Open Access
Add to BibSonomy
Abstract
Ion-sliced lithium niobate thin films offer high optical confinement over their bulk counter parts, enabling compact and efficient nonlinear optics. Quasi-phase matching by periodic poling of thin film lithium niobate remains challenging, however, due to large leakage currents, especially in magnesium oxide doped lithium niobate, which is used in bulk to reduce photorefractive damage. Here we present fabrication and poling details of 700 nm thick x-cut magnesium oxide doped lithium niobate thin films using electric field poling. We introduce a silicon dioxide insulation layer under co-planar electrodes to reduce leakage current. Rounded tips are utilized to encourage nucleation at the center of the electrodes. Uniform domains with 7.5 µm period and 50% duty cycle are achieved. The poling characteristics are compared to bulk lithium niobate, with and without the silicon dioxide insulation layer. The domains are characterized by piezoresponse force microscopy, hydrofluoric etching, and waveguide second harmonic generation at 1550 nm wavelength in a thin film lithium niobate waveguide loaded with a silicon nitride strip.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Lithium niobate (LN, LiNbO3) thin films enable an integrated optical platform for compact and high-efficiency nonlinear optics [1,2], with applications ranging from quantum optics [3], spectroscopy [4], and signal processing [5]. Bulk LN is a well-established second-order nonlinear optical material with attractive properties, including a wide transparency range spanning from 450 nm to 4.5 µm (α < 1 cm−1) [6], large electro-optic coefficients (r33 = 30.8 pm/V at λ = 0.63 µm) [7], and large nonlinear optical coefficients (d33 = −33.4 pm/V at λ = 1.15 µm) [8]. Also, LN’s ferroelectric nature enables quasi-phase matching (QPM) by periodically inverting the spontaneous polarization with the application of an electric field. Magnesium oxide (MgO) doping by 5 mol% further improves the material properties by reducing the coercive field (Ec) required for poling from 21 to 4 V/µm [9], as well as the sensitivity to green-induced infrared absorption (GRIIRA) [10] and photorefractive damage [11]. Films with sub-micrometer thicknesses are ion-sliced from a bulk wafer and are often bonded to a LN substrate with an intermediate SiO2 layer forming lithium niobate-on-insulator (LNOI) [12]. LNOI waveguides have a high refractive index contrast that supports optical mode field diameters one order of magnitude smaller than conventional titanium in-diffused or proton exchanged waveguides in bulk [2]. These highly confined modes lead to high optical intensities which motivate the use of MgO doped LN, as relatively low optical powers may exceed the photorefractive damage threshold. Furthermore, the high confinement reduces bend loss and enhances modal overlap, important for efficient nonlinear optical interactions.
Second harmonic generation (SHG) [13–20], and sum/difference frequency generation (SFG/DFG) [21,22] have been demonstrated in LNOI using a variety of phase matching techniques. Modal phase matching uses waveguide dispersion engineering between a specific set of modes with either different polarizations and/or higher-order modes [13,15,18,19]. Mode shape modulation achieves phase matching by periodically changing the width of the waveguide to modulate the nonlinear modal overlap [16,20]. These methods, however, are limited in conversion efficiency and design flexibility compared to QPM by periodic poling due to either reduced overlap between the different mode profiles, inability to access the largest nonlinear coefficient, or weak modulation and scattering losses in structured waveguides.
Poling has been successfully demonstrated in x-, y- and z-cut LN thin films with different electrode configurations. For z-cut films, the poling occurs in the vertical direction and the transverse-magnetic (TM) mode is used to access the largest nonlinear coefficient d33. The poling field can be applied through patterned electrodes [23–25], or with a conductive atomic force microscope (AFM) tip [25–27] on the +z surface. A drawback to this geometry is that a buried electrode is required under the film. While the AFM technique is capable of patterning nanometer scale domains, it is a serial process with limited write fields. Another option is to first perform the poling on a bulk wafer, and then proceed with ion-slicing a thin film, albeit less flexible [21,28,29]. For many devices, x-cut is the preferred orientation because the transverse-electric (TE) mode accesses d33 and co-planar electrodes can be deposited on the top surface. Poling of x- and y-cut thin films has been demonstrated by applying multiple high voltage pulses to co-planar electrodes [14,17,30,31], with periods as small as 2.4 µm [14].
