Open Access
Add to BibSonomy
Abstract
We present continuous wave laser activity in neodymium-doped sapphire ridge waveguides. The ridges were prepared using diamond blade dicing of thin Nd3+:sapphire films grown on sapphire substrates by pulsed laser deposition. Lasing was realized at wavelengths of 1092 nm and 1097 nm for ridge waveguide orientations addressing the σ- and π-polarization, respectively. A maximum slope efficiency with respect to incident pump power of 12% was achieved in σ-polarization with a ridge cross section of 40.8 × 2.6 µm2 and a ridge length of 8 mm. With an available incident pump power of 2.8 W from a Ti:sapphire laser, a maximum output power of 322 mW was realized.
© 2017 Optical Society of America
1. Introduction
A key technology of integrated-optical devices is the fabrication of thin films that allow for waveguiding due to a positive refractive index difference with respect to the substrate and the top layer (air or solid). Lateral structuring, for example via reactive ion etching [1] or fs-structuring [2,3] can further increase the functionality of initially planar waveguides allowing for two-dimensional confinement [4]. Alternatively, ridge waveguides can be prepared using a diamond blade dicing process where a high precision dicing blade with metal bonded diamond grains is used to carve grooves into a waveguiding thin film [5].
Pulsed laser deposition (PLD) is a method that allows for high-quality fabrication of thin waveguiding films. The working principle is to irradiate a ceramic target of the intended film stoichiometry with a high intensity laser pulse in order to create a plasma expanding normal to the target's surface. Ionic matter transported in the plasma plume then condensates on a substrate where a film starts to grow. The composition of the film is governed by the target stoichiometry. PLD has proven to be a versatile method to grow thin films suitable for laser oscillation in waveguide configuration in well-established materials such as garnets [6,7] as well as emerging materials like sesquioxides [8,9]. Recently, 16 W of output power with 70% slope efficiency were achieved in PLD grown Yb:YAG planar waveguides [10], demonstrating the quality of PLD grown crystalline films to be comparable to crystals grown by well-established methods such as the Czochralski technique.
The prospect of Nd3+:α-Al2O3 (Nd3+:sapphire) as a novel laser material was first suggested by [11], who successfully produced such films via molecular beam epitaxy. The main difficulty in growing RE3+:α-Al2O3 film is the large difference of the ionic radii of aluminum (Al3+) and the rare-earth elements (RE3+). In the sixfold-coordinated C3v-site of sapphire the ionic radius of RE3+ ions is about 1 Å, whereas the Al3+ ion exhibits a radius of 0.53 Å. Due to the significantly larger ionic radius of RE3+ ions, their distribution coefficient in α-Al2O3 is extremely low. Hence there is no report on the growth of rare-earth-doped crystalline sapphire from the melt with significant doping concentrations so far. To achieve higher doping concentrations, growth methods apart from the thermal equilibrium like PLD have to be applied.
PLD growth of Nd3+:α-Al2O3 was first reported in 2012 [12]. The very high Nd3+ emission cross-sections of up to 10−18 cm2 in Nd3+:α-Al2O3 and the increase of the refractive index due to the doping motivated the realization of Nd3+ doped waveguide lasers on undoped α-Al2O3 substrates with this material, which was achieved recently in planar waveguide geometry [13]. In this work, diamond blade dicing was applied to the thin films in order to realize lateral confinement in ridge waveguide lasers.
2. Waveguide preparation
Thin films of Nd3+ doped sapphire were epitaxially grown via PLD by deposition on undoped m-cut and a-cut sapphire substrates polished in epitaxial quality with the surfaces corresponding to the (10-10) and (11-20) plane, respectively (see Fig. 1). Sintered (1 at.%) Nd3+:α-Al2O3 ceramics were used as targets. The 25 ns long laser pulses for ablation were supplied by a KrF excimer laser at a wavelength of 248 nm with a fluence of 2.3 J/cm2 on the target. The substrate was mounted 9.65 cm opposite to the target and heated from the back side by an array of infrared laser diodes to a temperature of 1050° C. Stoichiometric material transfer from the target to the substrate was indicated by microprobe analysis. Growth in the intended α-modification of Al2O3 was verified by XRD and RHEED measurements. Waveguiding as well as lasing in planar waveguide geometry was demonstrated in these films recently with a maximum output power of 137 mW, a slope efficiency of 7%, and a threshold pump power of 96 mW in π-polarization [13]. In σ- polarization the maximum output power was 55 mW with a slope efficiency of 4% and a threshold pump power of 590 mW. For both polarizations output power, slope efficiencies and threshold pump powers were given with respect to incident pump power [13]. In order to profit from two dimensional confinement in a ridge waveguide geometry the thin films were processed using a precision dicing saw (model DISCO DAD322). Different dicing blades were tested for the dicing process. The best results concerning chipping and homogeneity of the two trenches carved to form the ridge waveguides were obtained using metal bond blades. Blades with 52 mm diameter were employed and a rotation speed of 30,000 rpm with a vertical velocity of 0.1 mm/s was used. The channels were diced 20 µm deep into the substrate, which is around ten times the thickness of the waveguiding ridges.
