Yb:Sc2SiO5 has been investigated in a low-power laser femtosecond oscillator pumped by 400-mW single-mode fiber-coupled diode at 976 nm. Pulses as short as 71 fs were achieved. The same crystal was later employed in a regenerative amplifier, with an output power as high as 4.7 W at 500 kHz and sub-300-fs pulses.
© 2015 Optical Society of America
Among interesting new Yb-doped crystals for ultrashort pulse lasers and high-power amplifiers, Yb:Sc2SiO5 (Yb:SSO) was identified as the most promising candidate belonging to the family of silicates [1,2]. Yb:SSO has a thermal conductivity of ≈70% compared to Yb:YAG, a much wider fluorescence bandwidth of ≈50 nm and a much smaller transparency pump intensity, about 2% compared to Yb:YAG . The crystal is particularly promising for high-power ultrafast sources and amplifiers as a replacement for Yb:YAG; indeed, it has been already investigated in a thin-disk setup  an in a picosecond regenerative amplifer . However, the role of the multi-peaked fluorescence spectrum in view of the full exploitation of the available gain bandwidth for generation of sub-50-fs pulses is not fully understood yet [6,7]. Hence, we decided to test an Yb:SSO crystal in an oscillator pumped by a single-mode fiber-coupled 400-mW laser diode. Please note that this oscillator setup had previously been employed to demonstrate 40-fs pulses with Yb:CALGO  and Yb:CALYO , another promising material with a comparable bandwidth but smoother fluorescence spectrum.
Our experiment shows that the laser has a strong tendency to operate at either one of the two fluorescence peaks, as in many multi-site disordered media with strongly structured emission spectrum. Mode-locking operation with spectral width encompassing a significant fraction of the fluorescence spectrum was not possible; however, pulses as short as 71 fs with spectral bandwidth as large as 17 nm full width half maximum (FWHM) was obtained at 1045 nm and 77 fs with almost the same spectral width at 1065 nm.
The same Yb:SSO crystal was later used as the gain element in a regenerative amplifier, generating up to 4.7 W at 24 W incident pump power. The relatively high doping level of the crystal was not suited to higher pumping power, however the regenerative amplifier performance was promising, and pulses as short as 296 fs were achieved after compression.
2. Oscillator experiments
The pump source was a 400 mW, single-mode fiber-coupled (FC) laser diode (JDSU S27-7602-400) emitting at 976 nm. Considering the optical transmission of the pump lenses and the cavity pump mirror M1 (see Fig. 1), the maximum incident pump power on the Yb:SSO sample was 390 mW. The active medium was 2.91-mm-long, 5%-doped, and antireflection-coated (AR) for both the pump and laser wavelength. The crystal was simply put on a metallic plate and oriented with a small tilt angle with respect to the normal incidence, in order to avoid unwanted etalon effects.
The material exhibits a large absorption peak centered at 976 nm (well matched to the pump wavelength). The emission spectrum shows two distinct, prominent peaks centered approximately around 1036 nm and 1062 nm, respectively .
The X-folded resonator for continuous wave (CW) and continuous wave mode-locked (CW-ML) experiments is depicted in Fig. 1. The FC pump laser diode output was collimated by means of the aspherical lens L1 (f = 15.3 mm, NA = 0.16, AR coated at 976 nm). To focus the pump in the active medium we employed a f = 50 mm, AR coated, spherical lens (L2 in the figure), yielding a pump mode waist in the active medium of radius of ≈11 μm. Given the diffraction limited pump beam quality, and the refractive index of Yb:SSO (n = 1.85), the pump beam confocal parameter was 2zR ≈1.5 mm, about half of the crystal length. In order to minimize the cavity mode astigmatism in the active medium, the angle of incidence on the folding mirrors M1 and M2 were kept as small as possible, about 2.2° owing to mechanical constraints.
