The monomode enhanced-index LiNbO3 waveguide excited at 1540 nm is reported. X-cut LiNbO3 crystals were implanted at room temperature by 6.0 MeV C3+ ions with a dose of 2.0×1015 ions/cm2. Low loss planar optical waveguides were obtained and characterized by the prism coupling technique. Four dark modes were observed for extraordinary light at 633 nm, while only one enhanced-index mode was observed at 1540 nm. The propagation loss of the waveguide is 1.01 dB/cm measured with the moving fiber method. Reflectivity calculation method (RCM) was applied to simulate the refractive index profiles in waveguide. The width of waveguide structure induced by carbon ion implantation is ~3.6 µm.
© 2004 Optical Society of America
Broad band optical communications have gained great importance in the past few years, and this is expected to continue to grow. One important component is a wide bandwidth external optical modulator. In many high-speed and long-haul telecommunications systems, data encoding is accomplished using a LiNbO3 modulator [1, 2]. LiNbO3 is used because of its low optical loss, large electro-optic coefficient, and the existence of a mature technology for crystal growth, characterization, and fabrication. Moreover it can combine high optical nonlinearities with the capability of incorporating high concentrations of laser active ions. The basic building block of a LiNbO3 modulator is an optical waveguide that is formed in the LiNbO3 crystal.
Several methods are now widely used for fabricating waveguides in LiNbO3, such as diffusion [3, 4], ion exchange  and ion implantation . In the past, the first method was preferred due to the relative easy feasibility shown by the local doping techniques mentioned. But the optimization of diffused waveguides is difficult because one not only needs to understand the diffusion process, but also how the index changes as a function of the composition. However the ion implantation process becomes an extremely attractive technique for waveguide formation because of the possibility of introducing any given impurity at low temperature with accurate control of both the depth and lateral concentrations of the dopant.
In most cases, MeV light ions such as He+ have been implanted into LiNbO3 crystal with a dose of ~1016 ions/cm2 to produce the waveguide. At the end of the ion track, a low refractive index optical barrier is built up because of the lattice disorder produced by the nuclear damage process of the ion implantation. The light can be confined in the layer between the optical barrier and the surface, so the barrier-confined waveguide is formed . The barrier-confined waveguide was also produced by the implantation of heavy ions, such as Ni, and large refractive index decreases were found in the waveguide regions . Another way to fabricate an optical waveguide is to increase the refractive index of the crystal surface. In this paper, we report the monomode LiNbO3 waveguide fabrication by MeV C ion implantation with low dose. The waveguides were formed by ne increase in the waveguide layer. The ion implantation dose was around 1015 ions/cm2.
1.55 µm is preferred in optical fiber communication systems due to a minimum of fiber loss . Therefore, the property studies of the waveguides at telecommunication wavelength (~1.55 µm) are desired. Then we carried out the property studies of optical waveguides at 1540 nm.
2. Experimental details
The x-cut LiNbO3 samples, with the size of 21×5×2 mm3, were optically polished and cleaned. Before the ion implantation, the refractive indices of the sample were measured. They were implanted by 6.0 MeV carbon ions with a dose of 2×1015 ions/cm2 in vacuum at room temperature. The ion beam was electrically scanned to ensure a uniform implantation over the sample. To avoid channeling effect the samples were titled 7° off the beam direction. The implantation was performed at 1.7 MV tandem accelerator of Peking University.
The effective refractive indices of the dark modes were measured by using the prism-coupling method with a resolution better than 0.0002. Laser beam with 633 nm and 1540 nm were used in the measurement, respectively. The waveguide loss was measured by the fiber probe technique. When a guide mode was excited, a fiber probe scanned down the whole length of the propagating streak to detect the exponential decay of the light. It was assumed that the intensity of the light scattered out of the waveguide was proportional to the transmitted light power in the guide layer. The dark modes and the loss measurement were performed with a model 2010 prism coupler (Metricon, USA).
