1. Introduction

Optical waveguides have been fabricated in a variety of materials for different purposes and applications. One of these applications is for the production of waveguide lasers and amplifiers. For this particular purpose the waveguide geometry allows greater pump intensity-length products than those that are obtained in a bulk crystal, which means that high optical gains per unit pump power can be obtained. In the case of optical waveguides made in crystals doped with rare earth ions, the high power densities result in lower laser thresholds and provide a great potential for the development of compact and efficient lasers and amplifiers for integrated optics.

Ion beam implantation has become a powerful technique to fabricate waveguides and has been proved in a variety of materials [1]. The damage caused by nuclear collisions during the implantation reduces the physical density of the crystal, which results in a reduction of the refractive index. The low density buried layer that is produced at the end of the ion track acts as an “optical barrier” which has a lower refractive index than the substrate. The region between this barrier and the surface is surrounded by regions of lower index and can act as a waveguide [2]. The different parameters of the implantation process are the specific projectile ion, ion energy, total dose and the angle at which the implantation is done. The manipulation of these parameters allows the fabrication of waveguides with a variety of properties.

Nd:YAG is a well-known bulk laser material and was the first material to be used as substrate to form an optical waveguide by ion implantation [3]. Previous work on ion-implanted Nd:YAG planar waveguide lasers reported that the projectiles used were He⁺ ions [4,5]. Subsequent work presented the formation of Nd:YAG waveguides by multi-implant of protons with energies close to 1 MeV [6], as well as single-implant of carbon ions with an energy of 7 MeV, showing that the luminescence from the waveguides is coincident with the bulk crystal [7]. Initial results of laser emission from these waveguides were described in Ref. [8]. Also the fabrication of channel waveguides in Nd:YAG crystals by using different implantation angles has been reported [9]. In the present work we report the performance at 1.06 μm of Nd:YAG waveguide lasers fabricated by proton implantation at different implantation angles and doses, and the main laser characteristics of these waveguides, such as slope efficiency and threshold power, are also presented.

2. Experimental methods

Planar waveguides were fabricated at room temperature on Nd:YAG substrates by the ion implantation technique at the Instituto de Física (UNAM) 9SDH-2 Pelletron Accelerator, using protons of energies around 1 MeV. The angle of implantation was varied in order to explore the waveguide size and the characteristics of the optical barriers formed.

Multi-implants were done in one sample in order to obtain several thin barriers very close between each other in a way that they all can be considered as a single broad barrier. The energies used were 1.0, 1.05, 1.1 and 1.15 MeV with a total dose of 5×10¹⁶ ions/cm². The implants were performed at an angle of 30° in order to produce a thinner waveguide than that made at normal incidence [6]. A second sample was implanted with protons at an energy of 1.0 MeV, angle of 60° and at a total dose of 2×10¹⁶ ions/cm². Both samples have a length of 1 cm.

The optical characterization was performed by the standard m-line technique [6], from which the propagation modes and the related index profiles were obtained. For this purpose, a rutile prism was used to couple polarized light from a He-Ne laser (632.8 nm) into the waveguide.

The absorption losses produced by implantation were diminished by annealing the samples in an electrical furnace at 400 °C during 30 minutes in open atmosphere. In the experimental setup used to measure the losses, the light provided by a solid state laser (635 nm) was coupled into the waveguide by means of an optical fiber and coupled out by a microscope objective (10x), the output light was measured focusing it on an integrating sphere detector and using an optical power meter.

To test the lasing properties of the waveguides, a CW Ti:Sapphire laser with a tuning range between 750–850 nm, was used as the excitation source. The waveguide laser cavity was formed by butting planar mirrors to the polished end-faces of the planar waveguides as sketched in Fig. 1. A 97.5 % reflectivity mirror at 1064 nm and transmission of 56.3 % at 809 nm was placed in the front face as the high reflector, while in the other face a 98.1 % at 1064 nm and 62.5 % at 809 nm reflecting mirror was used as the output coupler. The pump beam from the Ti:Sapphire source was coupled into the waveguide with a 10X microscope objective by the end-fire coupling technique. The output light was collected through a 20X microscope objective. In order to record the laser spectrum, the output light was directed through an entrance fiber to an Optical Spectrum Analyzer (HP 70951A) with a spectral range between 600–1700 nm. The output power at 1064 nm and the input pump power were measured with two power meters.

