Twin-hole fibers were thermally poled with different internal electrode configurations, including having only one anode wire in the hole, two anode wires in the two holes, one cathode wire, and two cathode wires in the holes, in comparison to the conventional one anode wire and one cathode wire combination. Second harmonic microscopy was utilized to visually reveal the spatial distribution and to measure the magnitude of the induced second-order optical nonlinearity within the poled fibers. It was found that both one- and two-anode configurations resulted in strong nonlinearity comparable with the conventional case but the two-anode configuration was more reproducible than the one-anode case; for the one-cathode-wire and two-cathode-wire configuration, strong nonlinearity in a ring shape concentric with the fiber outer surface was induced as if the cathode metal wire were in the center of the twin-hole fiber rather than substantially offset. These new results provide strong support for the proposed model of a “self-adjustment” mechanism and point the way to simplified and more repeatable experimental techniques.
©2012 Optical Society of America
Thermal poling, a technique first reported in 1991 , has been widely used to induce a large and stable second-order optical nonlinearity, χ(2), in otherwise amorphous silicate glasses and optical waveguides [2–8]. Such nonlinear optical glasses and waveguides can be potentially used in a wide range of applications such as second-harmonic generation [9–11], electro-optic modulation and switching [12, 13], and telecom-band polarization-entangled photon pair generation . The underlying physical mechanism of the thermal poling process has been generally accepted to be the creation of an internal electric field under the anode surface caused by mobile ions (mainly Na+) migration towards the cathode, and the subsequent action of the resultant frozen-in electric field on the intrinsic third-order nonlinearity of the poled glass through [15, 16]. The technique involves applying a high electric field across a glass sample at an elevated temperature. To thermally pole an optical fiber, the normal practice is to use a twin-hole fiber in which two holes are present at either side of the core for holding metal wires as internal anode and cathode respectively [6, 17]. Recently, W. Margulis et al. reported that a strong and stable χ(2) can still be induced in twin-hole fibers with two internal electrodes both as anodes . Such an apparently counter-intuitive poling configuration provides some advantages over the conventional anode-cathode poling configuration in terms of simplicity and less risk of electric breakdown in the poled glass region, and thus deserves further investigation.
In this paper, we report our comprehensive experimental results of thermally poling optical fibers with different electrode arrangements. Besides the anode-anode and single-anode arrangements investigated by W. Margulis et al. , additional electrode arrangements were also investigated in our experiment, i.e. single-cathode and dual-cathode combinations. The magnitude and distribution of the induced χ(2) in each electrode configuration was characterized by second-harmonic (SH) microscopy [19, 20], and compared with that in fibers poled with the traditional anode-cathode configuration. Several interesting features, including the concentricity of the χ(2) profile with fiber outer surface or electrode holes, were observed and these are discussed in comparison with a numerical simulation of the electric field distribution at the onset of thermal poling.
2. Twin-hole optical fiber, thermal poling arrangement, and characterization
The twin-hole fibers used in our experiment were fabricated by the modified chemical vapor deposition method. The diameters of the overall fiber and the two holes were ~188 and 67 μm, respectively. The hole-to-hole spacing was ~23 μm. The fiber core, ~4.5 μm × 6.0 μm and elongated in the hole-to-hole direction, was ~7.8 μm edge-to-edge to one hole and ~9.2 μm edge-to-edge to the other. The highly Ge-doped fiber core had a numerical aperture (NA) of 0.34. As the main purpose of this investigation was to determine the effects of different electrode arrangements, the exact waveguiding property of the fiber was unimportant to this study. Aluminum wires of 50 μm diameter were inserted in the holes as electrodes. In the experiment, fibers were thermally poled by being placed on top of a hot plate at 320°C and applying 3.5 kV dc voltage to the electrodes. The higher temperature of 320°C, rather than the usual 280°C, was adopted in this experiment in order to speed up the poling process and overcome the impeding effect to migrating charges from the core-cladding interface as well [21, 22]. The poling time was varied from 20 to 40 min.
Different electrode-wire combinations were investigated, namely, conventional anode-cathode (a positive voltage applied to one wire, with the other connected to the ground and the hot plate at a floating voltage), anode-anode (a positive voltage applied to both wires and the hot plate grounded), single-anode (single wire in only one hole with a positive voltage applied and the hot-plate grounded), single-cathode (single wire in one hole with a negative voltage applied and the hot plate grounded), and cathode-cathode (a negative voltage applied to both wires and the hot plate grounded).
