurt overall laser performance. Thermal maintenance, via the water-cooled mandrel, minimizes the re-absorption. Approximately half a meter of passive fiber with matching core and cladding is spliced to each end of the active fiber. This is another method of thermal management, which circumvents heating problems that can result from directly pumping active fiber. Each of the passive ends of fiber is mounted in a copper v-groove. The copper mount that holds the passive fiber through which pump light is coupled is also water-cooled to 140 C. A 790 nm fiber-coupled diode (LIMO) with a maximum power of 300 W is used to pump the fiber. The pump light is free-space coupled into the cooled passive fiber end using a 1:1 telescope and a dichroic mirror. The telescope consists of two 5 cm diameter 0.26 NA infinite-conjugate achromatic triplet lenses, and is ~80% transmissive. However, overall pump coupling efficiency is ~75% due to an additional ~5% mode matching loss. The dichroic mirror is high reflectivity at 790 nm and high transmission from 1.85 to 2.1 µm. Light from the pump passes through the telescope and is then reflected off the dichroic mirror and onto the fiber facet. The pumped end of the fiber is angle cleaved at ~80 to prevent parasitic lasing from Fresnel reflections. The opposite end is cleaved normally and the 4% Fresnel reflections provide the feedback, allowing this end to act as the laser’s output coupler. Light from the output end is collimated with an uncoated ~90% transmissive singlet lens with a 50 mm focal length.
Figure 2: Q-switched Laser Schematic
An aplanatic triplet lens with a focal length of 26 mm acts as an intra-cavity collimator for the light exiting the pumped fiber end. After collimation the signal then passes through a quarter waveplate (QWP), a half waveplate (HWP), and a 1000:1 polarizing beam splitter (PBS). These three elements ensure that the laser signal will have a single linear
polarization state. They also allow the polarization state to be rotated so that it matches both the stress rods of the PM fiber and maximizes the grating efficiency. After the PBS, the signal passes through the Q-switch before reaching the feedback element. The Q-switch is a TeO2 acousto-optic modulator (AOM) from NEOS Technologies which has a ~70% diffraction efficiency for the -1 order. The feedback element is set after the AOM and reflects back the -1 order. Three different feedback elements were investigated: a highly reflective mirror at 2 μm, a 600 l/mm gold-coated reflection grating (>90% reflectivity for horizontally polarized light), and a 5x5x6 mm3 volume Bragg grating optimized for feedback at 2052.5 nm (~95% reflectivity). 2.2
Laser performance
The laser’s Q-switched performance was examined using three different feedback elements. The first was an HR mirror at 2 µm, which allowed the laser to freely oscillate over a wide range of frequencies, spanning approximately 20 nm. The slope efficiency of the laser was found to be 33% at 20 kHz, and 40% at 100 kHz (Fig. 3a). The pulse energy and duration were also measured as a function of incident pump power (Fig. 3b), and it was noted that pulse energy increased as pulse duration decreased. The shortest pulse duration achieved using the HR mirror was 115 ns at a 100 kHz repetition rate, and had a pulse energy of 360 μJ.
a
b
Figure 3: a) Q-switch output power vs. launched power using an HR mirror, b) 4 pulse energy and pulse duration vs. launched npump power at 20 kHz
Using the 600 l/mm gold-coated reflection grating, the laser produced 325 µJ pulses at 20 kHz with a 125 ns duration and a wavelength of 1992 nm with sub-nanometer spectral width. Above 325 µJ the laser begins for oscillate at multiple wavelengths. These results are similar to those of Eichhorn et al. who generated 270 μJ pulses with 41 ns duration in non-polarization-maintaining thulium fiber laser system at 1983 nm [6, 7]. The laser had a slope efficiency of 28% with the grating (Fig. 4). Employing the VBG, the laser generated 225 µJ pulses with 200 ns duration and a 20 kHz repetition rate at 2052 nm, the peak reflectivity of the VBG. At energies higher than 225 µJ, parasitic lasing occurred around 1980 nm, which is thulium’s amplified spontaneous emission (ASE) peak. The laser achieved a 24% slope efficiency with the VBG (Fig. 4).