In CW and CW-ML experiments, the resonator was operated in the second stability region (larger spacing d2), hence we could control the cavity mode waist on the SESAM by properly adjusting the distance d1 and the curved mirrors separation d2. The cavity arms length were d1 = 175 mm, d2 = 97 mm, and M1-OC = 530 mm, both in CW and CW-ML experiments. By means of ABCD modeling of the resonator we could estimate a fundamental cavity mode radius in the active medium ranging from 12 to 14 μm within the stability region.
The results obtained in CW regime for a set of different output couplers (OCs), are shown in Fig. 2. With the optimum T = 5% OC we obtained up to 231 mW with an absorbed pump power of 375 mW (62% optical-to-optical efficiency), corresponding to a slope efficiency of 69%. A maximum slope efficiency of 72% was measured with the 10% transmission OC. In CW regime, the laser emission occurred at about 1066 nm with the T = 0.8% OC and moved towards 1062 nm with the T = 10% OC.
For the CW-ML experiments, we modified the resonator as depicted in Fig. 1. The HR flat mirror was replaced by a SESAM with 3% modulation loss and 140 μJ/cm2 saturation fluence. From ABCD modeling of the resonator, the cavity mode radius on the SESAM was ≈35 μm. For group delay dispersion (GDD) compensation we used a pair of SF10 dispersive prisms separated by a distance d3 of about 30 cm, corresponding to a maximum negative GDD of about −2500 fs2. In order to force the central output wavelength to move towards the blue side of the emission spectrum we inserted a sharp knife edge close to the second SF10 prism, as shown in Fig. 1. Without the insertion of the knife edge, CW-ML regime was obtained around 1066 nm. Optimizing the cavity alignment and the prisms insertion with T = 0.8% OC we obtained spectra as wide as ~17 nm. The corresponding pulse duration of 77 fs was close to Fourier transform-limit for sech2 shaped pulses (ΔνΔτ = 0.35). In these operating conditions the average output power was about 35 mW. The autocorrelation trace and spectrum of the pulses are shown in Fig. 3(a).
The suppression of strong residual CW components in the spectrum required a careful adjustment of prisms insertion, resonator alignment and pump focus positioning. Once the pure CW-ML regime was obtained it was stable, but not self-starting. Also, the average output power was generally lower when no CW components were present in the output spectrum. In order to force the central output wavelength to move towards the blue side of the emission spectrum, we progressively inserted a knife edge close to the second prism of the SF10 pair (see Fig. 1). The output power dropped to approximately 20 mW, CW-ML regime was still obtained at shorter wavelengths. The shortest pulse duration achieved was 71 fs with a 17 nm FWHM wide optical spectrum, corresponding to a time-bandwidth product ΔνΔτ = 0.33 (see Fig. 3(b)). Also, in this case the suppression of strong residual CW components in the spectrum was difficult. Once the pure CW-ML regime was obtained it was stable, but again not self-starting. It is worth noting that it was not possible to continuously tune the central output wavelength by progressively inserting the knife edge. At a certain moment, the central output wavelength jumped from about 1060 nm to 1040 nm and then it was necessary to re-adjust all the cavity parameters (alignment, prisms insertion, pump focus and crystal position…) to widen the spectrum again.
We also performed experiments employing a single SF10 prism for GDD compensation . With ABCD modeling of the resonator, we found that the virtual prism was positioned about 50 mm from mirror M1, towards the OC. The optimal distance between virtual and real prism was about 270 mm (corresponding to a maximum negative GDD of about −2200 fs2). With the same T = 0.8% OC employed in the previous experiments we obtained almost Fourier transform limited 109 fs-long pulses, with an average power of 45 mW; the corresponding optical spectrum was 11 nm FWHM wide and centered at 1067 nm.
In the single prism setup the laser worked around the 1065 nm emission cross section peak only. Although central output wavelength tuning was possible between 1060 and 1070 nm, the optical spectrum was always narrower than 10 nm, except for a small range around 1067 nm, as mentioned above. Once again it was very difficult to avoid strong CW components in the optical spectrum. This behaviour is in stark contrast to what we observed in low power Yb:CALYO and Yb:CALGO oscillators, where the tuning was much easier across about 40 nm, keeping the pulse width well below 100 fs [9,11].