The SH microscopy set-up was similar to the one we reported earlier . An inverted Leica DMI6000B microscope equipped with a Leica SP5II confocal and multiphoton system was used. The laser source was a Mai Tai Deep See tunable pulsed titanium sapphire laser operating at 810 nm with pulses in the 100-200 fs range. The microscope was also equipped with dual photomultiplier transmitted light detectors, with a DCLP dichroic mirror dividing the detectable spectrum (380-680 nm) at 495 nm between the two channels, with the longer wavelength channel (channel 1) detecting either two-photon 500-550 nm fluorescence or a 543 nm transmission non-confocal image of the sample and the shorter wavelength channel (channel 2, fitted with a 405/10 nm narrow-band pass filter) receiving only the second harmonic signal. The spatial resolution of the SH microscopy was estimated to be ~0.6 μm. The excitation laser beam was mainly linearly polarized along a direction ~15° counter-clockwise relative to the horizontal direction of the specimen platform. Measurement of the nonlinear coefficient d33 was calibrated with a y-cut quartz plate with d11 = 0.335 pm/V.
3. Experimental results
3.1 Conventional anode-cathode configuration
Typical SH microscopy micrographs from the cross-sections of the optical fibers poled with the conventional anode-cathode configuration are shown in Fig. 1 . The fiber core with a higher refractive index can be clearly identified in Fig. 1(a). The SH signal distribution profile is shown in Fig. 1(b). It can be seen that SH signals are only generated from a narrow ring-shaped layer that surrounds the anode hole. This can be seen more clearly in the overlay micrograph of Fig. 1(c). The circular χ(2) layer is almost concentric with the fiber hole. The impeding effect of the core-cladding interface on the progression of the χ(2) layer is observed again in this case [21, 22], but the fiber core has become partially SH active, indicating a near optimum choice of the poling condition. The measured nonlinear coefficient d33 in the core and in the ring-shaped layer outside the core was in the range of 0.19-0.22 and 0.11-0.15 pm/V, respectively. These values agree well with our previous experimental results .
3.2 Single-anode configuration
Typical SH micrographs are shown in Fig. 2 . The χ(2) magnitude and distribution profile resemble that of the conventional anode-cathode wire configuration. It is also found that the uniformity of the χ(2) layer and the repeatability of this poling configuration are not as good as the conventional configuration. In many cases, the narrow χ(2) layer did not form a regular continuous ring but had a wide gap, which, if its location coincided with the core, resulted in the core being SH inactive, as shown in Fig. 3 . This ineffectiveness in core poling might be due to non-optimal rotation of the fiber. Note also that surface charges at the inner wall of the anode hole also generated weak SH signals.
3.3 Anode-anode configuration
The anode-anode configuration is the same as that adopted by Margulis et al. . Compared with the single-anode configuration, the repeatability of this poling arrangement is greatly improved. In all such poled fibers, the χ(2) layer is almost the same in shapes and sizes, with typical SH micrographs shown in Fig. 4 . Outside the fiber core, the χ(2) layer is in a figure-of-eight shape, with the circular sections concentric with the corresponding electrode holes. It has the appearance that the two rings around the anode holes have crossed the fiber core and, after coming into contact with each other, have merged into the closed figure-of-eight shape. Furthermore, the fiber core remains SH active even if the figure-of-eight χ(2) layer is outside it. The nonlinearity coefficient d33 was also measured, with 0.2-0.22 pm/V for the figure-of-eight χ(2) layer outisde the core and ~0.21 pm/V in the core.
To investigate the χ(2) behavior in an earlier stage of thermal poling in this configuration, some fibers were either poled for a time duration shorter than 40 min or at a lower temperature than 320°C. Typical SH micrographs are shown in Fig. 5 . Now the two χ(2) rings have not joined together but are still separated, each resembling the χ(2) distribution in the anode-cathode case, as shown in Fig. 1(b). Similar to the case of Fig. 1, the χ(2) ring has reached the core-cladding interface and the core has become SH active.