Figure 4: The Q-switch output power vs. launched power at 20 kHz for HR, gold-coating grating, and VBG. The slope efficiency is 33% using HR, 28% using gold-coated grating, and 24% using VBG.
3. LASER-INDUCED BREAKDOWN SPECTROSCOPY 3.1
Experimental Setup
To perform laser-induced breakdown spectroscopy (LIBS) experiments, the reflection grating was used as the feedback element for the laser. Output from the laser was collimated using an uncoated ~90% transmissive singlet lens with a 50 mm focal length. After collimation the beam was focused onto the sample using an 11 mm focal length 0.3 NA aspheric lens with ~95% transmission at 2 µm. The sample was placed at a ~45° angle to beam propagation at the focal point of the lens to facilitate collection of the LIBS signal. Pieces of 0.25 mm-thick 99.98% copper foil (349178, Sigma-Aldrich) were used as target samples. Each sample was cleaned with acetone prior to irradiation to remove any oxidation layers that may have formed on its surface. Samples were then mounted on three motorized stages (VP-25XA) in an X-Y-Z configuration. The stages were operated with a Newport motion controller (ESP300). This configuration allowed the samples to be translated along the X (45° to the beam’s axis) and Y (vertical) axes such that a point on the sample’s surface was always located at the focal point of the lens. Samples were translated at a rate of 25 mm/s. Plasma was generated using 100 µJ pulses with durations of ~200 ns at 1992 nm and a 20 kHz repetition rate. The incident beam had a focal waist diameter of ~10 µm, yielding a target it irradiance of 600 MW/cm2. Plasma emission was collected with an f/2 UV transmissive collimating ball lens (BL) from Ocean Optics (74-UV) and coupled directly into a 1 meter UV transmissive fiber with a 200 µm core (HRE-FBR-1M, Princeton Instruments). The light was sent into a 250 mm Echelle spectrometer (Acton HRE, Princeton Instruments) and an iCCD camera. The GenII iCCD camera has 1024x1024 pixels with an MgF2 input window and a P46 phosphor (PI-MAX2, Princeton Instruments). The resolution of the system was 0.04 nm for the 200 – 900 nm scan range. Each of the spectra taken was a sum of 5000 single-shot spectra (-100 ns delay, 400 ns acquisition). Thirty spectra were taken to provide statistical measurement.
Fig. 5: LIBS experimental set up
3.2 Results The copper spectrum reveals the absence of several typical sources of signal noise (Fig. 6). First, the spectrum does not show signal from the laser. This is because the laser’s wavelength lay outside of the detection range, obviating the need for a high-speed camera to avoid its detection. No continuum emission is seen in the spectrum, which could imply a low electronic density in the plasma. Emission from atmospheric nitrogen and oxygen atomic transitions are typical of plasmas generated with these fluencies in normal atmosphere, but these are also absent from the spectrum. Calculations performed on the spectral data reveal a significant difference in the excitation temperature of atomic and singly-ionized copper, suggesting that the plasma is not in local thermal equilibrium. Details of the calculations may be found in reference 5.
Figure 6: LIBS spectrum of copper, showing the rich ultraviolet emission from singly-ionized copper. The inset gives a broader range, from 250 to 815 nm, showing mainly atomic transitions.
4.
CONCLUSIONS AND DISCUSSION
The fiber laser described here shows promise for applications that require high pulse energies, and narrow wavelength and polarization control. The highest slope efficiency at 20 kHz (33%) was found using the HR mirror which provided poor wavelength control and a spectral width of ~20 nm centered at 1980 nm. Both the reflection grating and the VBG offered sub-nanometer linewidth at their respective wavelengths (1992 and 2052 nm). However, the reflection grating demonstrated a higher slope efficiency (28%) than the VBG (24%). At a 20 kHz repetition rate the grating can generate 325 μJ pulses with a 125 ns duration, whereas with the VBG the laser could only generate 225 μJ pulses with a 200 ns duration. In both cases, the PER was >10 dB, and the upper limit on performance was the onset of parasitic lasing. These values could be improved further by implementing a higher efficiency reflection grating optimized for 2 μm light or a VBG with reflectivity centered at the ASE peak. The LIBS experiments performed, the first ever to use a thulium fiber laser as its ablation source, gave interesting results. The plasma generated had no continuum emission and showed no spectral lines from excited atmospheric species. This, in addition to the absence of the laser signal within the measured range, makes the 2 μm light a desirable ablation source.
