Pump-limited kW-class operation in a multimode fiber amplifier using adaptive mode control and a photonic lantern front end was achieved. An array of three single-mode fiber inputs was used to adaptively inject the appropriate superposition of input modes in a three-mode gain fiber to achieve the desired mode at the output. Mode fluctuations at high power were compensated by adjusting the relative phase, amplitude, and polarization of the single-mode fiber inputs. The outlook for further power scaling and adaptive-optic compensation is described.
© 2017 Optical Society of America
Solid-state laser power scaling has been limited at high power by dynamic thermal distortions that create birefringence and thermal lensing which ultimately results in transverse-mode oscillation . In solid-state lasers, deformable mirrors have been used to stabilize the beam and to improve beam quality . While fiber amplifiers are susceptible to similar distortions at high power, the use of free-space (non-waveguide) based deformable mirror components or their equivalent have limited utility as a result of (a) bandwidth limitations arising from limited mechanical response time, (b) power-handling limitations, and (c) the need for precise opto-mechanical alignment that is subject to drift.
In a fiber amplifier, area scaling increases the nonlinear threshold of unwanted effects such as stimulated Brillouin scattering (SBS), self-phase modulation (SPM), and Raman scattering that occur at high power . As the fiber area scales the introduction of higher-order mode content limits the beam quality and creates an environment for thermal mode-instability disturbances [3–5]. Furthermore, as the fiber-core area increases the effective index spacing between the lowest-order modes decreases making it increasingly challenging to achieve bend-loss-induced spatial-mode discrimination between the fundamental and higher-order modes . Bending also reduces the mode area of a fiber, which reduces its effectiveness of suppressing nonlinearities .
Recent research has focused on the design of specialty fiber that provides additional mode discrimination using either photonic crystal, photonic bandgap, or chirally-coupled cores to achieve mode filtering . In this paper, we demonstrate an alternative approach at high power using an all-fiber-based adaptive spatial mode control system . While this demonstration uses step-index fiber, it can also be extended to specialty fibers where dynamic disturbances degrade the stability and quality of the beam at high power.
2. Photonic lantern based mode control
Previously, we demonstrated the ability of a photonic lantern control system to selectively excite and maintain fundamental-mode output in a multimode fiber with a beam quality of 1.17 and a 97% combining efficiency . In this paper, we demonstrate that by adjusting the amplitude, phase, and polarization of the input-fiber array we can excite any mode supported by a few-mode fiber. We also report on scaling the output power of the photonic-lantern amplifier to 1.27 kW. This proof-of-principle demonstration shows promise toward future power scaling.
The enabling component in this demonstration is a photonic lantern [7–11]. It consists of an array of single-mode fibers which are adiabatically tapered to form a multimode fiber output. In order to achieve low loss for arbitrary modal excitation the number of input fibers in the photonic lantern must match the number of modes in the multimode fiber. This condition is necessary but not sufficient in that it provides the same number of input degrees of freedom (single-mode amplitude, phase, and polarization) as the output degrees of freedom (mode amplitude, polarization, and phase). The actual mapping of the inputs to the outputs is determined by the photonic lantern design and is described by a transfer matrix [7,11].
By using a photonic lantern on the seeding stage of an amplifier we can inject arbitrary superpositions of the modes supported by the gain fiber to compensate for disturbances in the fiber. Scaling to high power in this demonstration is achieved by transmitting 10 W through a 3-channel photonic lantern that is spliced into a loosely-coiled ytterbium-doped kW amplifier. The details of the photonic lantern design have been previously reported . The photonic lantern used in this experiment was measured to have a 0.8 dB insertion loss in mapping the single-mode fiber inputs into the fundamental mode. The overall system efficiency is driven largely by the efficiency in the final gain stage and is fairly insensitive to the insertion loss in the seeding stage so long as sufficient power is provided to saturate the final gain stage.
The ~20 dB, kW gain stage displayed mode fluctuations, referred to here as dynamic mode disturbances, at an 800 W threshold when using more conventional seeding consisting of a single-mode 10 W preamplifier. By achieving stable, pump-limited, 1.27 kW fundamental-mode performance using our adaptive photonic lantern front end, we demonstrated the ability of our system to achieve beam stability at a factor of 1.57 above the mode-disturbance threshold.
