1. Introduction

The integrated optical devices have become more and more important with the development of multimedia and internet, which in fact has stimulated a substantial interest towards optical waveguide research. Because of the excellent optical and ferroelectric properties, the niobate waveguides with tungsten bronze structure, such as Sr_xBa_1-xNb₂O₆ (SBN) and Ca_xBa_1-xNb₂O₆ (CBN), are the subject of intensive investigations [1–5]. Ca_xBa_1-xNb₂O₆ (x = 0.2–0.4) exhibits huge superiority due to the higher Curie temperature (greater than 280°C instead of 80°C for SBN), and this makes it more suitable for the high speed optical applications, in which the high density optical data packets generates plenty of heat [4]. Ndione et al. reported that the film waveguide of the CBN-28 had relative good waveguide properties at telecommunication wavelengths [5]. Gao et al. demonstrated an efficient diode-pumped Nd:CBN-28 ferroelectric crystal laser [1]. Tan et al. fabricated the waveguide structure on Nd:CBN-28 crystal by ion implantation method [6]. Up to now, all published papers mainly focus to the study of CBN with X_Ca = 0.28 [1, 5, 6], but the waveguide structure and optical properties of CBN crystal strongly depends on the calcium contents. Although CBN-20 single crystals have been grown, little work has been conducted on them.

Ion implantation is a powerful controllable approach to fabricate the optical waveguide [7–10]. It has been carried out on CBN-28 crystal successfully [6, 7]. In this work, we demonstrate the waveguide formation using 5 MeV carbon ions implantation with the fluence of 1 × 10¹⁵ ions/cm² on CBN-20 crystal. Micro-Raman imaging is used to provide the information about the crystal structure of CBN-20, which is a powerful nondestructive approach to examine lattice modification and local micro-structural change and lots of similar work in this area have been carried out [11–14]. In Ref [12], imaging micro-Raman spectroscopy was used to investigate the radiation damage and local strain in congruent LiNbO₃ as a result of high-energy (~MeV) He⁺ irradiation; point defects and compositional changes after irradiation and surface deformation are observed in the ion penetration region.

2. Experimental

Single crystal of a-cut Ca_0.20Ba_0.80Nb₂O₆ (CBN-20) was used in our work, which was provided by the Institute of Crystal Materials of Shandong University of China. The density of the sample is 5.351 g/cm³ and the Curie temperature (T_C) is about 305°C, which is higher than CBN-28 crystal. CBN-20 belongs to the point group P4bm and space group 4mm respectively. Prior to implantation, the (1 0 0) face of the sample was optically polished, and then implanted with 5 MeV C ions with the fluence of 1 × 10¹⁵ ions/cm² at a 7° tilt angle from normal to prevent ion channeling effect, which was performed on a 2 × 1.7 MV tandem accelerator at Peking University. After implantation, the end faces ((0 1 0) face) of the sample were optically polished and then prepared for the end-face coupling and Raman measurements.

The effective refractive index of the guided mode was measured by prism-coupling equipment. It has two laser sources corresponding to 633 nm and 1539 nm respectively. The polarization of the light can be modulated corresponding to excitation transverse magnetic (TM) or transverse electric (TE) guided mode. The end-face coupling equipment was used to get the near field light intensity profile and its optical arrangement has been introduced in Ref [15]. The confocal micro-Raman experiments were performed through a multichannel modular triple Raman system with confocal microscopy and an XY motorized stage at the School of Chemistry and Chemical Engineering, Shandong University. The laser source with the wavelength of 473 nm was used in our experiments and the diameter of the light spot could up to ≤ 1μm through a 100 microscope objective. The Raman probing was done by scanning along polished end face ((0 1 0) face) to ensure the Raman signal has been collected from the accurate waveguide and substrate region.