The reflectivity calculation method (RCM) , which was developed by Chandler and Lama, has been carried out to reconstruct the refractive index profiles in the observed waveguide structures.
3. Results and discussion
When the laser beam was employed with extraordinary polarization, dark modes were detected, see Figs. 1 (a) and (b). They show the measured relative intensity of the light from the prism versus the effective refractive index (ne) of the incident light. When the light was coupled into the waveguide, a lack of reflected light would result in a dip in intensity. Fig. 1 (a) shows that when 1540 nm laser beam is employed, only one mode is observed in the waveguide, and the effective refractive index of the mode (neff=2.1441) is higher than that of the substrate (nsubs=2.1383). Because there is only one mode, we cannot use the RCM to simulate the refractive index profile. When 633 nm laser beam is employed, there are four modes in the sample, as indicated in Fig. 1(b). The first two are very sharp, which may correspond to the waveguide modes, the following dips become broader, which may represent the substrate modes. This means that the light in the first two modes is well defined. The effective refractive index of TE0 mode (neff0=2.2075) is higher than the refractive index of the substrate (nsubs=2.2023), while the neff1 of TE1 mode (neff1=2.2016) is lower than the nsubs. Except for the first four dips, the other dips become much broader; this may result from the multiple optical reflections occurring at the interfaces between the waveguide and the substrate .
Since the information of the refractive index profile of the waveguides is especially important for application, many attempts have been done to reconstruct the index profile. Reflectivity calculation method (RCM) has been proved to be suitable for the simulation of the refractive index profile of ion-implanted waveguides. In this case a least-square fitting program based on RCM was available to calculate the refractive index profile by adjusting certain parameters until the theoretical modes indices match the experimental ones within a satisfactory error.
Figure 2 indicates the refractive index profiles based on RCM of the LiNbO3 waveguide measured with 633 nm laser beams. The refractive index of the substrate is marked in Fig. 2. The refractive index is increased by about 0.31% in the guiding region, and about 0.19% refractive index decrease occurs at the optical barrier. Therefore, the region between the surface and the barrier becomes a waveguide layer. Table 1 gives the comparison of the measured mode indices with the fitted values of the indices based on RCM for carbon ion-implanted waveguide. It is found that the measured effective refractive index is in agreement with the calculated values within 10-4.
The lattice damage produced by ion implantation is considered to be the main reason for refractive index change in the LiNbO3 waveguide. There are two kinds of damage produced by carbon implantation: near-surface damage correlated to electronic stopping, which causes an increase of the extraordinary refractive index, and end-of ion range damage generated by collision cascades, which decreases the extraordinary refractive index values . Then we have used TRIM, 98 (Transport of Ions in Matter) code to simulate the process of the implantation with 6.0 MeV C ions into LiNbO3. Figure 3 shows the electronic and nuclear energy loss as a function of penetration depth. Comparing Fig. 3 with Fig. 2, it seems reasonable that the near-surface (~3.2 µm) damage correlated to electronic energy deposition may cause a positive change of the extraordinary refractive index, while the nuclear collisions at the end of the ion track may decrease the extraordinary refractive index.
Figure 4 shows the light (633 nm) propagation in the waveguide fabricated by carbon ion implantation at a dose of 2×1015 ions/cm2 after annealing at 290 °C for 20 min in air. As we can see, there is a clear light propagation line in the waveguide.
The propagation loss of the as-implanted waveguide in LiNbO3 is 3.52 dB/cm. Then we gave heat treatments to the waveguide. After annealing at 260 °C for 20 min, the measured loss is 1.89 dB/cm. Figure 5 shows the loss measurement of the LiNbO3 waveguide after annealing at 260 °C for 20 min and then 290 °C for 20 min. And the loss is 1.01 dB/cm. Therefore, heat treatment under suitable conditions could be used as a practical method to decrease waveguide propagation loss.