3. Regenerative amplification experiments
The same crystal as employed for above-mentioned femtosecond oscillator experiments was also tested in a regenerative amplifier. The experimental setup was similar to the one described in , consisting of a Yb:CALGO femtosecond oscillator seeding a diode-pumped regenerative amplifier operating at 500 kHz repetition rate. Pulse stretching occurred inside the cavity owing to normal dispersion of the BBO Pockels cell . Pulse compression after amplification was realized by means of a high efficiency transmission grating with 1250 lines/mm. The seeder was optimized to operate at 1050 nm central output wavelength with ≈90 fs-long pulses at 63 MHz repetition rate, but given the peculiar characteristics of the gain spectrum of Yb:SSO, we tuned the central output wavelength as close as possible to the peaks of the Yb:SSO emission spectrum. The shorter wavelength available from the seeder was about 1045 nm, where it provided 312 fs long pulses with a FWHM bandwidth of 4.2 nm and an average output power of ≈260 mW. The longer wavelength was about 1054 nm, with 191 fs-long pulses (6.9 nm FWHM optical spectrum) and ≈400 mW average output power. It is worth noting that it was not possible to red-shift the central output wavelength closer to the gain peak at 1063 nm, still maintaining a stable mode-locking regime. We estimated a stretched pulse duration after amplification ranging from ~2 to 3 ps, depending on the seed pulse duration.
Preliminarily we tested the system in CW regime. The pump diode could provide up to 120 W at 980 nm, but we did not operate it at the maximum power level in order to prevent damages in the active medium. In fact, the pump power was limited to the onset of significant aberration in the output beam. The pump and laser beam diameter in the active medium were ≈375 µm and 310 µm, respectively. At a maximum incident pump power of 24 W, we obtained up to 5.8 W with a slope efficiency of 34.5% at about 1060 nm with the optimum T = 5% OC. Pump absorption was very high, resulting in 93% under lasing condition.
For pulse amplification experiments, the seeder was initially set to 1045 nm. At 24 W incident pump power and with 100 cavity round trips, we obtained an average output power of about 1.88 W. After compression, we measured 487 fs-long pulses with an 3.2 nm FWHM wide optical spectrum centered at 1042.5 nm (corresponding time-bandwidth product of ΔνΔτ ≈0.43), and an average power of 1.5 W (compressor efficiency ≈80%). Autocorrelation trace and optical spectra of seeder, regenerative amplifier and (non-seeded) cavity-dumped operation are shown in Fig. 4. The output beam quality was very good with a measured M2x × M2y ≈1.17 × 1.13.
Later we tuned the seeder to 1054 nm and investigated the amplifier performance when seeded closer to the second fluorescence spectrum peak. The amplifier efficiency was significantly improved owing to the higher seed power: we obtained up to 4.7 W average power before compression. After compression, we measured 430 fs-long pulses with an optical spectrum of 3.4 nm FWHM, centered at 1062 nm (ΔνΔτ ≈0.39). The compressor efficiency was still about 80%, yielding a maximum average output power of 3.8 W (corresponding pulse energy = 7.6 μJ). The result of the beam quality measurement is shown in Fig. 5(a). Up to the maximum absorbed pump power of about 22 W the close-to-diffraction-limited seed beam quality was well preserved by the amplifier (M2x × M2y = 1.18 × 1.09).
By tuning the central output wavelength of the amplifier towards the seeder central output wavelength we were able to further broaden the pulse spectrum and significantly shorten the compressed pulse duration, at the expense of a reduction in output power. The spectrum and autocorrelation trace of the shortest pulses are shown in Fig. 5b. We measured a minimum pulse duration of 296 fs with an average power of 1.7 W after compression. The corresponding optical spectrum was centered at 1057 nm and was 7.6 nm-wide, yielding a time bandwidth product ΔνΔτ ≈0.60.