3.4 Single-cathode configuration
In this case only one hole had a cathode Al wire inside with the hot plate grounded. Typical SH micrographs are shown in Fig. 6 . It is interesting to note that now the induced χ(2) is mainly distributed in a ring concentric with the fiber outer surface. The χ(2) magnitude, d33, was measured to be ~0.12-0.14 pm/V, very similar to that achieved in the conventional anode-cathode case. The distance of the χ(2) ring away from the fiber outer surface is ~11 μm.
This result shows that, although there was only one cathode wire in the twin-hole fiber, the χ(2) layer progressed into the fiber cladding as if the fiber were poled with the cathode wire located at the geometrical center of the fiber.
3.5 Cathode-cathode configuration
Finally, for the dual-cathode configuration, where both holes had an Al wire at −3.5 kV with the hot plate electrically grounded, typical SH microscopy results are shown in Fig. 7 . There is no significant difference from the single-cathode configuration. The ring-shaped χ(2) profile, almost concentric with the fiber outer surface, was again observed here. The position of the χ(2) layer and the χ(2) magnitude are almost identical to those achieved in the single-cathode configuration.
4. Discussion and numerical simulation
The experimental results show that, when the hot plate is at ground potential, the single-anode wire configuration could be, although less repeatable, almost as effective as the conventional anode-cathode configuration in χ(2) magnitude and distribution. The dual-anode internal wire configuration is even better than the traditional anode-cathode configuration with slightly larger χ(2) achieved. According to the space-charge field model [15, 16], the creation of a strong internal electric field relies on the depletion of mobile charges (Na+), activated by thermal stimulus and driven by the applied poling electric field, from the region under the anode surface. It has been observed that the poling voltage seems to have a threshold below which no detectable χ(2) could be induced [24, 25]. With this in mind, one might expect ineffective poling from such an anode arrangement, because, theoretically, in the dual-anode configuration, the electric field at the central point of the two holes would be very weak. Assuming a fully filled electrode hole, we numerically calculated the distribution profile of both the electric field E and voltage V within the fiber cross section at the outset of thermal poling (before any charge migration occurs), with the field distribution profile along the shortest line connecting the two electrode holes (and across the fiber core) shown in Fig. 8 . For a fiber of the same geometry and poled under the same conditions as this experiment, the anode-core voltage difference is <55 volts and the initial electric field experienced by the fiber core is virtually zero. Yet strong χ(2) has been achieved in the fiber core with the dual-anode configuration.
It is also interesting that in the experiment the χ(2) layer around the anode hole seems to progress at a relatively uniform pace in all directions around the hole, even though the simulation shows that the initial applied poling field varies vastly in the fiber cross section. We calculated the initial E magnitude along two circles, one concentric to the fiber cross-section but 5 μm below the fiber outer surface and the other concentric to the left hole but 0.1 μm bigger in diameter. The result is shown in Fig. 9 .
This poling behavior could be explained by a “self-adjustment” mechanism, similar to the “avalanche-like positive feedback” mechanism proposed earlier by W. Margulis et al. . As soon as the poling voltage is applied to the electrodes, charge migration begins. At regions where the initial electric field is stronger, charges move faster and the depletion region there is developed faster. This can lead to an initially non-circular χ(2) shape around the anode hole. But the creation of the depletion layer immediately increases the electric resistivity in the depletion glass layer. A major part of the applied voltage is dropped across this depletion layer and the charge migration slows down, allowing charge migration at other areas where the initial poling field is weaker to catch up. As thermal poling goes on, injection of hydrogen-related charges (e.g. H3O+) into the depletion layer and subsequent charge neutralization will follow. This will result in the expansion of the depletion layer and push the front of the depletion layer farther towards the cathode. Through the self-regulation to the charge migration process, the final front of the depletion layer around the anode hole seems to progress at the same pace during fiber poling; a phenomenon experimentally observed in all the electrode configurations. The striking similarity between the simulated distribution profile of the mobile Na+ ions in Fig. 5 of W. Margulis’s work  and the χ(2) profile in Fig. 4 of this paper provides strong experimental support for the proposed “self-adjustment” mechanism. The observed “figure-of-8” narrow-layer χ(2) profile in this experiment provides further support to the charge-neutralization mechanism and once again demonstrates that the induced χ(2) is mainly formed at the very front edge of the depletion layer .