Fig. 1 Hexagonal unit cell illustrating the orientation of the substrates and the grown films. Waveguiding occurs with polarizations parallel to the film surfaces. The films have been grown on m-cut (depicted in green) and a-cut (depicted in orange) substrates. The ridge waveguides were cut in a way that their orientation was parallel to the c-axis in the film on m-cut substrates and perpendicular to the c-axis in the a-cut substrates allowing waveguiding for σ-polarization and π-polarization, respectively.
With sapphire exhibiting a Mohs hardness of 9 it proved difficult to initiate the dicing process directly on the edge of the film without deviations in the width of the carved trenches. This problem could be solved by starting the dicing process on uncoated sapphire substrates placed at the polished facets of the film. As shown in Fig. 2 this led to straight cuts with no strong occurrence of chipping at the ridge edges.
Fig. 2 Microscope images of a 40 µm wide ridge prepared using diamond dicing in a 2.6 µm thick Nd(1%):sapphire film. Top view (left), front view (right).
3. Loss measurements
Scattering losses of the ridge-type waveguides were measured via absorption dependent fluorescence measurements as proposed in [14]. This non-destructive technique allows for the calculation of scattering losses independent of the incoupling conditions. The ridges are excited by a Ti:sapphire laser beam coupled in through an end facet. The exponential decay of fluorescence light emitted through the ridges surface is recorded for different excitation wavelengths with a CCD camera. With knowledge of the relative absorption cross sections obtained from excitation spectra, the combined absorption and scattering losses are extrapolated to zero absorption, revealing the pure scattering losses. With this method the scattering losses for different ridge waveguides were estimated to be in the range between 6 dB/cm and 16 dB/cm. As an example, Fig. 3 shows the loss data for ridge waveguides of different widths prepared from a 2.6 µm-thick Nd:sapphire film grown on (10-10) oriented substrate. The scattering losses tentatively increase for smaller ridge widths. This behavior is expected from a classical ray-optical model because more reflections of the sidewalls of the narrower waveguide ridges occur when compared with wider ridges.
Fig. 3 Scattering losses of diamond diced channel waveguides with different widths prepared on 2.6 µm thick Nd(1%):sapphire films on the a-cut substrate.
4. Laser experiments
The laser setup is depicted in Fig. 4. The light of a Ti:sapphire laser set to a wavelength of 833 nm was focused onto the polished front facet of the samples with an aspheric lens having a numerical aperture of 0.26. No mirrors were applied, thus feedback is solely provided by the Fresnel reflections of 7.4% at both the front and end facets of the ridges resulting from the refractive index of 1.75 for sapphire. Therefore, the total output coupling losses of the waveguides are as high as 22.6 dB. Lasing was achieved in the ridges presented in Sec. 2 despite the high scattering losses of up to 16 dB/cm (cf. Fig. 3). The compensation of such high total losses is possible due to the very high small signal gain of more than 1000 dB/cm derived from the maximum emission cross sections of 5 × 10−19 cm2 and 10 × 10−19 cm2 in σ- and π-polarization, respectively [12], at full inversion (1% Nd doping corresponds to a Nd density of 4.71 × 1020 cm−3). The measured laser output power was corrected for the transmission of both the microscope objective as well as the RG1000 filter glas and multiplied by two, assuming symmetrical output for both sides of the waveguide.
Contrary to the planar waveguide laser experiments presented in [13], which yielded higher slope efficiency and higher maximum output power for samples allowing for the π-polarized laser operation, here the best results were achieved with ridge waveguides guiding σ-polarized light.
The results of the laser experiments for 8 mm long ridge waveguides prepared on the (10-10) oriented film are depicted in Fig. 5. With respect to the incident pump power the best results were obtained in a 40.8 µm wide ridge waveguide. In this case, a slope efficiency of 12% at a threshold pump power of 59 mW was observed. At the maximum available incident pump power of 2850 mW an output power of 322 mW was achieved. Compared to the results for planar waveguides presented in [13] the slope efficiency is increased by a factor of three, while the threshold pump power decreased by a factor of two.