 

Fig. 1. (a) Experimental setup used to produce laser oscillation in waveguides. (b) Photograph of laser cavity during laser performance.

Download Full Size | PDF

3. Results and discussion

According to Transport of Ions in Matter (TRIM) calculations [10], the ion range distributions for 1.0–1.15 MeV are located around 8–11 μm for an implant angle of 30° (sample 1), shown in Fig. 2. The purpose of this multi-implant strategy was to prevent tunneling losses and to improve waveguide confinement, in order to increase internal power and therefore reduce the laser threshold.

Moreover, from TRIM calculations a greater angle of implantation produces broader barriers than those obtained with smaller angles, although at the expense of reduced barrier height, the ion range being of around 5 μm with an implant angle of 60° for the second sample. The expected effect of this implant was then to reduce the size of the waveguide (see Fig. 2), and in addition, to generate a broader barrier than that produced at a smaller angle with the same dose.

The propagation modes of each waveguide were obtained by the standard m-line technique and the corresponding index profile was calculated by fitting it to the observed modes [11] and to optical microscopy measurements of the width and depth of the barrier [12]. The results are shown in Fig. 3 where the barrier width of sample 1 is greater than that of sample 2, as was expected. We found that waveguide 1 presents an optical barrier height (decrease in refractive index relative to the substrate) of ~ 0.6 % and a depth of about 9 μm; waveguide 2 shows an optical barrier with a height of approximately ~ 0.27 % and a depth of about 4 μm. These differences are consequences of the different implantation conditions as expected: the depth of the barriers is a result of the distinct angles at which the implants were done and the variations in barrier height are caused both by the different total doses and the implantation angles used.

 

Fig. 2. Depth distributions of the protons implanted in Nd:YAG substrates obtained from TRIM calculations. Sample 1: multi-implant of protons with energies between 1.0–1.15 MeV at an angle of 30° and total dose of 5×10¹⁶ ions/cm² (continuous line). Sample 2: protons with an energy of 1 MeV at an angle of 60° and total dose of 2×10¹⁶ ions/cm² (dotted line).

Download Full Size | PDF

 

Fig. 3. Waveguide refractive index profiles fitted to reproduce the experimental modes (symbols). Waveguide 1 corresponds to the continuous line and waveguide 2 to the dotted line.

Download Full Size | PDF

When waveguides are formed by ion implantation, color centers are generated during the process, which implies absorption losses in the waveguide [1]. Thus, an annealing step is necessary in order to reduce these losses. In order to determine the reduction in waveguide losses, the transmission intensity was measured. After annealing at 400°C during 30 minutes, the losses measured from the waveguides 1 and 2 were 5.3 and 5.0 dB/cm, respectively. These values correspond to the total waveguides losses (i.e., launch efficiency, tunneling, scattering and mode coupling).

We proceeded to analyze the main characteristics of laser performance. The wavelength used to pump the waveguides was 808.6 nm, this line was selected because neodymium ions in YAG present the best absorption bands around 810 nm, shown in Fig. 4, and after optimizing the laser output at 1064 nm the 808.6 line nm was selected. Employing the broad spectrum of the Ti:Sapphire laser in pulsed mode (around 40 nm), it was possible to observe the waveguide absorption bands, and at the wavelength of 808.6 nm the waveguide exhibits the best absorption in coincidence with the excitation range, this line corresponds to a transition between the energy levels ⁴I_9/2 → ⁴H_9/2: ⁴F_5/2 of the neodymium ions in YAG hosts.