One of the challenges in poling MgO:LN is the leakage current that flows through the charged domain walls (CDW). These domain walls separate regions with different directions of spontaneous polarization and exhibit bound charge discontinuities. CDWs in bulk MgO:LN has been modeled numerically [32] and measured experimentally [33], with both reporting orders of magnitude increase in conductivity. The leakage current increases with the amount of MgO doping and no leakage current is observed in undoped bulk LN [9,34]. In effect, the domains create a conductive path that reduces the applied electric field, thereby stunting domain growth, and resulting in non-uniform domains [9]. Off-axial cut substrates [9] and SiO2 insulation layers [35,36] have been shown to reduce the leakage current in bulk samples by blocking the connection between electrodes. While leakage currents were not observed in undoped bulk LN, they have been reported during poling of undoped LN thin films, including both x- [14] and z-cut [37] LNOI. More recently, the conductivity at domain walls in undoped z-cut LNOI has been studied by AFM techniques and found to be 5 orders of magnitude larger than surrounding areas [38]. Techniques to minimize the leakage current are therefore pertinent to both MgO- and undoped LNOI. In this work, we report the fabrication and poling of ion-sliced x-cut MgO:LN thin films with an SiO2 insulation layer under the electrodes to reduce the leakage current. The poling characteristics are compared to bulk samples with and without the SiO2 insulation layer by means of poling voltage and current measurements. The domains are characterized by piezoresponse force microscopy (PFM), hydrofluoric (HF) etching, and second-harmonic generation.
2. Fabrication
MgO doped LNOI wafers are fabricated in-house via ion-slicing. We start with two congruent x-cut LN wafers. One wafer is doped with 5 mol% MgO and serves as the thin film donor while the other undoped wafer is used as a supporting handle. A SiO2 layer is deposited by plasma enhanced chemical vapor deposition (PECVD) on the LN handle and chemically-mechanically polished (CMP) to enable direct bonding. The donor wafer is implanted with He+ ions at 230 kV and a dose of 3.5 × 1016 He+/cm2 to form a damage layer below the +x surface. The depth of the damage layer depends on the implantation energy and is calculated using Stopping and Range of Ions in Matter (SRIM) simulations [39]. After implantation, both wafers are bonded at room temperature via wet chemical surface activation. The bonded wafers are then annealed to exfoliate a film along the ion implant damage layer and to repair ion implant induced damage [40]. Lastly, the thin films are polished to a final thickness of 700 nm with less than 0.5 nm RMS surface roughness and diced into ∼2 cm2 pieces.
Poling electrodes consisting of 160 nm evaporated chromium (Cr) on top of 100 nm PECVD SiO2 are defined by electron beam lithography (EBL) and plasma etching. Hydrogen silsesquioxane (HSQ) resist serves as the Cr etch mask and the Cr is subsequently used to mask the underlying SiO2. The HSQ is fully removed during the SiO2 etching. Figure 1(a) shows a schematic of the electrodes and Fig. 1(b) provides a zoomed in view of an electrode with a rounded tip. We include a number of electrode designs on bulk and LNOI samples, with and without the SiO2 insulation. The poling periods Λ range from 6 to 10 µm and the duty cycle spans 30−65%, while the gap size is fixed at 15 µm. Designs also include round and square electrode tips. The electrode length L is 80 µm for poling optimization tests, but is extended to 2 mm for SHG devices. Additionally, the LNOI samples used for poling optimization have a 4 µm thick buried oxide (BOX), while SHG devices are fabricated on samples with a 1 µm thick BOX. Lastly, the SHG devices are based on LN thin films loaded with a silicon nitride (SiN) strip that is patterned between the electrodes. The SiN is deposited by reactive sputtering and is patterned with EBL and plasma etching. The total waveguide length is 5.8 mm after dicing and edge polishing.