In the anode-anode wire configuration, the fact that the fiber core retains a large χ(2) even after the poling fronts have moved through it, possibly due to the existence of low-mobility H3O+ ions at the core-cladding interface that cannot be driven away by the relatively weak poling field in this case, is significant. This implies that the fiber cannot be over poled in this configuration, or at least that it is more resistant to over poling, thus greatly easing the control of poling time.
Both single- and dual-cathode configurations are effective in inducing a strong χ(2) in the twin-hole fiber. It is interesting to see that in both cases the induced χ(2) is distributed in a narrow circular layer concentric with the fiber outer surface, as if there were only one wire located at the very geometrical centre of the fiber. The phenomenon can also be explained by the “self-adjustment” mechanism discussed above. Knowledge of this poling behavior can be used to greatly simplify the negative-voltage poling procedure, as only one electrode (wire) is required and the wire thickness relative to the hole size is not an issue. Such electrode arrangement is particularly suitable to poling optical fibers in which the core is closer to the outer surface.
This cathode-wire poling feature is not confined to a cylindrical fiber; it also applies to fibers with less geometrical symmetry. In our earlier experiments, D fibers with only one hole in the cladding were poled with a cathode wire in the hole. Typical SH results are shown in Fig. 10 . Once again we can see that the χ(2) layer is almost at the same distance everywhere beneath the fiber outer surface. It is expected that the same behavior applies to other fibers of more irregular shapes.
According to the space-charge field model, the χ(2) region corresponds to an internal electric field that is frozen-in after thermal poling is finished [15, 16]. The etching rate of the poled silica fiber in HF is affected by the presence of this frozen-in electric field and the resultant observable features under a scanning electron microscope (SEM) reflect the distribution of the frozen-in electric field. To test this, another polished fiber section from the same poled fiber as in Fig. 10 was etched in diluted HF and then observed with SEM, with a typical micrograph shown in Fig. 11 . An excellent agreement in location between the etched trench and the χ(2) layer can be clearly seen.
Twin-hole fibers have been thermally poled with different internal electrode wire configurations in addition to the conventional anode-cathode configuration. Experimental results show that both the single-anode and dual-anode configurations can achieve a large χ(2) comparable to that in the conventional case, but the dual-anode configuration possesses far better repeatability. The induced χ(2) is in a narrow circular layer almost concentric with the anode hole. For the dual-anode case, large χ(2) can be attained in the fiber core even after the χ(2) layers have passed the core and formed a “figure-of-eight” shape. Single-cathode and dual-cathode configurations achieve almost identical results, both with the induced χ(2) distributed in a thin circular layer concentric to the fiber outer surface as though the cathode wire were put in the geometrical center of the fiber cross-section.
These findings have direct impact on the design and subsequent thermal poling of specialty multi-core and/or multi-hole optical fibers. For multi-core fibers, to ensure efficient χ(2) creation through thermal poling, the cores can be positioned in a circle concentric with the hole when poled with positive voltage, or in a circle concentric with the fiber outer surface when poled with negative voltage. For multi-hole fibers, when a negative voltage is used, all that is needed is only one cathode wire. What is more, it seems that the relative position of the cathode hole in the fiber cross-section is not of any real significance, thus greatly easing the constraints on fiber design and facilitating fiber fabrication.
This research was supported under the Australian Research Council’s Discovery funding scheme (project number DP0986237). The authors acknowledge the facilities as well as scientific and technical assistance from staff in the Australian Centre for Microscopy & Microanalysis (ACMM), the University of Sydney.
