Open Access
Add to BibSonomy
Abstract
In this paper, we report on the investigation of two different approaches based on grating waveguide structures (GWS) to generate radially polarized ultra-short pulses in thin-disk laser resonators. In previously reported mode-locked thin-disk oscillators emitting radially polarized pulses, a partly reflective GWS was employed as an outcoupling element for polarization shaping. This partly reflective GWS introduced significant losses for the outcoupled radiation, resulting in a reduced optical efficiency. The aim of the investigations presented here is to explore the possibility of improving the optical efficiency by using a highly reflective GWS in different resonator configurations, which circumvents the loss port present for the partly reflective GWS. The investigations show that using a highly reflective GWS as a folding mirror of the resonator enables a significant improvement of the optical efficiency compared to the previously reported configurations.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Various scientific and industrial applications benefit from the unique properties of axially-symmetric polarized (e.g. radially and azimuthally polarized) laser beams [1–5]. Intra-cavity generation of radially and azimuthally polarized emission was achieved e.g. by using triple-axicon retroreflectors [6,7], exploiting thermally induced birefringence in the laser crystal [8] and by using different grating based optical elements [9–13]. Especially for laser material processing with axially-symmetric polarized laser beams, high average power is required to scale the throughput to an industrially relevant level. In continuous wave (cw) operation, radially polarized radiation with an output power of either 3 kW or 1 kW was obtained from a CO2 [14] laser and from an Yb:YAG thin-disk laser [15], respectively. In both cases, a circular Grating Waveguide Mirror (GWM) was implemented as the end mirror of the resonator to select the radially polarized ring-shaped LG*01 mode [16]. With regard to ultra-short (<10 ps) radially polarized pulses, the highest average power of 1 kW was demonstrated with a two-stage thin-disk multipass amplifier (TDMPA) [17]. However, the excellent power scaling capability of the TDMPA technology comes at the cost of complexity, size of the footprint and costs of the overall laser system. Therefore, scaling the output power of mode-locked thin-disk laser oscillators emitting radially polarized pulses is still of major interest. So far, the maximum output power achieved with a mode-locked thin-disk oscillator emitting radially polarized pulses was 125 W [18] using an approach where a semiconductor saturable absorber mirror (SESAM) [19] was implemented as the resonator’s end mirror to stabilize soliton mode locking. To select the radially polarized LG*01 mode, a partly transmissive Grating Waveguide Output Coupler (GWOC) was used as output coupling element. A significant disadvantage of the GWOCs as used in [18,20] compared to the highly reflective GWMs is the diffraction of a significant amount of the outcoupled power into a higher diffraction order, reducing the power available in the generated output beam. Transmission measurements indicated that about 40% of the outcoupled power is diffracted into a higher diffraction order [18] as illustrated in Fig. 1.
As an alternative, two different approaches that enable the generation of radially polarized pulses in a thin-disk resonator without utilizing a GWOC have been investigated and are reported on in the following. In the first approach, a highly reflective GWM was implemented as the end mirror of the cavity and a semiconductor saturable output coupler (SESOC) was used as the output coupling mirror. A numerical model was employed to assess the potential of the SESOC with regard to scaling of the average power. In the second approach, a standard SESAM was used as cavity’s end mirror and a highly reflective GWM was implemented as folding mirror. The numerical model described in [21] was used to investigate the obtainable laser performance in this configuration taking into account the residual losses of the GWM.
2. Motivation
To investigate the impact of the losses on the laser performance, a thin-disk resonator that allowed for a direct comparison between the implementation of a GWOC or a GWM for polarization shaping was set up. In the first configuration, the GWM with a reflectivity of 99.3% served as the end mirror of the resonator and the beam was coupled out through a dielectric output coupler with a reflectivity of approximately 96%. In the second configuration, the GWM was replaced with a GWOC that exhibits a reflectivity of 95.2% for radial polarization and the dielectric output coupler was replaced with a highly reflective mirror. As can be seen in Fig. 2, (a) much higher optical efficiency can be obtained by using a GWM instead of a GWOC for polarization shaping. In this configuration, the slope efficiency was almost 20 percentage points higher as compared to the configuration with the GWOC as the polarization selective element.