ACKNOWLEDGEMENTS This work is supported by the JTO-HEL MRI program on high power fiber lasers (contract #W911NF-05-1-0517), and the State of Florida.
REFERENCES [1] P. Sprangle, A. Ting, J. Penano, R. Fischer, and B. Hafizi, "Incoherent Combining and Atmospheric Propagation of High-Power Fiber Lasers for Directed-Energy Applications," Quantum Electronics, IEEE Journal of 45, 138-148 (2009). [2] G. J. Koch, J. Y. Beyon, B. W. Barnes, M. Petros, J. Yu, F. Amzajerdian, M. J. Kavaya, and U. N. Singh, "Highenergy 2 µm Doppler lidar for wind measurements," Optical Engineering 46, 116201-116214 (2007). [3] L. Shah, J. Zeller, M. Richardson, M. L. Talbert, G. S. Seetharaman, and T. Manzur, "Wideband lasers on manned/unmanned platforms under poly-environment," in Unmanned Systems Technology XI, (2009), 733203733205. [4] F. Hanson, P. Poirier, D. Haddock, D. Kichura, and M. Lasher, "Laser propagation at 1.56 μm and 3.60 μm in maritime environments," Appl. Opt. 48, 4149-4157 (2009). [5] M. Baudelet, C. Willis, L. Shah, and M. Richardson, "Laser Induced Breakdown Spectroscopy of Organic Materials with a mid-IR Thulium Fiber Laser Nanosecond Pulse at 2 µm," in North American Symposium on Laser Induced Breakdown Spectroscopy 2009, (New Orleans, LA, 2009). [6] M. Eichhorn, S. D. Jackson, “High-pulse-energy actively Q-switched Tm3+-doped silica 2 μm fiber laser pumped at 792 nm,” Optics Letters 32, 2780-2781 (2007). [7] M. Eichhorn, S. D. Jackson “Actively Q-switched Tm3+-doped and Tm3+,Ho3+-codoped Silica Fiber Lasers,” in Conference on Lasers and Electro-Optics and Quantum Electronics and Laser Science Conference 2008, (San Jose, CA, 2008). [8] S. Agger, J. J. Povlsen, P. Varming, “Single-frequency thulium-doped distributed-feedback fiber laser” Optics Letters 29, 1503-1505 (2004). [9] W. A. Clarkson, N. P. Barnes, P. W. Turner, J. Nilsson, D. C. Hanna “High-power cladding-pumped Tm-doped silica fiber laser with wavelength tuning from 1860 to 2090 nm,” Optics Letters 27, 1989-1991 (2002). [10] A. F. El-Sherif, T. A. King “High-energy, high-brightness Q-switched Tm3+-doped fiber laser using an electro-optic modulator,” Optics Communications 218, 337-344 (2003). [11] S. D. Jackson, "Cross relaxation and energy transfer upconversion processes relevant to the functioning of 2 µm Tm3+-doped silica fibre lasers," Optics Communications 230, 197-203 (2004). [12] J.Y. Allain, M. Monerie, H. Poignant, “High-efficiency CW thulium-sensitised holmium-doped fluoride fibre laser operating at 2.04 μm” Electronics Letters 27, 1513-1515 (1991).
[13] T. S. McComb, L. Shah, R. A. Sims, C. C. C. Willis, P. Kadwani, V. Sudesh, M. Richardson “Portable, high power, narrow linewidth widely tunable thulium fiber Lasers system” Pending submission (2010).