Fig. 5 Output power vs. incident pump power in 8 mm long ridge waveguides of different widths dRidge prepared from a 2.6 µm thick Nd(1%):sapphire film grown on the (10-10) oriented substrate. PThr and slope denote threshold pump power and slope efficiency, respectively. Lasing occurred at 1092 nm in σ-polarization.
The set of ridge waveguides prepared from a 2 µm-thick Nd(1%):sapphire film, grown on a (11-20) oriented substrate guiding only π-polarized light, allowed for lasing as well. However, when compared to the results obtained in planar waveguide experiments, a different behavior is observed. One immediately recognizes the strong variations in slope efficiency in the laser results displayed in Fig. 6. The slope efficiencies have been determined with respect to incident power. They range from 0.6% to 6.8% for different widths of the ridges. The best result with a slope efficiency of 6.8% was realized in a 23.5 µm wide ridge waveguide. Lasing started at a threshold pump power of 23 mW. Due to the confinement the threshold pump power is decreased significantly by a factor of 6 when compared to the planar waveguide results [13] (see also Sec. 2). The slope efficiency remained nearly at the same value indicating similar internal laser losses in the planar and ridge waveguides.
Fig. 6 Output power vs. incident pump power in 6 mm long ridge waveguides of different widths dRidge prepared from a 2 µm thick Nd(1%):sapphire film grown on (11-20) oriented sapphire. PThr and slope denote threshold pump power and slope efficiency, respectively. Lasing occurred at a wavelength of 1097 nm in π polarization.
The slope efficiencies for all measured channel waveguides are depicted in Fig. 7. The data feature large scattering and no clear common dependency on the ridge widths, but for all cases the channel waveguides on the (10-10) oriented substrates show the better performance. For the channel waveguides on the (10-20) substrate one could suppose a tendency of decreasing slope efficiencies with increasing ridge widths.Figure 8 displays the σ-polarized laser spectrum and mode profile of the Nd(1%):sapphire ridge waveguide prepared on the (10-10) oriented film. The spectral width of the laser was measured with an optical spectrum analyzer with a spectral resolution of 0.034 nm to be 0.11 nm. The axial mode spacing of the 8 mm long waveguide is about 0.037 nm. Therefore, the measured laser bandwidth is obviously caused by several axial and transversal modes, which could not be resolved. The near-field image taken at the end facet of the waveguide shows the laser mode to be well confined in the ridge waveguide. In the far field, mode diameters of 2.5 cm in horizontal and 4.6 cm in vertical direction were measured 21.5 cm away from the end facet, which corresponds to a half divergent angle of 3.3° in the slow axis and 6.1° in the fast axis, or a numerical aperture of 0.1 and 0.06, respectively.
Fig. 7 Measured slope efficiencies ηSlope vs. ridge widths dridge for all examined channel waveguides on (10-10) and (10-20) oriented substrates.
Fig. 8 Laser spectrum (left) and near field image of the laser mode in a 40 µm-wide ridge (right) of a Nd:sapphire laser in σ-polarization. The white silhouette serves as a guide for the eye.
It has to be noted, that the values for slope efficiency and threshold pump power refer to incident pump power. With the low heights of the ridges of 2.6 µm and 2 µm, focusing with the aspheric lens did not result in an efficient incoupling. In fact, the spatial overlap of the ~10 µm pump spot with the 2.6 µm high ridge waveguide end facets should not allow for more than 50% of incoupling efficiency ηinc. Moreover, despite the high output coupling losses of 22.6 dB the scattering losses >6 dB/cm (≈10 dB per roundtrip in the 8 mm long ridge waveguide) would not allow for resonator extraction efficiencies ηres exceeding 70%. Taking finally into account the Stokes efficiency ηStokes of 75% between the pump and the laser photons, the maximum slope efficiency ηslope, max that could be expected in our setup amounts to
Therefore, the best ridge waveguide performance with a slope efficiency of 12% in Fig. 5 can be understood reasonably well, also taking into account the unknown amount of pump absorption as well as the comparably high doping concentration and the high inversion level in the ridge waveguides, which may give rise to upconversion and/or cross-relaxation losses.5. Conclusions
We have shown that using optimized processing parameters, a diamond blade dicing process can be applied to thin epitaxially grown Nd3+ doped sapphire films without strong chipping, cracking, or separation of the film from the substrate. The achieved two-dimensional optical confinement allowed for superior lasing performance of fabricated ridge waveguides when compared to their planar counterparts. For the sample guiding σ-polarized light, significantly lower lasing thresholds and increased slope efficiencies were found. However, another sample with lower thickness and guiding of the π-polarization did not show the same level of improvement, likely due to inhomogeneous thickness or quality of the initial planar thin film and fabricated ridges. With the poor estimated incoupling efficiency, actual slope efficiencies versus absorbed pump power are expected to be much higher under improved conditions. It is therefore of strong interest to significantly increase the film thickness to some 10 µm in order to validate the potential of Nd:sapphire as well as further rare-earth doped sapphire films as laser gain media.