3. Results and discussion

When we performed the prism-coupling measurement, if the light coupled into the prism, there would be a guided mode correspondingly. Figures 1(a) and 1(b) show the effective refractive index (n_eff) of guided modes about C-implanted CBN-20 crystal with the wavelength of 633 nm and 1539 nm, respectively. The refractive indices (n_sub) of the sample substrate are 2.3175 for n_o and 2.2454 for n_e respectively as measured by the prism coupling method at a wavelength of 633 nm, and 2.2319 for n_o and 2.1755 for n_e at the wavelength of 1539 nm correspondingly. As seen from Fig. 1(a), the n_eff of TM₀ mode is 2.3056 and the n_eff of TE₀ mode is 2.2504 at the wavelength of 633 nm. Obviously, the n_eff of TM₀ mode is lower than the n_sub of n_o, whereas the n_eff of TE₀ mode is higher than the n_sub of n_e, which demonstrates the ion implantation process will reduce the ordinary refractive index and raise the extraordinary refractive index. This phenomenon is similar with the previous studies on ion-implanted LiNbO₃ crystal [16] and we think the similar physics mechanism occurs in the CBN-20 crystal. In other words, properly reduction of spontaneous polarization in CBN-20 crystal will reduce the n_o and raise the n_e. The n_eff of TM₀ mode is 2.2123 and the n_eff of TE₀ mode is 2.1760 at the wavelength of 1539 nm as shown in Fig. 1(b). Notice that there is only one mode in the n_e direction.

 

Fig. 1 The effective refractive index (n_eff) versus mode number of the CBN-20 planar waveguide for n_o and n_e: (a) at a wavelength of 633 nm, (b) at a wavelength of 1539 nm.

Download Full Size | PDF

Figure 2 is the microscope image of the polished end face of C-implanted CBN-20 crystal, which shows the obvious distinction among air, waveguide and substrate region is obviously. The location of waveguide region (implanted region) has been indicated by red arrow. The depth between the substrate edge and waveguide end is about 3.2 μm. In order to analyze the formation mechanism of the CBN-20 waveguide, we simulated the process of C-implanted CBN-20 crystal by SRIM 2010 [17] and the calculated electronic and nuclear energy loss profiles are shown in Fig. 3.The electronic energy loss (S_e) occupies a dominate position at the sample surface and simultaneously, the end of ion path is governed by Nuclear energy loss (S_n). The peak position of S_n is 3.17 μm and this is in accord with Fig. 2.

 

Fig. 2 Microscope picture of the end face of the CBN-20 after 5 MeV C ions implantation.

Download Full Size | PDF

 

Fig. 3 Electronic (divided by 10) and nuclear energy losses as a function of the penetration depth for 5 MeV C ions implanted into CBN-20 based on the SRIM 2010 simulation.

Download Full Size | PDF

As shown in Fig. 2, the color of the C-implanted region has changed and this phenomenon depends on the changes of refractive index. We reconstructed the refractive index profiles by reflectivity calculation method (RCM) [18] for multi-mode waveguide and pre-set parameter method [19] for single-mode waveguide to study the formation mechanism of the waveguide more clearly. The reconstructed refractive index profiles at the wavelength of 633 nm and 1539 nm are shown in Fig. 4(a) and 4(b) separately. The depth of optical barrier is 3.17 μm equal to the peak position of S_n. As shown in Fig. 4, the refractive index profile for n_e is “well + barrier” type, and the refractive index profile for n_o is pure “barrier” confined type. From Fig. 4(a), we can see that Δn_w = −0.01 in the waveguide region and an optical barrier (Δn_b = −0.08) for n_o when Δn_w = + 0.0076 and Δn_b = −0.005 for n_e at the wavelength of 633 nm, which have changed to Δn_w = −0.0074 and Δn_b = −0.042 for n_o, Δn_w = + 0.006 and Δn_b = −0.004 for n_e at the wavelength of 1539 nm.

 

Fig. 4 The n_o (dashed blue line) and n_e (solid red line) profile of the C ions implanted planar waveguide with energy of 5 MeV and fluence of 1 × 10¹⁵ ions/cm²: (a) at a wavelength of 633 nm; (b) at a wavelength of 1539 nm.