Fig. 2. Comparison between the obtained performance using either a GWM or a GWOC in the same thin-disk resonator. (a) Laser output power versus pump power and (b) optical efficiency versus pump power obtained in cw operation.
In the following, we therefore report on the investigation of two different approaches that allow the use of a GWM instead of a GWOC for polarization shaping in a passively mode-locked thin-disk oscillator emitting radially polarized pulses.
3. Approach 1: semiconductor saturable output coupler combined with a GWM as the end mirror of the resonator
In order to increase the efficiency of mode-locked thin-disk oscillators emitting radially polarized pulses, the approach described in this section aims at replacing the GWOC with a highly reflective GWM as the resonator’s end mirror. Starting and stabilizing of soliton mode locking was enabled by a Semiconductor Saturable Output Coupler (SESOC). The operating principle of a SESOC is similar to that of a SESAM. The sole difference is that the Distributed Bragg Reflector (DBR) of the SESOC is designed to be only partly reflective in order to couple out a part of the incident radiation. Figure 3 shows the principal structure of a SESOC. Since the SESOC cannot be cooled from the backside, the heat is dissipated through the edges of the element. Hence, a strong thermal gradient is expected inside the SESOC leading to thermal lensing and higher-order aberrations for the beam transmitted through the SESOC. In order to estimate the limitations of the SESOC for its use in thin-disk laser resonators, we developed a model which is able to predict the beam quality of the outcoupled radiation as a function of the incident power. In the first part of this section, we describe the numerical model together with the therewith calculated results. In the second part of this section, we present first experimental results using a saturable output coupler in a thin-disk laser resonator. The characteristics of the SESOC used for the experiments are listed in Table 1.
Table 1. Parameters of the SESOC used in the experiment.
3.1 Simulations
There are two different effects leading to a heating of the SESOC. On the one hand, each semiconductor saturable absorber device (SESAM or SESOC) exhibits residual non-saturable losses. The state-of-the-art SESAMs used for the latest records in terms of average output power were reported to have non-saturable losses of 0.1% [22]. On the other hand, in the case of the SESOC, the outcoupled part of the beam propagates through the GaAs substrate, where it experiences linear and non-linear absorption. This effect presents an additional heat source that is not present in the case of the SESAM, where almost 100% of the power is reflected at the DBR.
The temperature field inside the SESOC was calculated by means of a Finite Element Model (FEM) implemented in Comsol Multiphysics (see Fig. 4). The heat source representing the non-saturable losses occurring within the layer structure comprising the SESOC was implemented as a surface heat source on top of the GaAs substrate by taking into account the ring-shaped intensity distribution of the beam oscillating within the resonator as well as the non-saturable losses ${A_{ns}}$ of the SESOC. The DBR was approximated as a thin layer with a thickness of 1.6 µm (corresponding to 10 alternating layers of GaAs/AlAs) and a heat conductivity of 45 W/m·K [23]. Furthermore, the heat generated by linear and non-linear absorption of the outcoupled beam inside the GaAs substrate was implemented as a volume heat source. The supplier of the SESOCs (Batop GmbH) measured a linear absorption coefficient of $\alpha = 1.22\; c{m^{ - 1}}$ and a non-linear absorption coefficient of $\beta = ({31 \pm 9.3} )\frac{{\; cm}}{{GW}}$. The thickness and the diameter of the GaAs substrate was 625 µm and 5 mm, respectively. In order to calculate the beam quality of the beam transmitted through the SESOC, the optical path difference (OPD) induced by the heated SESOC was calculated as shown in [24]. Subsequently, the electric field of an ideal LG*01 [16] mode was multiplied with the calculated OPD as a phase modulation. Finally, the modulated electric field was propagated by means of Fourier optics [25] to different planes to calculate the intensity distribution as well as the beam radius, which, in turn, allowed the calculation of the beam quality factor of the transmitted beam.
Fig. 4. Arrangement considered for the FEM simulation employed to calculate the temperature field inside the SESOC.
In order to estimate the influence of the SESOC on the beam quality of the outcoupled beam, the fluence was fixed to 2500 µJ/cm2 by adapting the beam radius for each power level accordingly. For a saturation fluence of 50 µJ/cm2 (as measured for the available SESOC samples), this fluence corresponds to a typical saturation level used in previous experiments utilizing SESAMs in Yb:YAG thin-disk oscillators [18–20]. The pulse duration was assumed to be 800 fs, which corresponds to a typical pulse duration obtained with SESAM mode-locked thin-disk oscillators.