The monomode LiNbO3 waveguide at 1540 nm was fabricated by low-dose MeV carbon ion implantation. The waveguide was formed with the extraordinary refractive index enhancement in the guiding region. The annealing at moderate temperatures was performed, which reduced the propagation loss down to 1.01 dB/cm. The refractive index profiles in the waveguide were simulated by RCM. The present result suggests a potential application of the carbon ion-implanted LiNbO3 waveguide.
This work was supported by the National Natural Science Foundations of China (Grant No. 10035010) and the MOE Key Laboratory of Heavy Ion Physics, Peking University.
References and links
1. Osamu Mitomi, Kazuto Noguchi, and Hiroshi Miyazawa, “Design of ultra-broad-band LiNbO3 optical modulators with ridge structure,” IEEE Transactions on Microwave Theory and Techniques 43, 2203–2207 (1995). [CrossRef]
2. A. L. Campillo, J. W. P. Hsu, C. A. White, and C. D. W. Jones, “Direct measurement of the guided modes in LiNbO3 waveguides,” Appl. Phys. Lett. 80, 2239–2241 (2002). [CrossRef]
3. Eli Arad, Shlomo Ruschin, and David Nir, “Buried modes in combined Ti diffused and Li outdiffused LiNbO3 slab waveguides,” Appt. Phys. Lett. 62, 2194–2916 (1993). [CrossRef]
4. B. Herreros and G. Lifante, “LiNbO3 optical waveguides by Zn diffusion from vapor phase,” Appl. Phys. Lett. 66, 1449–1451 (1995). [CrossRef]
5. Yu. N. Korkishko, V. A. Fedorov, S. M. Kostritskii, E. I. Maslennikov, M. V. Frolova, A. N. Alkaev, C. Sada, N. Argiolas, and M. Bazzan, “Proton-exchanged waveguides in MgO-doped LiNbO3 : Optical and structural properties,” J. Appl. Phys. 94, 1163–1170 (2003). [CrossRef]
6. S. S. Sarkisov, E. K. Williams, D. Ila, P. Venkateswarlu, and D. B. Poker, “Vanishing optical isolation barrier in double ion-implanted lithium niobate waveguide,” Appl. Phys. Lett. 68, 2329–2331 (1996). [CrossRef]
7. G. V. V ázquez and P. D. Townsend, “Improvements of ion implanted waveguides in Nd:YAG and LiNbO3 using pulsed laser anneals,” Nucl. Instr. Meth. B 191, 110–114 (2002). [CrossRef]
8. H. Hu, F. Chen, F. Lu, J. Zhang, J. Liu, K.-M. Wang, B.-R. Shi, D. Shen, and X. Wang, “Optical waveguide formation in LiNbO3 by 2.6 MeV Nickel Ions Implantation,” Chin. Phys. Lett. 18, 242–244 (2001). [CrossRef]
9. D.-L. Zhang and E. Y. B. Pun, “Accurate measurement of 1.5 µm of Er3+ in LiNbO3 crystals and waveguides,” J. Appl. Phys. 94, 1339–1345 (2003). [CrossRef]
10. P. J. Chandler and F.L. Lama, “A new approach to the determination of planar waveguide profiles by means of a non-stationary mode index calculation,” Opt. Acta 33, 127–142 (1986). [CrossRef]
11. P. D. Townsend, P. J. Chandler, and L. Zhang, Optical Effects of Ion Implantation (CUP, 1994).
12. G. G. Bentini, M. Bianconi, M. Chiarini, L. Correra, C. Sada, P. Mazzoldi, N. Argiolas, M. Bazzan, and R. Guzzi, “Effect of low dose high energy O3+implantation on refractive index and linear electro-optic properties in X-cut LiNbO3 : Planar optical waveguide formation and characterization,” J. Appl. Phys. 92, 6477–6483 (2002). [CrossRef]