 

Fig. 4. Excitation range at 1064 nm of the Nd:YAG waveguide 1 where the bands corresponding to the energy level transitions ⁴I_9/2 → ⁴F_7/2 and ⁴I_9/2 → ⁴F_5/2 are clearly observed.

Download Full Size | PDF

Figure 5 shows the recorded spectrum of the laser emission from sample 1 for pumping at ~ 808.6 nm, it was recorded using the maximum resolution of the optical spectrum analyzer which is 0.08 nm. A narrow band centered at ~1064 nm with a full width at half maximum (FWHM) of ~ 0.17 nm is observed. This laser emission corresponds in fact to the maximum gain transition of the Nd³⁺ ions in YAG crystals (⁴F_3/2 → ⁴I_11/2). The spectrum also shows small emission bands with peaks located at 1063.89, 1063.95 and 1064.04 nm which can be attributed to the cavity modes (longitudinal modes). The separation between adjacent longitudinal modes is given by the distance between mirrors and the refractive index of the material inside of the cavity [13]. In the case of Nd:YAG crystals and the cavity used, the calculation for this separation gives 0.03 nm approximately.

 

Fig. 5. Recorded laser emission spectrum from waveguide 2. In the insets we show a photograph of the laser output at 1064 nm and a zoom the same spectrum.

Download Full Size | PDF

Figure 6 shows the slope efficiency curve for both samples. From this curve we obtained the maximum output power, the laser threshold and the slope efficiency of the laser. For waveguide 1 we obtained up to 14.2 mW of signal at 1064 nm, a laser threshold of about 50 mW with a slope efficiency of ~ 18.1 % relative to the absorbed pump power. For waveguide 2 the maximum output power was 16.1 mW at 1064 nm, the laser threshold was approximately 50.8 mW and the slope efficiency about 21.38 %. The absorbed pump power was calculated taking into account the microscope objective and entrance mirror transmittances, as well as the light coupling efficiency into the waveguide. Given the absorption coefficient at the pump wavelength and the waveguide length, it is safe to assume that all the pump power launched is absorbed in the waveguide (αL~4 at the absorption peak).

 

Fig. 6. Slope efficiency curves for both samples: (a) Waveguide 1, (b) Waveguide 2. The very similar behavior in power threshold and slope efficiency between the waveguides can be noted.

Download Full Size | PDF

The use of protons has several advantages in the design of ion implanted planar waveguides when we compare the results with those obtained when He⁺ ions are used. One advantage is the reduced deleterious effects on the optical properties of Nd:YAG crystals [6], another advantage arises from the fact that protons are the lightest ions, permitting them to reach greater depths than other ions for a given energy, and thus making possible the development of deeper waveguides [6].

Although the laser performance is very similar comparing the two waveguides, it must be emphasized that the second one was fabricated just by one implant leading to less than half the total dose of the first sample. This is relevant because one cause of losses in waveguides is the tunneling through the barrier used to confine the light. To prevent these tunneling losses higher and broader barriers must be produced. The latter requires higher doses, increasing the lattice damage and potentially modifying the optical properties of the crystal. The alternative that has been considered and explored here with satisfactory results is to make broader barriers using smaller doses and increasing the angle of implantation.

4. Conclusions

Planar waveguide lasers in Nd:YAG crystals have been obtained by means of proton implantation. The waveguide lasers have been produced by two different processes: a proton multi-implant at an angle of 30° and a proton single-implant at an angle of 60°, resulting in broad optical barriers at depths of 9 and 4 μm, respectively. The waveguides present a very good laser performance, great stability at continuous wave operation and even with the marked differences in the optical barriers, they exhibit similar laser characteristics: a threshold of ~ 50 mW and slope efficiency up to 21.38 %. The results have shown that it is possible to achieve efficient laser oscillation in waveguides fabricated by proton implantation with a relatively small dose by increasing the angle of implantation.