The results of this simulations assuming different values for the non-saturable losses are shown in Fig. 5. For SESOCs with a very low amount of non-saturable losses of 0.1%, a maximum average output power of approximately 13 W with good beam quality (M2<2.5) can be expected. However, for higher non-saturable losses of 0.5%, the average output power with good beam quality is limited to around 7 W. It is important to note that the maximum achievable power with good beam quality calculated with the numerical model represent an upper limit. Small deviations from the idealized thermal behavior assumed in the simulations will result in a reduced beam quality.
Fig. 5. Beam quality factor of the beam transmitted through the SESOC versus output power calculated for a fluence of 2500 µJ/cm2 at the SESOC and a pulse duration of 800 fs. The beam radius was adapted for each power level to maintain the same fluence.
A further limitation of the SESOC for the laser performance is absorption of the outcoupled beam inside the GaAs substrate. Even the linear part of the absorption amounts to approximately 7% for a substrate thickness of 625 µm. At the desired fluence of 2500 µJ/cm2, the calculated non-linear absorption leads to an increased absorption of 11.8%. Figure 6 shows the overall absorption of the outcoupled beam inside the SESOC as a function of the pulse peak power for a beam radius of 225 µm. The upper abscissa in Fig. 6 shows the corresponding average power assuming a pulse duration of 800 fs and a repetition rate of 52.2 MHz, as obtained in the experiment.
Fig. 6. Calculated overall absorption of the outcoupled beam inside the GaAs substrate as a function of peak power for a beam radius of 225 µm at the SESOC. The upper abscissa shows the laser output power calculated from the pulse peak power for a pulse duration of 800 fs and a repetition rate of 52.2 MHz.
3.2 Experimental results
In order to test the SESOC in laser operation with a radially polarized output, the resonator was set up as shown in Fig. 7. The thin-disk crystal with a thickness of 215 µm and a doping concentration of 7 at. % was glued on a diamond heat sink and mounted in a pumping module allowing for 12 reflections of the pump radiation (940 nm) at the crystal. The diameter of the pumped area was set to 2.4 mm and the resonator was designed such that the beam radius at the SESOC and at the thin-disk crystal was 225 µm and 950 µm, respectively. A Grating Waveguide Mirror (GWM) was implemented as the end mirror of the cavity to favor the radially polarized LG*01 mode. The reflectivity of this GWM was measured to 99.3 ± 0.2% for a radially polarized beam. In cw operation with a standard dielectric output coupler with a reflectivity of 96%, a maximum output power of 56.4 W at an optical efficiency of 46.5% and a beam quality factor of M2<2.3 was obtained.
Fig. 7. Setup of the laser resonator employing a GWM as the end mirror of the resonator and a SESOC as the output coupler.
With the SESOC implemented as the output coupler, self-starting mode locking was observed for output powers exceeding 8 W. Figure 8 shows the output power and optical efficiency versus incident pump power. The maximum optical efficiency of 19.5% was observed at an output power of around 10 W. With increasing output power, the optical efficiency dropped by roughly 4 percentage points. This drop of the optical efficiency can be attributed to an increased absorption caused by non-linear absorption of the outcoupled beam inside the GaAs substrate.
Fig. 8. Output power and optical efficiency as a function of pump power obtained in mode-locked operation.
Figure 9 (a) shows the beam quality factor measured for different output powers and the far-field intensity distributions as insets. The measurement uncertainty of the beam profiling system (DataRay WinCam) used for this experiment is specified to be ±10%, illustrated by the error bars in Fig. 9 (a). At an output power of around 2 W (cw operation), the beam quality was close to diffraction limited and the far-field intensity distribution was ring-shaped. At the threshold to stable cw mode locking (8 W of output power), the beam quality degraded to M2=2.65 and the far-field intensity distribution showed noticeable distortions. At the highest output power of 12.2 W, a highly degraded beam quality of around M2=4.5 was measured as the far-field intensity distribution exhibited multiple diffraction rings. Furthermore, Fig. 9 (a) shows the beam quality factor calculated with the numerical model assuming a constant beam radius of 225 µm at the SESOC. At the maximum output power of 12.2 W, the discrepancy between the measured beam quality factor and the calculated beam quality factor was approximately 25%. This discrepancy can be attributed to slightly higher non-saturable losses than specified for this SESOC and worse thermal properties than assumed in the numerical model. The polarization was analyzed qualitatively by inserting a rotatable polarizer into the beam path. Despite the distorted intensity profiles observed at higher powers, well separated lobes aligned along the transmission axis of the polarizer can be observed in Fig. 9 (c) and Fig. 9 (d), indicating a high degree of radial polarization.