3. Experimental results
A proof-of-concept high-power amplifier experiment is shown in Fig. 1. A 1064 nm seed laser is linewidth broadened to 35 GHz using a lithium niobate phase modulator. The linewidth broadening is sufficient to suppress stimulated Brillouin scattering (SBS). The spectrally broadened output is then split into three fibers. Motorized path length delay lines are used to adjust the path lengths of the individual arms to within their coherence length to enable coherent beam combination. The output of the delay lines feed into lithium niobate phase modulators. All the fiber components described including the phase modulators are polarization maintaining (PM). The PM phase modulators feed into polarization controllers which output non-PM single-mode fiber into three 10-W preamplifiers with optically isolated outputs. The preamplifier pump current can be adjusted to achieve the desired power distribution among the photonic-lantern inputs.
The output of the photonic lantern is typically set to 10 W to insure saturating the final kW amplifier gain stage. The passive fiber out of the photonic lantern consists of a 25/400 double-clad 0.065 NA fiber which was spliced into a (6 + 1)x1 pump-signal combiner that serves to combine the seed signal and pump light (976 nm) into the double-clad ~12 m long gain fiber. The gain fiber consists of a 25/400 ytterbium-doped 0.065 NA fiber. The output gain fiber is loosely coiled to a 200 mm diameter to avoid filtering of the higher-order modes.
In this experiment, the output of the gain fiber is sampled by a pinhole detector. A stochastic parallel gradient descent (SPGD) algorithm dithers the input polarization, amplitude, and phase of the inputs and measures the on-axis response for optimizing the LP01 mode [12–14]. Since this gain fiber only supports the LP01, LP11e, and LP11o modes, measuring the on-axis intensity is equivalent to measuring the LP01 mode since it is the only mode that contains an on-axis component . The algorithm measures the on-axis response to the dither, and iteratively applies a correction to maximize the intensity on axis and therefore the LP01 mode. In other experiments, the detector and pinhole are moved off-axis to excite the LP11 mode and to steer the beam. The LP11 modes are maximized when the pinhole detector position coincides with a peak of the LP11 modes. The output of the passive photonic lantern is shown in Fig. 2 as the SPGD detector is placed at three discrete locations to excite the LP11 modes and LP01 mode. This output is taken prior to the (6 + 1)x1 pump signal combiner labeled “double clad fiber splice” illustrated in Fig. 1.
For higher-order mode fibers, which contain LP0m modes, an appropriate mask may be used to project the desired mode onto the detector. As an example, a sample of the output fiber could be imaged onto a spatial light modulator (SLM). The desired mode is then created on the spatial light modulator. Using a lens, the Fourier transform of the product of the output of the fiber and the spatial light modulator mode is obtained in the far field. Since this product corresponds to a correlation in the far field, the on-axis component is proportional to the overlap integral between the modes out of the fiber and the mode on the SLM  and could be maximized using a detector and the SPGD algorithm. This technique has been utilized in the three-moded fiber to selectively excite the LP11 modes (see Visualization 1). The flicker shown in Visualization 1 may be in part attributed to the refresh rate of the SLM. In principle, this technique is extendable to fibers with larger numbers of modes.
Further characterization of the efficiency with which the mode-control system selectively excites the higher-order modes will be reported in future studies. One metric may be obtained by measuring the ratio of the on-axis intensity when the control system generates an LP11 mode to the on-axis intensity when generating an LP01 mode. This metric conveys the ability of the control system to extinguish the LP01 mode when commanded to generate a LP11 mode. In Fig. 2, the ratio of the null on-axis to the saturated on-axis signal shown in comparing Fig. 2 (left) and (center) with Fig. 2 (right) suggest this metric is better than 1 part in 256 for an 8-bit image. This metric however does not quantify the ability of the SLM to selectively excite the desired LP11 orientation among the degenerate LP11 orientations which all contain an on-axis null. While SPGD simulations to be described in future reports suggest high-efficiencies can be obtained in principle for small dithers, the efficiencies obtained in practice remain to be quantified and ascertained.
