The beam quality of large-mode-area fiber amplifiers was investigated at the 10 W power level using a tunable ring cavity, that is also used in the laboratory system of GEO600 as a pre-mode-cleaner for mode filtering. More than 98% of the overall output power were contained within the polarized (200:1) TEM00 mode with an appropriate choice of coiling diameter. With the high sensitivity ring cavity analysis, the beam quality improvement caused by decreasing the coiling diameter was verified, while this could not be seen within conventional M2-measurements. The results are compared with the properties of single-mode fibers.
©2003 Optical Society of America
There is a growing need for stable, low noise, single-frequency, single-mode sources for applications like ultrahigh-precision interferometry for gravitational-wave detection as one of the most demanding examples. In the last years it was demonstrated, that fiber amplifiers are suitable for the generation of high output powers . However, for single-frequency generation, stimulated Brillouin scattering imposes an upper limit on the output power due to a long interaction length and a high power density in the single-mode fiber core. The use of multi-mode fibers led to the generation of several watts of output power  and recently the amplification in a large-mode-area (LMA) fiber above the 100 W level has been reported . According to these results and prior investigations , a high beam quality (M2<1.1) can be achieved by suppressing higher order modes due to bending losses in these fibers. While scanning ring cavities have already been used for the mode analysis of solid state laser systems , the beam quality analysis of fiber based systems has only been performed by conventional M2-measurements to date. However, for high-precision metrology applications only the polarized output power contained within the fundamental Gaussian mode can be used.
We report, to the best of our knowledge for the first time, about high sensitivity measurements of the beam quality of both single-mode and large-mode-area fiber amplifiers. For this purpose, we used a non-confocal scanning ring cavity, that is also used in the laboratory system of GEO600 as a pre-mode-cleaner for mode filtering. For comparison we carried out conventional M2-measurements.
2. Experimental setup
In our setup, see Fig. 1, the radiation of a 1064 nm-Nd:YAG single-frequency nonplanar ring oscillator was focused into the active core of an Ytterbium-doped double-clad fiber. The amplifier was counter-pumped by a fiber coupled diode laser at 975 nm. On the pump side of the active fiber the amplified signal was separated from the pump light by a dichroic mirror. It was then imaged onto an iris aperture to cut off radiation from the pump core. Both ends of the active fiber were angle polished (8°) in order to suppress backreflections from the fiber ends. All fibers used in our experiments were non-polarization-maintaining, hence the output radiation was elliptically polarized due to the static birefringence of the optical fibers. Therefore, the polarization of the amplified signal was adjusted by a combination of a quarter-and a half-wave plate and passed through a polarizing beam splitter. For the beam quality analysis the signal was attenuated by a partially reflecting mirror to a power level suitable for the final photodetection and then mode-matched to the fundamental mode of the ring cavity. Both the transmitted and reflected signals were detected by photodetectors. Additionally, the excited mode of the ring cavity was monitored with a CCD-camera placed behind one of the highly reflecting ring-cavity mirrors in order to ensure excitation of the fundamental mode.
The detailed optical setup of the ring cavity used in these experiments is described elsewhere . The finesse of the ring cavity depended on the input polarization of the analyzed signal. We chose the input signal to be p-polarized which resulted in a finesse of F=200. The free spectral range of the ring cavity was 714 MHz.
For the beam quality characterization the reflected power signal from the ring cavity was analyzed, as this is more accurate than the analysis of the transmitted power signal. When the fundamental mode of the cavity is scanned, all power, that is mode-matched to this mode, will be passed through the cavity. Hence the decrease of reflected power directly gives a lower limit for the amount of power within the fundamental mode. In contrast, the analysis of the transmitted power signal is more difficult. The overall power is distributed over all modes that are excited while the cavity is scanned. Hence the sum of the heights of all transmission peaks corresponds to the overall input power. Very small higher order peaks can not be taken into account as they may not be visible due to measurement noise. Furthermore, the measured value can be inaccurate due to any offsets in the photodetection.
For comparison with the scanning ring cavity measurements we carried out conventional M2-measurements of the polarized radiation behind the PBS. For this purpose an additional lens was placed behind the partially reflecting mirror and the resulting beam diameter was measured along the beam propagation through the focus over 3–4 Rayleigh ranges using a commercial beam analyzer (BeamAlyzer by Melles Griot).
In addition, we determined the polarization ratio by the ratio of the maximum and minimum power that could be passed through the polarizer by adjustment of the waveplates. Furthermore, the amount of amplified spontaneous emission (ASE) in the optical spectrum of the polarized light behind the PBS was recorded for all fibers under test with an alignment of the waveplates for maximum transmission.