Fig. 9. (a) Beam quality factor M2 versus output power and far-field intensity distributions as insets obtained in mode-locked operation (black symbols). The red symbols show the beam quality factor calculated with the numerical model. (b) Beam caustic measured at the maximum output power of 12.2 W. (c) Qualitative polarization analysis at an output power of 2 W and (d) qualitative polarization analysis at the maximum output power (the orientation of the transmission axis of the polarizer is indicated by the arrows). (e) Intensity distribution of the intra-cavity beam measured behind a folding mirror.
It is important to note that, as shown in Fig. 9 (e), the intra-cavity beam (observed behind a folding mirror) did not show any signs of aberrations for all power levels. This fact shows that the degradation of the beam quality is introduced by the transmission through the SESOC and should be mitigated by an improved design of this element as discussed further below.
Figure 10 summarizes the properties of the pulses at the maximum output power of 12.2 W. Close to transform-limited pulses with a duration of 786 fs and a spectral bandwidth of 1.94 nm were obtained at a repetition rate of 52.2 MHz. The corresponding pulse energy and pulse peak power was 238 nJ and 257 kW, respectively.
Fig. 10. Pulse diagnostics at 12.2 W of output power. (a) Autocorrelation trace, (b) spectral intensity, (c) pulse train and (d) radio frequency signal measured with a span of 45 kHz and a resolution bandwidth (RBW) of 30 Hz.
To summarize, the numerical model indicates that the output power with good beam quality (M2<2.5) that can be obtained with the presently available SESOCs in a thin-disk resonator is limited to approximately 6-10 W, depending on the non-saturable losses. A first experimental test of a SESOC as the output coupler confirmed these limitations. In fact, stable soliton mode locking with a beam quality of M2=2.65 was demonstrated at an output power of 8 W. For higher output powers, the beam quality degraded significantly. To scale the output power of radially polarized thin-disk lasers with a SESOC, the thermal management of this element therefore requires further improvements.
3.3 Outlook: improvement of the thermo-optical properties of the SESOCs
The experiments showed that the outcoupled beam transmitted through the SESOC suffers from aberrations, leading to a significant degradation of the beam quality. However, the simulations show that two measures may be followed to significantly improve the thermo-optical properties of the SESOCs. First, the heat generated in the absorber section at the intra-cavity surface of the SESOC should be removed by a heat spreader. Second, the overall absorption of the outcoupled beam inside the GaAs substrate should be reduced by reducing the thickness of the substrate. The influence of both measures was theoretically investigated by means of the aforementioned model. A single-crystal diamond heat spreader with a thickness of 500 µm was assumed at the intra-cavity side of the SESOC as shown by Fig. 11. At the same time, the thickness of the GaAs substrate was assumed to be reduced to 200 µm. To take into account the heat barrier between the diamond heat spreader and the SESOC caused by the dielectric anti-reflection coatings, a thin layer with a thickness of 1 µm and a heat conductivity of 1.4 W/m·K was implemented in the FEM model.
Fig. 11. SESOC with the suggested 500 µm thick single-crystal diamond heat spreader at the intracavity side and a reduced thickness of the GaAs substrate.
The beam quality calculated for a beam transmitted through this optimized SESOC is shown in Fig. 12. Compared to the standard SESOC, the beam quality can be significantly improved. Assuming a non-saturable absorption of 0.3%, the model predicts an output power as high as 150 W with good beam quality (M2<2.5). Furthermore, the overall absorption calculated for this improved SESOC at a fluence of 2500 µJ/cm2 was reduced to 4%.
Fig. 12. Calculated beam quality versus laser output power calculated for an optimized SESOC with a diamond heat spreader on top and a thinner GaAs substrate with a thickness of 200 µm. The fluence on the SESOC was fixed to 2500 µJ/cm2 and the pulse duration was assumed to be 800 fs.