Having performed higher-order mode proof-of-principle demonstrations using the photonic lantern and SPGD control, the experimental configuration shown in Fig. 1 was used to test the adaptive-optic (AO) compensation capability of the photonic lantern. First, a conventional single-mode 10 W preamplifier was used to seed the kW gain stage. The output of the single-mode 10 W preamplifier was spliced into the double-clad pump-signal combiner which seeds the loosely coiled, ~20 dB, kW gain stage, which supports the LP01 and LP11 modes. For the proof-of-principle experiments described below, a gain fiber that had failed quality control inspection for exhibiting mode fluctuations at low power (<1 kW) using tighter coil diameters (~100 mm) was selected as a candidate to demonstrate adaptive-mode control.
At the 200 mm coil diameters used in this experiment, the output beam was stable for a fixed pump power up to an output power of 800 W on the time scale of ~seconds. As the pump power increased toward 800 W, the beam profile changed, indicating that higher-order modes were injected. Beyond 800 W, mode oscillations were observed on a far-field camera (see Visualization 2) and the on-axis signal displayed oscillation at ~100 Hz as shown in Fig. 3 (a). The oscillations grew in magnitude as the power level increased above and around 800 W suggesting a threshold dependence similar to previously reported thermally-induced mode-instability behavior [3–5]. The power was not increased significantly beyond 800 W to avoid premature failure. Since the fundamental mode is the only mode with an on-axis intensity component in a three-mode fiber, the on-axis signal oscillation suggests energy transfer from the fundamental mode.
Furthermore, the output power as a function of pump power for the conventional amplifier configuration shown in blue in Fig. 4 illustrates a decreased efficiency in signal extraction which may be indicative of increased thermal dissipation. The red trace illustrates the output power with a photonic lantern front end and suggests that improved signal extraction may be possible by controlling the mode content to stabilize the fundamental mode.
While the ~100 Hz disturbance is lower than the 1-2 kHz characteristic frequency signature of transverse-mode instabilities typically reported for similar fiber, recent reports have shown that the co-propagating pump geometries used in these experiments exhibit fluctuations at ~200 Hz which is lower than the ~2 kHz reported using the same fiber in a counter-pumped geometry . It is beyond the scope of this publication to determine the origin of the observed mode fluctuations. Instead, we refer to the sudden threshold onset of mode fluctuations as a mode disturbance while remaining control-system agnostic to its origin. It should be noted that other fibers have been tested of similar geometry that did not exhibit the mode disturbances shown in Fig. 3 (a) for similar pump powers tested up to pump-limited 1.56 kW operation. While the adaptive mode control system applied to these fibers also stabilized the output, a stable beam with good beam quality could be obtained by alternative means such as coiling, and is therefore not as good a candidate to demonstrate the merit of the adaptive mode control described below.
The role of the control system in Fig. 1 is to perform adaptive correction using the principle of wavefront conjugation that is successfully employed in adaptive-optic applications. In particular, we can envision sending the desired mode (i.e. fundamental mode) in reverse for the purposes of illustration . In practice, the control system does not require reverse propagation. In the presence of mode instability or other mode disturbances, a superposition of the modes supported in the few-mode fiber are generated. This superposition then evolves into the single-mode fiber inputs which can be described by a combination of amplitude, polarization, and phase. The job of the control system is to determine the amplitude, conjugate phase, and polarization to faithfully reproduce the desired mode in the forward direction. The control system must respond on the time scale of the disturbance. The FPGA-based SPGD controller used in these experiments has a dither rate of 800 kHz which is capable of compensating ~kHz class disturbances . The bandwidth of the control system is proportional to the dither rate . In principle, other control algorithms such as LOCSET may be used .
Next, in order to demonstrate the control system, the photonic lantern front end was spliced into the pump-signal combiner. The gain was increased to pump-limited 1.27 kW output operation for a measured 1.56 kW of available pump power. With the control system turned on, the on-axis intensity was stabilized and reached a maximum as shown in Fig. 3 (b) using the pinhole detector method without an SLM as shown schematically in Fig. 1. When the control system was turned off, the intensity on axis decreased and fluctuated as indicated in Fig. 3 (b). Accompanying Visualization 3 illustrates captured images from a far-field camera observing the output with the control system on and off. The effectiveness of the photonic-lantern control system to stabilize the output beam may be determined by comparing the left of Fig. 3 (b) with Fig. 3 (a) at pump-limited output power. Similar stability was achieved at lower power including 800 W using the photonic lantern based adaptive mode control front end.