In order to investigate the influence of the fiber parameters on the beam quality, we analyzed three different types of active fibers. As a reference measurement for the validity of our results we measured the beam quality of a single-mode fiber (length: 14 m) with a 4 µm core and a numerical aperture (NA) of 0.16. The active core of the second fiber (length: 20 m) had a diameter of 10 µm with a NA of 0.07. The cut-off wavelength of the second fiber was well below 1064 nm, hence a single-mode output was expected from this fiber as well. Finally, we measured the beam quality of a large-mode-area fiber (length: 12.5 m) with a core-diameter of 29 µm and a NA of 0.054. The operation of the LMA fiber amplifier was studied for three different coiling diameters (21 cm, 31 cm and 62 cm) in order to investigate the influence of bending losses for the higher order modes on the output beam quality. The dopant concentrations of the fibers were 6500 mol ppm, 1000 mol ppm and 500 mol ppm Yb2O3, respectively. The pump cladding was D-shaped for improved pump light absorption and the cladding diameter was 400 µm with a NA of 0.37 for all fibers used in the experiments.
3. Experimental results
Even for a standard single-mode fiber, the fundamental mode shows small deviations from a Gaussian shape causing a theoretical limit for the transmission through the ring cavity. In order to get an estimate for the maximum achievable beam quality, the measured refractive-index distributions of the fiber-preforms were scaled down to the fiber dimensions and the intensity distribution for the fundamental mode was numerically calculated for each fiber. Finally, the best overlap integral between this mode and a fundamental Gaussian mode was determined. According to our calculations we expected a maximum transmission through the ring cavity of 99.2% for the 4 µm-amplifier and 99.6% for the 10 µm- and the 29 µm-amplifier.
The 4 µm single-mode fiber amplifier was studied at an output power of 1 W. This corresponds to a gain of ~16, as about 60 mW of the NPRO radiation were coupled into the fiber core.
At this power level we measured an M2 <1.05+/-0.05. The polarization ratio was higher than 200:1 and the amplifier output contained only 0.03% of ASE, see Fig. 2, blue trace. The reflected power signal from the ring cavity decreased below 2.5% when the TEM00 mode was scanned. This confirmed that more than 97.5% (theoretical limit: 99.2%, see above) of the output power were contained within the fundamental Gaussian mode. The remaining deviation can be explained by the slightly elliptical beam profile due to the angled fiber ends.
The 10 µm-amplifier was operated at an output power level of 7 W. This corresponds again to a gain of ~16 as the NPRO output power was increased up to an injected power of about 430 mW. The amount of ASE in the polarized amplified radiation behind the PBS was measured to be below 0.005% (Fig. 2, red trace). The results of the M2-measurements (M2<1.05+/-0.05) were identical to the case of the 4 µm single-mode amplifier. In the scanning cavity analysis the reflected power dropped below 1.3%. These results confirm, that the output of the 10 µm-amplifier provided single-mode beam quality, as one would predict from the fiber parameters.
From the differences in the results of the two single-mode fiber amplifiers, the fundamental advantages and disadvantages of the two beam quality measurement methods can be seen. On the one hand, the actual beam quality will always be equal or better than measured with the scanning cavity analysis, as this method will always give a lower limit for the amount of power within the fundamental Gaussian mode. However, the scanning cavity analysis is very sensitive to the mode-matching. Hence the actually measured value will strongly depend on the quality of the mode-matching and will converge to the final value with improved adjustment. On the other hand, the M2-measurements are relatively simple to perform, but they will always have a measurement error in both directions. Furthermore, any misalignment in the measurement setup will increase the uncertainty in the M2-results, while the scanning cavity analysis will always give a lower limit for the beam quality.
For the investigations on the LMA-amplifier the NPRO was operated at the same output power as for the 10 µm-amplifier. The maximum output power of the amplifier depended on the coiling diameter of the LMA-fiber. For the maximum coiling diameter of 62 cm we achieved a maximum output power of 11.8 W while it decreased to 11 W for a coiling diameter of 21 cm as a result of the increased bending losses. All measurements on the LMA-amplifier were carried out at an output power of 10 W. The polarized amount of ASE in the output radiation did not depend on the coiling diameter and was measured to be 0.35%, see Fig. 2, black trace.
This relatively high amount of ASE compared to the other amplifiers can be explained by the quasi-three-level structure of the Yb-ion. With increasing core diameter the product of Yb-concentration and fiber length decreased for the tested fiber configurations. As the gain was roughly constant, this results in an increase of the inversion. Hence, the reabsorption decreases and the gain maximum is shifted towards shorter wavelengths . At the emission-cross-section maximum at about 1040 nm the reabsorption remained strong enough to suppress this peak in the 4 µm- and the 10 µm-amplifier. However, for the 29 µm-amplifier the gain maximum is shifted in this 1040 nm range and the emission dominates over the absorption. This results in a strong ASE emission at this wavelength as the gain is higher than for the signal-wavelength .
For the coiling diameter of 62 cm the polarization ratio was below 20:1, see Fig. 3(a). Interestingly, the polarization ratio showed a maximum value at an amplifier power of about 4–5 W. This can be explained by an interaction between the fiber gain and bending losses.
Higher order modes suffer higher bending losses than the fundamental mode. With increasing amplifier output power, the gain for the higher order modes can overcome these losses. Thus any power launched into these modes e.g. by spontaneous emission or by mode conversion due to fiber bending will be amplified along with the fundamental mode. As each individual fiber mode experiences a different effective fiber static birefringence due to its particular intensity distribution, higher order modes show different output polarizations than the fundamental mode resulting in a decreased overall polarization ratio.