To summarize, one promising approach to improve the thermal management of the SESOC was analyzed using the numerical model. In this approach, a diamond heat spreader is proposed to be mounted at the intra-cavity side of the SESOC to dissipate the heat generated in the absorber section and a reduction of the thickness of the substrate is proposed to reduce the absorption of the outcoupled beam inside the substrate. The numerical model predicts that with such an optimized SESOC an output power in excess of 100 W with good beam quality is feasible.
An alternative approach is presented in the following section.
4. Approach 2: grating waveguide mirror as folding element
Another approach that allows the generation of radially polarized pulses in a thin-disk resonator without using a GWOC is presented in this section. To start and stabilize soliton mode locking, a standard SESAM was implemented as the end mirror of the resonator. Polarization shaping was achieved by implementing the same GWM used for the experiment presented in the previous section as a folding mirror of the resonator. The reflectivity of the GWM was measured to be 99.2 ± 0.2% for a radially polarized beam and an angle of incidence (AOI) of 0.75°, which is slightly lower than for normal incidence (99.3 ± 0.2%). This difference of the measured reflectivities is within the measurement uncertainty. To minimize the losses and to avoid polarization distortions, the resonator was therefore designed to allow for an AOI of <0.5° for the beam incident on the GWM. Furthermore, the oscillating beam experiences the residual losses of the GWM twice per resonator round trip. Consequently, the residual losses of the GWM are more critical when it is operated as a folding element of the resonator. In order to assess the benefits of this configuration with regard to the optical efficiency of the laser, a resonator that allowed a direct comparison of different configurations in cw operation was set up as shown in Fig. 13. In configuration 1, which was analyzed as a reference, the GWM served as the resonator’s end mirror at position 1. The beam was coupled out by a standard dielectric output coupler (OC) with a reflectivity of 96% at position 3. In configuration 2, the optical elements at position 1 (GWM) and position 2 (HR mirror) were swapped to test the GWM as a folding mirror. Finally, in configuration 3 both the output coupling and the polarization shaping was performed with a GWOC at position 1. In this configuration, plane HR mirrors were placed at the positions 2 and 3. The reflectivity of the GWOC was measured to be 95.2% for radial polarization.
The performances of the three setups are compared in Fig. 14. The best performance with a slope efficiency of 58% was obtained with the GWM as the resonator’s end mirror. With the GWM used as a folding mirror, the slope efficiency dropped to 47%. Using a GWOC (configuration 3) resulted in a slope efficiency of only 38%. These experiments showed that using this GWM as a folding element enables a significantly more efficient laser operation as compared to the configuration using a GWOC. However, the residual losses of the GWM implemented as a folding mirror have a significantly larger impact on the laser performance compared the configuration where the GWM is used as the end mirror of the resonator. To assess this approach from a more general point of view, it is important to know how the laser performance obtainable in this configuration depends on the residual losses of the GWM. This was investigated using a numerical model [21] capable of calculating the output power and the optical efficiency of the oscillating LG*01 mode to fit the experimental results.
Fig. 14. Output power (a) and optical efficiency (b) versus incident pump power for three different configurations in cw operation. The round symbols and the solids lines represent the measured values and the calculated values, respectively.
The pump light absorption as well as the reflectivity of the folding mirrors were adjusted in the simulation in order to match the calculated values of the output power to the ones measured with the GWM as the end mirror. A pump light absorption of 99% and a reflectivity of the resonator’s folding mirrors of 99.95% yielded a good agreement between the measured and the calculated output power (blue line in Fig. 14). These values of the two fitting parameters (pump light absorption and reflectivity of the resonator’s folding mirrors) were fixed for all further simulations. In configuration 2, where the GWM was implemented as a folding mirror, the AOI of the oscillating beam at the GWM was approximately 0.4° and therefore the reflectivity of this element was slightly reduced. Assuming a reflectivity of 99.2% for the simulation yielded a good agreement between measured and calculated output power (red solid line in Fig. 14). To fit the simulation to the measurement with the GWOC as the output coupler, the diffraction losses experienced by the outcoupled beam were set to 42% (black solid line in Fig. 14). Table 2 lists all the parameters used for the simulations.