It is worth noting that the dynamics with the control system off using a photonic lantern front end are different than that observed using a single-mode preamplifier as evident in comparing Figs. 3 (a) and (b). Experimentally, in the case of Fig. 3 (a), a single-mode preamplifier seeds the loosely coiled multimode gain stage. At the splice interface between the preamplifier and the high-power gain stage, the injected modal content is anticipated to be static given the single-mode characteristic of the preamplifier. In Fig. 3(b), when the control system is turned off, random phase fluctuations on the input fibers to the photonic lantern generate random mode superpositions on the output. This dynamic mode injection is sufficient to disrupt the periodic oscillation observed in Fig. 3(a). While the output is less regular with the control system off, it is nonetheless unstable with the random photonic lantern mode injection. With the control system turned on, the input phase and polarization are controlled to stabilize and maximize the on-axis intensity. The amplitudes are set to their near approximate optimum values but are not varied in the high-power demonstration. The output power as a function of pump power is shown in Fig. 4 with the photonic lantern control system on, and representative output mode profiles are shown in Fig. 5.
As illustrated in Fig. 5 at 1.27 kW, the on-axis intensity is lower when the control system is turned off (left) relative to the case with the control system turned on (right). The mode is also distorted with the SPGD control off as shown in a typical frame illustrating elongation and distortion in Fig. 5 (left); from frame-to-frame, the mode content and beam distortion varies on the camera. With the SPGD control on, the on-axis intensity increases as shown on the right in Fig. 5, and the beam is observed to be stable on the camera from frame-to-frame (see Visualization 3). These observations are consistent with the on-axis signal trace shown in Fig. 3 where SPGD control is shown to stabilize and maximize the on-axis intensity.
We have demonstrated the ability of the adaptive spatial mode control technique to compensate for mode disturbances in a fiber amplifier at high power and achieved pump-limited 1.27 kW output power. This is a critical milestone toward the development of high-power amplifiers using all-fiber based AO compensation. As the core area scales in a fiber, the onset of intensity-limited nonlinearities increases. However, the introduction of higher order mode content degrades the beam quality and creates mode instabilities at high power. Furthermore, as the core area increases the mode-instability threshold decreases . The prospects for further mode area and power scaling are promising based on the results achieved in this proof-of-concept demonstration which demonstrates the ability of the adaptive control system to compensate for mode disturbances irrespective of their origin. While the disturbances in this demonstration contained frequency content of ~100 Hz, the SPGD algorithm can compensate for higher frequency content with bandwidths proportional to the dither rate  which was 800 kHz in our control system.
In particular, it is worth noting that as the mode area increases, the mode-instability dynamics decrease in bandwidth which in turn reduce the overall bandwidth requirements on a control system. Furthermore, as the fiber core area increases the number of modes increases linearly. By means of an example, by doubling the core diameter to 50 µm, the number of modes would increase by a factor of ~4 which would require ~4x more fibers in a photonic lantern to achieve full control. Assuming a linear increase in power with area, the output power could increase by ~4x. It is also worth noting that the fiber could be coiled to reduce the number of modes and the number of input control channels. However, the mode-shrinking with bending needs to be considered if coiling is to be used.
It should also be noted that free-space systems have previously demonstrated active control of mode instability and have successfully scaled the mode-instability threshold by a 3x factor . Here we presented a system developed toward obtaining a similar goal in an all-fiber based format that alleviates the need for free-space optics. Furthermore, the principle of operation has been described from the vantage point of an adaptive-optics description based on reversibility and wavefront conjugation. This builds on prior demonstrations and is applicable toward generalized mode control using a photonic lantern and SPGD .
Lastly, in general, a photonic-lantern amplifier may also be of benefit in applying compensation for atmospheric disturbances. In particular as the core increases more modes may be used to compensate for propagation through atmosphere or other turbulent environments.
U. S. Air Force (FA8721-05-C-0002 and/or FA8702-15-D-0001).
This material is based upon work supported by the High Energy Laser Joint Technology Office (HEL-JTO) under Air Force Contract No. FA8721-05-C-0002 and/or FA8702-15-D-0001. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the HEL-JTO.