References
1. P. Nunn, J. Olivares, L. Spadoni, P. Townsend, D. Hole, and B. Lu, “Ion beam enhanced chemical etching of Nd:YAG for optical waveguides,” Nucl. Instrum. Meth. Phys. Res. Sec. B 127–128, 507–511 (1997). [CrossRef]
2. K. M. Davis, K. Miura, N. Sugimoto, and K. Hirao, “Writing waveguides in glass with a femtosecond laser,” Opt. Lett. 21(21), 1729–1731 (1996). [CrossRef] [PubMed]
3. T. Calmano and S. Müller, “Crystalline waveguide lasers in the visible and near-infrared spectral range,” IEEE J. Sel. Top. Quantum Electron. 21(1), 1 (2015). [CrossRef]
4. H. Kogelnik, “An introduction to integrated optics,” IEEE Trans. Microw. Theory Tech. 23(1), 2–16 (1975). [CrossRef]
5. M. F. Volk, S. Suntsov, C. E. Rüter, and D. Kip, “Low loss ridge waveguides in lithium niobate thin films by optical grade diamond blade dicing,” Opt. Express 24(2), 1386–1391 (2016). [CrossRef] [PubMed]
6. C. Grivas, T. C. May-Smith, D. P. Shepherd, and R. W. Eason, “Laser operation of a low loss (0.1 dB/cm) Nd:Gd3Ga5O12 thick (40 µm) planar waveguide grown by pulsed laser deposition,” Opt. Commun. 229(1–6), 355–361 (2004). [CrossRef]
7. J. A. Grant-Jacob, S. J. Beecher, T. L. Parsonage, P. Hua, J. I. Mackenzie, D. P. Shepherd, and R. W. Eason, “An 11.5 W Yb:YAG planar waveguide laser fabricated via pulsed laser deposition,” Opt. Mater. Express 6(1), 91 (2016). [CrossRef]
8. A. Kahn, S. Heinrich, H. Kühn, K. Petermann, J. D. Bradley, K. Wörhoff, M. Pollnau, and G. Huber, “Low threshold monocrystalline Nd:(Gd, Lu)2O3 channel waveguide laser,” Opt. Express 17(6), 4412–4418 (2009). [CrossRef] [PubMed]
9. T. L. Parsonage, S. J. Beecher, A. Choudhary, J. Grant-Jacob, P. Hua, J. I. Mackenzie, D. P. Shepherd, and R. W. Eason, “7 W diode-end-pumped PLD-grown Yb:Lu2O3 planar waveguide laser,” in Advanced Solid State Lasers, OSA Technical Digest (Optical Society of America, 2015), paper AW1A.6.
10. S. J. Beecher, J. A. Grant-Jacob, P. Hua, J. J. Prentice, R. W. Eason, D. P. Shepherd, and J. I. Mackenzie, “Ytterbium-doped-garnet crystal waveguide lasers grown by pulsed laser deposition,” Opt. Mater. Express 7(5), 1628 (2017). [CrossRef]
11. R. Kumaran, S. E. Webster, S. Penson, W. Li, T. Tiedje, P. Wei, and F. Schiettekatte, “Epitaxial neodymium-doped sapphire films, a new active medium for waveguide lasers,” Opt. Lett. 34(21), 3358–3360 (2009). [CrossRef] [PubMed]
12. S. Heinrich, T. Gün, and G. Huber, “Neodymium and ytterbium doped sapphire films grown by pulsed laser deposition,” in AIOM, Talk 1F2A.5, San Diego, USA (2012).
13. S. H. Waeselmann, S. Heinrich, C. Kränkel, and G. Huber, “Lasing of Nd3+ in sapphire,” Laser Photonics Rev. 10(3), 510–516 (2016). [CrossRef]
14. A. Kahn, Y. Kuzminykh, H. Scheife, and G. Huber, “Nondestructive measurement of the propagation losses in active planar waveguides,” J. Opt. Soc. Am. B 24(7), 1571 (2007). [CrossRef]