Table 2. Parameters used for the simulations.
The impact of the residual losses of the GWM on the laser performance was investigated by varying the reflectivity of the GWM in the numerical model with the aforementioned values of the fitting parameters. Figure 15 shows the calculated slope efficiency as a function of the reflectivity of the GWM used as the resonator’s end mirror or as a folding mirror, respectively. Using a GWM with a reflectivity of 99.8% (as reported in [15,26]) as a folding mirror enables a slope efficiency >60%. With this reflectivity, the difference to the configuration where the GWM is used as the resonator’s end mirror is minor (2.6 percentage points). With decreasing reflectivity of the GWM, however, the difference of the slope efficiency obtainable with the two configurations increases. For a reflectivity <98.4%, the slope efficiency obtainable with the GWM as folding mirror is <40%, which is comparable to the slope efficiency of 38% obtainable with the currently available GWOCs. To summarize, for the outcoupling rate of the dielectric output coupler (4%) and of the GWOC (4.8%) used in the experiment and in the simulations, the laser performance can be significantly increased by using a GWM as a folding mirror compared to a GWOC, provided that the reflectivity of the GWM is >98.4%. For reflectivities of the GWM of <98.4%, the GWOC yields a better laser performance.
Fig. 15. Slope efficiency calculated for GWM as folding mirror and GWM as end mirror for different reflectivities of the GWM.
To demonstrate mode-locked operation in the configuration with the GWM as a folding mirror, the resonator setup that was previously used for the cw experiments was adapted slightly (see Fig. 16). The plane dielectric end mirror at position 1 (Fig. 13) was replaced with a concave mirror to focus the beam onto the SESAM which served as the resonator’s end mirror. The nominal beam radius on the SESAM was 307 µm. To compensate for the non-linear phase shift accumulated by the pulse oscillating in the cavity, three dispersive Gires Tournois Interferometer (GTI) type mirrors introducing a total negative group delay dispersion of −5400 fs2 per roundtrip were implemented as folding mirrors. The SESAM had a modulation depth of 0.64%, a saturation fluence of 60 µJ/cm2 and non-saturable losses of approximately 1.4%.
Figure 17 shows the output power and optical efficiency versus pump power. Self-starting soliton mode locking was observed above an output power of around 7 W and a maximum output power of 14.8 W was attained. The optical efficiency measured at the maximum output power was 23% and the slope efficiency was 32.5%.
Fig. 17. Output power and optical efficiency obtained in mode-locked operation with the GWM implemented as folding mirror.
The pulse duration at maximum power was 755 fs and the spectral bandwidth was 1.6 nm, indicating almost transform limited pulses (Fig. 18 (a) and (b)) with a time-bandwidth product of 0.34. Figure 18 shows the pulse diagnostics measured at the maximum output power.
Fig. 18. Pulse diagnostic measured at maximum output power. (a) Intensity autocorrelation, (b) spectral intensity, (c) pulse train recorded with a fast photodiode and an oscilloscope and (c) radio frequency signal measured with a span of 45 kHz and a resolution bandwidth (RBW) of 30 Hz.
The beam quality at the maximum power was close to diffraction limited (M2=2.1 along the horizontal and vertical direction) and the intensity distribution showed no signs of aberrations, as can be seen in Fig. 19 (a). The qualitative polarization analysis shown in Fig. 19 (b)-(d) revealed well separated intensity lobes aligned along the transmission axis of a polarizer, indicating a high polarization purity.
Fig. 19. (a) Far-field intensity distribution and (b)-(d) far-field intensity distribution recorded behind a polarizer with the transmission axis indicated by the white arrows obtained in mode-locked operation.
5. Conclusion
In this paper, two different approaches to generate radially polarized ultra-short pulses in a thin-disk laser resonator were investigated. The aim of these investigations was to optimize the optical efficiency of the laser by replacing the GWOC, which was implemented in the previously reported thin-disk oscillators to generate radially polarized pulses. The first approach used a saturable output coupler and a highly reflective GWM as the resonator’s end mirror. Simulations and experiments showed that thermo-optical effects inside the substrate of the SESOC limit the output power to a few Watts. However, according to our numerical investigations, the thermal properties of the SESOCs can be significantly improved by adding a diamond heat spreader and by utilizing a thinner substrate. The numerical model predicts output powers in excess of 100 W with good beam quality for the improved SESOCs.