References and links
1. S. Piehler, T. Dietrich, P. Wittmüss, O. Sawodny, M. A. Ahmed, and T. Graf, “Deformable mirrors for intra-cavity use in high-power thin-disk lasers,” Opt. Express 25(4), 4254–4267 (2017). [PubMed]
2. C. Jauregui, J. Limpert, and A. Tünnermann, “High power fibre lasers,” Nat. Photonics 7, 861 (2013).
3. F. Jansen, F. Stutzki, H. J. Otto, T. Eidam, A. Liem, C. Jauregui, J. Limpert, and A. Tünnermann, “Thermally induced waveguide changes in active fibers,” Opt. Express 20(4), 3997–4008 (2012). [PubMed]
4. T. Eidam, C. Wirth, C. Jauregui, F. Stutzki, F. Jansen, H. J. Otto, O. Schmidt, T. Schreiber, J. Limpert, and A. Tünnermann, “Experimental observations of the threshold-like onset of mode instabilities in high power fiber amplifiers,” Opt. Express 19(14), 13218–13224 (2011). [PubMed]
5. A. V. Smith and J. J. Smith, “Mode instability in high power fiber amplifiers,” Opt. Express 19(11), 10180–10192 (2011). [PubMed]
6. S. Wielandy, “Implications of higher-order mode content in large mode area fibers with good beam quality,” Opt. Express 15(23), 15402–15409 (2007). [PubMed]
7. J. Montoya, C. Aleshire, C. Hwang, N. K. Fontaine, A. Velázquez-Benítez, D. H. Martz, T. Y. Fan, and D. Ripin, “Photonic lantern adaptive spatial mode control in LMA fiber amplifiers,” Opt. Express 24(4), 3405–3413 (2016). [PubMed]
8. S. G. Leon-Saval, T. A. Birks, J. Bland-Hawthorn, and M. Englund, “Multimode fiber devices with single-mode performance,” Opt. Lett. 30(19), 2545–2547 (2005). [PubMed]
9. D. Noordegraaf, P. M. Skovgaard, M. D. Nielsen, and J. Bland-Hawthorn, “Efficient multi-mode to single-mode coupling in a photonic lantern,” Opt. Express 17(3), 1988–1994 (2009). [PubMed]
10. S. G. Leon-Saval, A. Argyros, and J. Bland-Hawthorn, “Photonic lanterns: a study of light propagation in multimode to single-mode converters,” Opt. Express 18(8), 8430–8439 (2010). [PubMed]
11. N. K. Fontaine, R. Ryf, J. Bland-Hawthorn, and S. G. Leon-Saval, “Geometric requirements for photonic lanterns in space division multiplexing,” Opt. Express 20(24), 27123–27132 (2012). [PubMed]
12. M. A. Vorontsov, G. W. Carhart, and J. C. Ricklin, “Adaptive phase-distortion correction based on parallel gradient-descent optimization,” Opt. Lett. 22(12), 907–909 (1997). [PubMed]
13. S. M. Redmond, D. J. Ripin, C. X. Yu, S. J. Augst, T. Y. Fan, P. A. Thielen, J. E. Rothenberg, and G. D. Goodno, “Diffractive coherent combining of a 2.5 kW fiber laser array into a 1.9 kW Gaussian beam,” Opt. Lett. 37(14), 2832–2834 (2012). [PubMed]
14. J. Montoya, S. J. Augst, K. Creedon, J. Kansky, T. Y. Fan, and A. Sanchez-Rubio, “External cavity beam combining of 21 semiconductor lasers using SPGD,” Appl. Opt. 51(11), 1724–1728 (2012). [PubMed]
15. D. Flamm, D. Naidoo, C. Schulze, A. Forbes, and M. Duparré, “Mode analysis with a spatial light modulator as a correlation filter,” Opt. Lett. 37(13), 2478–2480 (2012). [PubMed]
16. B. Yang, H. Zhang, C. Shi, R. Su, P. Ma, X. Wang, P. Zhou, X. Xu, and J. Chen, “Experimental Study of the Transverse Mode Instability in a 3kW-level Bidirectional-pumped All-fiber Laser Oscillator,” in Laser Congress 2017 (ASSL, LAC), OSA Technical Digest (online) (Optical Society of America, 2017), paper JM5A.15.
17. A. Brignon, Coherent Laser Beam Combining (Wiley-VCH, 2013).
18. H. J. Otto, C. Jauregui, F. Stutzki, F. Jansen, J. Limpert, and A. Tünnermann, “Controlling mode instabilities by dynamic mode excitation with an acousto-optic deflector,” Opt. Express 21(14), 17285–17298 (2013). [PubMed]