Furthermore, the polarization ratio depended on the coupling of the seed-radiation into the fiber core. If the coupling was adjusted for maximum output power, the polarization ratio dropped below 6:1, corresponding to a power loss of about 14% at the PBS, but as the reflected power from the ring cavity dropped below 1.8%, more than 98% of the polarized output power were contained within the fundamental mode. In total, less then 86% of the overall amplifier output power were contained within the polarized fundamental mode.
On the other hand, if the amplifier was adjusted for maximum polarization ratio (20:1), the output power dropped about 3% and only about 95.7% of the polarized light was within the fundamental mode. Thus, in this case about 88% with respect to the maximum achievable output power (with worse polarization ratio) was contained within the polarized fundamental mode.
In summary, the total power loss by means of amplifier adjustment and transmission through the PBS and the ring cavity was about 12–14%. This degeneration of beam quality could not be seen within the M2-measurements which resulted in a value of M2=1.06+/-0.05 independently of amplifier adjustment.
The polarization ratio increased to 30:1 for a coiling diameter of 31 cm and a further reduction of the coiling diameter down to 21 cm improved the polarization ratio to more than 200:1. The polarization increased with amplifier output power (compare Fig. 3(b) for a coiling diameter of 21 cm) and did not show any local maximum as in the case of the larger coiling diameter due to the increased bending losses. For both the 31 cm and 21 cm coiling diameters, the adjustment for maximum output power corresponded to the maximum polarization ratio enabling a simple adjustment of the amplifier.
The scanning cavity results of the LMA amplifier were identical for the 31 cm and 21 cm coiling diameters and are shown for an output power of 10 W (coiling diameter: 21 cm) in Fig. 4. It can be seen, that the reflected power dropped below 1.6% (Fig. 4(a)). In comparison, an amount of 1.2% that was not within the polarized fundamental mode was obtained from adding the heights of the intermediate peaks in the transmitted power signal (Fig. 4(b)). Both results are close to the theoretical limit of 0.4% and confirm a single transverse mode output from the LMA amplifier.
In conclusion we have shown that large-mode-area fiber amplifiers are well suited for the generation of high power, single-frequency polarized radiation. With a proper choice of the coiling diameter, high order modes can effectively be suppressed and a true single-mode output can be obtained with these fibers. Additionally, a high level of output power and a very high degree of polarization above 200:1 can be maintained. Furthermore, we have shown that conventional M2-measurements are not sensitive to small deviations from the ideal beam shape and are not suited for high-precision measurements of the beam quality.
In addition we have demonstrated, that for high power applications the evolution of the polarization ratio with the amplifier output power has to be monitored in order to achieve a maximum polarization ratio. With a sufficient reduction of the coiling diameter, a simple adjustment procedure of the amplifier by simply optimizing the output power will directly yield optimal results.
The investigations were partially supported by the Deutsche Forschungsgemeinschaft within the Sonderforschungsbereich 407. We gratefully acknowledge the collaboration with the Institut für Physikalische Hochtechnologie (Dr. H.-R. Müller et al.) concerning the supply of the used fibers and the measured refractive index data. Furthermore, we thank our colleague P. Adel for the supply of the program for the calculation of the intensity distribution of the fiber fundamental mode. Finally, we thank the GEO600-Team (B. Willke et al.) for the provision of the ring cavity.
References and links
1. D. Hammons, J. Anderegg, S. Brosnan, H. Komini, and M. Wickham, “High power fiber amplifiers for kilowatt fiber arrays,” in Conference on Lasers and Electro-Optics 2003, Technical Digest (Optical Society of America, Washington, D.C., 2003), Postdeadline paper CThPDB10.
2. I. Zawischa, K. Plamann, C. Fallnich, H. Welling, H. Zellmer, and A. Tünnermann, “All-solid-state neodymium-based single-frequency master-oscillator fiber power-amplifier system emitting 5.5 W of radiation at 1064 nm,” Opt. Lett. 24, 469–471 (1999). [CrossRef]
4. J.P. Koplow, D.A.V. Kiner, and L. Goldberg, “Single-mode operation of a coiled multimode fiber amplifier,” Opt. Lett. 25, 442–444 (2000). [CrossRef]
5. B. Willke, N. Uehara, E.K. Gustafson, R.L. Byer, P.J. King, S.U. Seel, and R.L. Savage, Jr., “Spatial and temporal filtering of a 10-W Nd:YAG laser with a Fabry-Perot ring-cavity premode cleaner,” Opt. Lett. 23, 1704–1706 (1998). [CrossRef]
6. H.M. Pask, R.J. Carman, D.C. Hanna, A.C. Tropper, C.J. Mackechnie, P.R. Barber, and J.M. Dawes, “Ytterbium-Doped Silica Fiber Lasers: Versatile Sources for the 1–1.2 µm Region,” IEEE J. Sel. Top. Quantum Electron. 1, 2–13 (1995). [CrossRef]
7. P. Weßels, M. Auerbach, and C. Fallnich, “Narrow-linewidth master oscillator fiber power amplifier system with very low amplified spontaneous emission,” Opt. Commun. 205, 215–219 (2002). [CrossRef]