The second approach employed a standard SESAM as resonator end mirror and a highly reflective GWM as folding element. In this configuration, residual losses of the GWMs have a significantly stronger effect on the laser performance. However, the corresponding simulations predict an improved laser performance, provided that the GWM’s reflectivity is >98.4%. With the available GWM and GWOC, an increase of the optical slope efficiency of 9 percentage points was demonstrated in cw operation using the GWM as folding mirror. However, an improved GWM with a reflectivity >99.8% used as folding mirror will enable an optical slope efficiency larger by >20 percentage points compared to the configuration using a GWOC. In a first proof-of-principle experiment, radially polarized pulses with an average power of 14.8 W were obtained from a thin-disk oscillator using a GWM as folding mirror. The beam quality obtained in this configuration was close to diffraction limited.
Further investigations will be devoted to the development of more efficient GWMs, i.e. with a reflectivity as high as 99.8% and the implementation of the single crystal diamond heat spreader to the SOC for further power scaling of radially polarized mode-locked thin-disk oscillators.
Funding
Bundesministerium für Wirtschaft und Technologie.
Acknowledgments
This work was part of the research project Ripple (ZF4592401DF8) which was funded by the BMWi (Bundesministerium für Wirtschaft und Energie) within the ZIM (Zentrales Innovationsprogramm Mittelstand) framework.
Disclosures
The authors declare no conflicts of interest.
References
1. Q. Zhan, “Cylindrical vector beams: from mathematical concepts to applications,” Adv. Opt. Photonics 1(1), 1 (2009). [CrossRef]
2. M. Kraus, M. A. Ahmed, A. Michalowski, A. Voss, R. Weber, and T. Graf, “Microdrilling in steel using ultrashort pulsed laser beams with radial and azimuthal polarization,” Opt. Express 18(21), 22305 (2010). [CrossRef]
3. R. Weber, A. Michalowski, M. Abdou-Ahmed, V. Onuseit, V. Rominger, M. Kraus, and T. Graf, “Effects of Radial and Tangential Polarization in Laser Material Processing,” Phys. Procedia 12, 21–30 (2011). [CrossRef]
4. V. G. Niziev and A. V. Nesterov, “Influence of beam polarization on laser cutting efficiency,” J. Phys. D: Appl. Phys. 32(13), 1455–1461 (1999). [CrossRef]
5. E. Skoulas, A. Manousaki, C. Fotakis, and E. Stratakis, “Biomimetic surface structuring using cylindrical vector femtosecond laser beams,” Sci. Rep. 7(1), 45114–11 (2017). [CrossRef]
6. M. Endo, “Azimuthally polarized 1 kW CO2 laser with a triple-axicon retroreflector optical resonator,” Opt. Lett. 33(15), 1771 (2008). [CrossRef]
7. B. Li, Y. Hu, Y. Hu, and J. Zhao, “1.5 kW radially polarized beam irradiated from a FAF CO2 laser based on an intracavity triple-axicon retroreflector and quarter wave phase retarders,” Appl. Opt. 56(12), 3383 (2017). [CrossRef]
8. I. Moshe, S. Jackel, and A. Meir, “Production of radially or azimuthally polarized beams in solid-state lasers and the elimination of thermally induced birefringence effects,” Opt. Lett. 28(10), 807–809 (2003). [CrossRef]
9. T. Kämpfe, S. Tonchev, A. V. Tishchenko, D. Gergov, and O. Parriaux, “Azimuthally polarized laser mode generation by multilayer mirror with wideband grating-induced TM leakage in the TE stopband,” Opt. Express 20(5), 5392 (2012). [CrossRef]
10. T. Moser, J. Balmer, D. Delbeke, P. Muys, S. Verstuyft, and R. Baets, “Intracavity generation of radially polarized CO2 laser beams based on a simple binary dielectric diffraction grating,” Appl. Opt. 45(33), 8517–8522 (2006). [CrossRef]