References and links
6. D. Wong, W. Xu, S. Fleming, M. Janos, and K. M. Lo, “Frozen-in electrical field in thermally poled fibers,” Opt. Fiber Technol. 5(2), 235–241 (1999). [CrossRef]
7. J. Arentoft, M. Kristensen, K. Pedersen, S. I. Bozhevolnyi, and P. Shi, “Poling of silica with silver-containing electrodes,” Electron. Lett. 36(19), 1635–1636 (2000). [CrossRef]
8. W. T. Li, H. An, and S. Fleming, “Second-order optical nonlinearity in thermally poled multilayer germanosilicate thin films,” Electron. Lett. 44(10), 639–641 (2008). [CrossRef]
9. R. Kashyap, G. J. Veldhuis, D. C. Rogers, and P. F. McKee, “Phase-matched second-harmonic generation by periodic poling of fused silica,” Appl. Phys. Lett. 64(11), 1332–1334 (1994). [CrossRef]
10. S. Chao, H.-Y. Chen, Y.-H. Yang, Z.-W. Wang, C. T. Shih, and H. Niu, “Quasi-phase-matched second-harmonic generation in Ge-ion implanted fused silica channel waveguide,” Opt. Express 13(18), 7091–7096 (2005), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-13-18-7091. [CrossRef] [PubMed]
11. A. Canagasabey, C. Corbari, A. V. Gladyshev, F. Liegeois, S. Guillemet, Y. Hernandez, M. V. Yashkov, A. Kosolapov, E. M. Dianov, M. Ibsen, and P. G. Kazansky, “High-average-power second-harmonic generation from periodically poled silica fibers,” Opt. Lett. 34(16), 2483–2485 (2009). [CrossRef] [PubMed]
12. M. Abe, T. Kitagawa, K. Hattori, A. Himeno, and Y. Ohmori, “Electro-optic switch constructed with a poled silica-based waveguide on a Si substrate,” Electron. Lett. 32(10), 893–894 (1996). [CrossRef]
14. L. G. Helt, E. Y. Zhu, M. Liscidini, L. Qian, and J. E. Sipe, “Proposal for in-fiber generation of telecom-band polarization-entangled photon pairs using a periodically poled fiber,” Opt. Lett. 34(14), 2138–2140 (2009). [CrossRef] [PubMed]
15. P. G. Kazansky and P. St. J. Russell, “Thermally poled glass: frozen-in electric field or oriented dipoles?” Opt. Commun. 110(5-6), 611–614 (1994). [CrossRef]
16. T. G. Alley, S. R. J. Brueck, and R. A. Myers, “Space charge dynamics in thermally poled fused silica,” J. Non-Cryst. Solids 242(2-3), 165–176 (1998). [CrossRef]
17. N. Myrén, H. Olsson, L. Norin, N. Sjödin, P. Helander, J. Svennebrink, and W. Margulis, “Wide wedge-shaped depletion region in thermally poled fiber with alloy electrodes,” Opt. Express 12(25), 6093–6099 (2004), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-12-25-6093. [CrossRef] [PubMed]
18. W. Margulis, O. Tarasenko, and N. Myrén, “Who needs a cathode? Creating a second-order nonlinearity by charging glass fiber with two anodes,” Opt. Express 17(18), 15534–15540 (2009), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-17-18-15534. [CrossRef] [PubMed]
19. W. Margulis, F. Laurell, and B. Leschel, “Imaging the nonlinear grating in frequency-doubling fibres,” Nature 378(6558), 699–701 (1995). [CrossRef]
20. H. An, S. Fleming, and G. Cox, “Visualization of second-order nonlinear layer in thermally poled fused silica glass,” Appl. Phys. Lett. 85(24), 5819–5821 (2004). [CrossRef]
21. D. Faccio, A. Busacca, D. W. J. Harwood, G. Bonfrate, V. Pruneri, and P. G. Kazansky, “Effect of core-cladding interface on thermal poling of germano-silicate optical waveguides,” Opt. Commun. 196(1-6), 187–190 (2001). [CrossRef]
22. H. An and S. Fleming, “Hindering effect of the core-cladding interface on the progression of the second-order nonlinearity layer in thermally poled optical fibers,” Appl. Phys. Lett. 87(10), 101108 (2005). [CrossRef]
23. H. An and S. Fleming, “Creating large second-order nonlinearity in twin-hole optical fibre with core at the centre of the two holes,” Electron. Lett. 43(4), 206–207 (2007). [CrossRef]
24. Y. Quiquempois, N. Godbout, and S. Lacroix, “Model of charge migration during thermal poling in silica glasses: Evidence of a voltage threshold for the onset of a second-order nonlinearity,” Phys. Rev. A 65(4), 043816 (2002). [CrossRef]
25. H. An and S. Fleming, “Second-order optical nonlinearity in thermally poled borosilicate glass,” Appl. Phys. Lett. 89(18), 181111 (2006). [CrossRef]