We report on our investigation of a Yb:SSO crystal both in a low-power femtosecond oscillator and in a regenerative amplifier. The peculiar shape of the emission spectrum of the material - characterized by a wide and almost flat gain bandwidth containing two strong and relatively narrow peaks at 1036 and 1062 nm - did not permit to fully exploit the fluorescence spectrum in mode-locked regime. Nevertheless, almost Fourier transform-limited pulses with duration well below 100 fs were obtained at central output wavelengths close to either gain peak in a SESAM mode-locked oscillator.
In addition, we present for the first time to the best of our knowledge a femtosecond regenerative amplifier based on Yb:SSO. Even if the combination of length and doping of the available sample was clearly not optimized for high average power amplification, we were able to demonstrate multi-Watt operation at 500 kHz repetition rate with excellent beam quality and sub-500 fs re-compressed pulses at 1054 nm. Even shorter pulses (below 300 fs) were obtained, proving the potential of the material for ultrashort pulses amplification.
References and links
1. S. Campos, A. Denoyer, S. Jandl, B. Viana, D. Vivien, P. Loiseau, and B. Ferrand, “Spectroscopic studies of Yb 3+ -doped rare earth orthosilicate crystals,” J. Phys. Condens. Matter 16(25), 4579–4590 (2004). [CrossRef]
2. R. Gaume, B. Viana, J. Derouet, and D. Vivien, “Spectroscopic properties of Yb-doped scandium based compounds Yb:CaSc2O4, Yb: SrSc2O4 and Yb: Sc2SiO5,” Opt. Mater. 22(2), 107–115 (2003). [CrossRef]
3. L. Zheng, J. Xu, G. Zhao, L. Su, F. Wu, and X. Liang, “Bulk crystal growth and efficient diode-pumped laser performance of Yb3+:Sc2SiO5,” Appl. Phys. B 91(3-4), 443–445 (2008). [CrossRef]
5. J. Li, P. Gao, X. L. Zheng, L. Su, J. Xu, and X. Liang, “Diode-pumped Yb:SSO chirped pulse amplifier with 1 ps pulse duration,” Chin. Opt. Lett. 13(1), 011403 (2015). [CrossRef]
6. W. De Tan, D. Tang, X. Xu, J. Zhang, C. Xu, F. Xu, L. Zheng, L. Su, and J. Xu, “Passive femtosecond mode-locking and cw laser performance of Yb3+:Sc2SiO5,” Opt. Express 18(16), 16739–16744 (2010). [CrossRef] [PubMed]
7. C. Xu, D. Tang, J. Zhang, H. Zhu, X. Xu, L. Zheng, L. Su, and J. Xu, “Sub-100 fs pulse generation in a diode-pumped Yb:Sc2SiO5 laser,” Opt. Commun. 294, 237–240 (2013). [CrossRef]
8. A. Agnesi, A. Greborio, F. Pirzio, G. Reali, J. Aus der Au, and A. Guandalini, “40-fs Yb3+:CaGdAlO4 laser pumped by a single-mode 350-mW laser diode,” Opt. Express 20(9), 10077–10082 (2012). [CrossRef] [PubMed]
9. F. Pirzio, S. D. Di Dio Cafiso, M. Kemnitzer, A. Guandalini, F. Kienle, S. Veronesi, M. Tonelli, J. Aus der Au, and A. Agnesi, “Sub-50 fs widely tunable Yb:CaYAlO4 laser pumped by 400-mW single-mode fiber-coupled laser diode,” Opt. Express 23(8), 9790–9795 (2015). [CrossRef]
11. A. Agnesi, A. Greborio, F. Pirzio, E. Ugolotti, G. Reali, A. Guandalini, and J. Aus der Au, “Diode-pumped passively mode-locked tunable Yb:CALGO solid state laser,” J. Opt. Soc. Am. B 30, 1513–1516 (2013).
12. E. Caracciolo, M. Kemnitzer, A. Guandalini, F. Pirzio, J. Aus der Au, and A. Agnesi, “28-W, 217 fs solid-state Yb:CaGdAlO4 regenerative amplifiers,” Opt. Lett. 38(20), 4131–4133 (2013). [CrossRef] [PubMed]