11. M. Rumpel, M. Haefner, T. Schoder, C. Pruss, A. Voss, W. Osten, M. A. Ahmed, and T. Graf, “Circular grating waveguide structures for intracavity generation of azimuthal polarization in a thin-disk laser,” Opt. Lett. 37(10), 1763 (2012). [CrossRef]
12. M. A. Ahmed, A. Voss, M. M. Vogel, and T. Graf, “Multilayer polarizing grating mirror used for the generation of radial polarization in Yb:YAG thin-disk lasers,” Opt. Lett. 32(22), 3272 (2007). [CrossRef]
13. D. Lin, J. M. O. Daniel, M. Gecevičius, M. Beresna, P. G. Kazansky, and W. A. Clarkson, “Cladding-pumped ytterbium-doped fiber laser with radially polarized output,” Opt. Lett. 39(18), 5359 (2014). [CrossRef]
14. M. A. Ahmed, J. Schulz, A. Voss, O. Parriaux, J.-C. Pommier, and T. Graf, “Radially polarized 3 kW beam from a CO_2 laser with an intracavity resonant grating mirror,” Opt. Lett. 32(13), 1824 (2007). [CrossRef]
15. T. Dietrich, M. Rumpel, F. Beirow, C. M. Mateo, C. Pruss, W. Osten, M. Abdou Ahmed, and T. Graf, “Thin-disk oscillator delivering radially polarized beams with up to 980 W of CW output power,” Opt. Lett. 43(6), 1371 (2018). [CrossRef]
16. W. Koechner, Solid-State Laser Engineering (Springer, 2006).
17. A. Loescher, J. Negel, C. Röcker, F. Beirow, and T. Graf, “Thin-disk multipass amplifier delivering radially polarized ultrafast pulses with an average output power of 1 kW,” in Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference (CLEO/Europe-EQEC) (2019), paper ca_5_5.
18. F. Beirow, M. Eckerle, B. Dannecker, T. Dietrich, M. A. Ahmed, and T. Graf, “Radially polarized passively mode-locked thin-disk laser oscillator emitting sub-picosecond pulses with an average output power exceeding the 100 W level,” Opt. Express 26(4), 4401 (2018). [CrossRef]
19. C. J. Saraceno, C. Schriber, M. Mangold, M. Hoffmann, O. H. Heckl, C. R. Baer, M. Golling, T. Südmeyer, and U. Keller, “SESAMs for High-Power Oscillators: Design Guidelines and Damage Thresholds,” IEEE J. Sel. Top. Quantum Electron. 18(1), 29–41 (2012). [CrossRef]
20. M. Eckerle, T. Dietrich, F. Schaal, C. Pruss, W. Osten, M. A. Ahmed, and T. Graf, “Novel thin-disk oscillator concept for the generation of radially polarized femtosecond laser pulses,” Opt. Lett. 41(7), 1680 (2016). [CrossRef]
21. P. Wittmuess, S. Piehler, T. Dietrich, M. A. Ahmed, T. Graf, and O. Sawodny, “Numerical modeling of multimode laser resonators,” J. Opt. Soc. Am. B 33(11), 2278 (2016). [CrossRef]
22. F. Saltarelli, I. J. Graumann, L. Lang, D. Bauer, C. R. Phillips, and U. Keller, “Power scaling of ultrafast oscillators: 350-W average-power sub-picosecond thin-disk laser,” Opt. Express 27(22), 31465 (2019). [CrossRef]
23. P. Zhang, M. Jiang, R. Zhue, D. Zhang, and Y. Song, “Thermal conductivity of GaAs/AlAs distributed Bragg reflectors in semiconductor disk laser: comparison of molecular dynamics simulation and analytic methods,” Appl. Opt. 56(15), 4537 (2017). [CrossRef]
24. S. Piehler, C. Thiel, A. Voss, M. Abdou Ahmed, and T. Graf, “Self-compensation of thermal lensing in optics for high-brightness solid-state lasers,” in High Power Laser Materials Processing: Lasers, Beam Delivery, Diagnostics, and Applications, E. Beyer and T. Morris, eds. (SPIE, 2012), 8239, p. 82390Z.
25. A. E. Siegman, Lasers (University Science Books, 1986).
26. M. A. Ahmed, M. Haefner, M. Vogel, C. Pruss, A. Voss, W. Osten, and T. Graf, “High-power radially polarized Yb:YAG thin-disk laser with high efficiency,” Opt. Express 19(6), 5093 (2011). [CrossRef]