Download Full Size | PDF

The He-Ne laser was used to couple into a polished end face of CBN-20 waveguide and the output light was collected from another polished end face through the end-face coupling arrangement. Numerical simulations based on beam propagation method (BeamPROP software) [20] were used to study the modal properties of the waveguide using the refractive index profiles represented in Fig. 4. Experimental and theoretical fundamental mode profiles for CBN-20 planar waveguide at 633 nm are depicted in Fig. 5.The 3D near field light intensity profile for TM₀ mode (n_o direction) captured by a CCD camera is shown in Fig. 5(a), which shows that the light could be confined in the CBN-20 waveguide and the ability of optical confinement is quite well (there is almost no light in the air layer and substrate area). The profile could regard as two half Gaussian curve, and the full width at half maximum (FWHM) of substrate edge (right) is slightly wider than that of air edge (left). “The difference of the refractive index between waveguide and air is much higher than that between waveguide and optical barrier” is responsible for this distinction, and this convinces that the n_o direction of CBN-20 crystal after C ion implantation could support a real guided mode and it provides the direct evidence for the formation of the waveguide structure between the air and optical barrier. The waveguide could not propagate the TE polarized light through our experimental measure. In other words, the n_e direction of the waveguide could not carry a guided mode and this result is the same as C-ion implanted Na: CBN crystal introduced in Ref [21]. As reported in Ref [5, 6], the propagation loss of the fundamental TE mode is lower than that of TM mode in the c-axis orientation CBN-28 film waveguide fabricated by pulse laser deposition; Nd: CBN-28 (c-cut) waveguide formed by He-ion implantation could only carry TE polarized mode. In a word, the waveguide structure of n_o direction could be more easily obtained in CBN series crystal. Notice that in the inset of Fig. 5(a), the 2D near field light intensity profile is shown for clearly. Theoretical fundamental mode profiles for n_o and n_e direction of the waveguide are depicted in Figs. 5(b) and 5(c). The Fig. 5(b) could match the experimental counterpart. As depicted in Fig. 5(c), the refractive index profile of n_e could carry a real guided mode and this could not compare with our experimental result, which demonstrates that the simulation process exists some limitation. We also simulated the propagation process of light with the wavelength of 1539 nm using the refractive index profiles depicted in Fig. 4 (b), and the simulation results show that the CBN-20 crystal could not propagate the light with the wavelength of 1539 nm, which is consistent with our experiment measurement.

 

Fig. 5 The near field light intensity profile of the planar waveguide: (a) TM₀ mode captured by a CCD camera correspond to n_o direction; (b) simulated result of TM₀ mode based on n_o profile; (c) simulated result of TE₀ mode based on n_e profile.

Download Full Size | PDF

As reported in Ref [12], if the point and extended defects generate in a crystal, there would be some changes in the Raman spectra. We studied the micro-Raman properties of the planar waveguide using a confocal microscope to explore the influence of the lattice structure by ion implantation process. We measured the Raman spectra at different position of the polished end face to obtain the signal of the waveguide and substrate area along the depth direction, and the back scattered signal were collected by a microscope object. Figure 6(a) is the micro-Raman spectra obtained from the C-implanted CBN-20 waveguide (at the depth of 1.5 μm) and substrate area (several microns away from the waveguide). The research on Nd: CBN-28 crystal introduced in Ref [6]. reveals that the intensity and location of the Raman peak almost wasn’t affected by the ion-implantation process. Compared with this, our experimental data on CBN-20 crystal are very different. The peaks shown in Fig. 6(a) correspond to characteristic peaks of Nb-O octahedral. As a general feature, it can be observed that the signal intensity decreases after the implantation process and the position of the peaks also shifts. The shift of the peak (648 cm⁻¹) is indicated in the inset of Fig. 6(a) by the dashed lines. In addition, the linewidth of the peak also broadens. The normalized intensity profiles are shown in Fig. 6(b) more clearly. The simultaneous Raman shows that the broadening and intensity decrease at the surface of CBN-20 crystal, which unequivocally indicates that the crystal lattice has been partially damaged and disordered by C-ion implantation. It is clear that the disorder results in a peak redshift. The presence of the shifted spectrum indicates that the change of the local lattice constants, and this phenomenon can be attributed to a slight local dilatation of the CBN-20 network. Since the broadening can also result from the local strain, which was reported in Ref [22], a more detailed set of experiments must be undertaken before more definitive discussions. With these variations of Raman spectra, the implantation-induced modifications indicate that C-ion implantation causes the distortion of Nb-O octahedral and lattice disorder.

 

Fig. 6 Raman spectra obtained at end face of the sample with different probing locations: waveguide region labeled by solid red line, substrate region labeled by dashed blue line. (b) is normalized Raman spectra by use of (a).

Download Full Size | PDF

4. Conclusions

The barrier waveguide on CBN-20 crystal has been formed after our implantation condition. We simulated the light propagation process at the wavelength of 633 nm by BeamPROP software using the reconstructed refractive index profile calculated by RCM method. The simulation results could compare with our experimental measurement on the light propagation in the n_o direction, but there are some limitations in the n_e direction. SRIM 2010 and Raman spectra were used to analyze the lattice changes induced by ion implantation process, and finally the lattice disorder has been proved clearly through the obvious changes in the intensity, peak shift, and linewidth of Raman peaks.