Fig. 1. (a) Schematic of poling electrodes and silicon nitride (SiN) strip on 700 nm thick x-cut MgO:LN thin film. (b) Zoomed in view of poling electrode tip showing 160 nm chromium (Cr) on top of 100 nm SiO2. (c) Poling circuit; pulses generated from an arbitrary waveform generator (AWG) are amplified by a high voltage amplifier (HVA) and applied to a sample covered with silicone oil. An oscilloscope monitors the poling voltage (VLN) and current (Ipol) through a voltage divider (R1 and R2) and series resister R3, respectively. Resistor RS provides over-voltage protection to the oscilloscope in the case of a short circuit. R1 = 20 MΩ, R2 = 2.26 MΩ, R3 = 330 kΩ, RS = 2 MΩ, and the oscilloscope input resistance Rin = 10 MΩ.
Poling is performed at room temperature using the circuit in Fig. 1(c). The sample is coated with silicone oil to prevent air breakdown / electrode burning and is connected using contact probes. Pulses are created by an arbitrary waveform generator and amplified by a high-voltage amplifier. The voltage and current are monitored with an oscilloscope using a voltage divider and a series resistor. After poling, the domains are imaged by PFM with platinum coated silicon probes. The conductive probe is scanned on the +x surface in contact mode while applying a ± 10 V AC signal. The applied electric field produces local oscillations through the converse piezoelectric effect which are measured by the probe deflection and read out by a lock-in amplifier. The electric field from the probe tip is strongly inhomogeneous and produces in-plane forces arising from the d22 and d33 piezoelectric tensor elements. We choose to align the probe’s cantilever with LN’s z-axis such that the d33 element causes the probe to buckle and ultimately be recorded as vertical (out-of-plane) movements [41]. Negative deflections correspond to -z crystal axis and positive deflections correspond to +z crystal axis. While PFM offers non-destructive, high-resolution imaging, it does not provide information on the depth the domains. Therefore, we also relied on focused ion beam (FIB) milling and HF etching.
3. Results and discussion
Many parameters were varied in the process of optimizing the electrode and waveform design, including voltage, duration, ramp rates, number of pulses, duty cycle and electrode tip style, for both bulk and LNOI. For the electrode designs, we have observed that with rounded tips, the domains tend to nucleate at the apex and spread out laterally from the center, as opposed to square tips which nucleated at the corners. The end result is that rounded tips produce narrower domains. Also, the width of the domains was seen to be directly proportional to the width of the electrode fingers, and therefore the duty cycle of the electrodes can simply be adjusted to compensate for unwanted domain spreading. For the waveform, the pulse voltage and duration were the most important. At voltages near Ec we observed low yields, with some domains failing to nucleate. Raising the voltage steadily improved the yield while also causing the domain width to grow. Similarly, the domain width is largely controlled by the pulse duration, however, if the pulse is too short, the individual filaments do not merge to form uniform domains. Multiple pulses are commonly employed as they have been reported to suppress sidewise domain growth compared to a single pulse with the same total duration [42]. The time interval between pulses is thought to reduce current induced heating effects. Our measurements, however, showed that the amount of domain spreading, as imaged by PFM, was comparable with single and multiple pulses, given our SiO2 insulation layer which reduces the current and our relatively short pulse durations. Lastly, it is reported in bulk LN that the poling can spontaneously backswitch if the voltage is reduced too quickly [43], however, we did not observe this effect with ramp-down times as short as 250 µs. The absence of backswitching was also observed in undoped z-cut LNOI and it was theorized that domain pinning by the interface of the thin film and substrate was responsible [26]. Nevertheless, we continued to use an asymmetrical pulse design with longer ramp-down than ramp-up times with backswitching in mind.
Overall, our approach to optimizing the poling is to first adjust the voltage of a single pulse to achieve 100% poling yield. Then with the voltage fixed, tune the pulse duration such that it is long enough to form uniform domains, but short enough to avoid significant domain widening and merging. Finally, the duty cycle of the electrodes is adjusted to compensate for any remaining domain spreading.
3.1 Poling
After optimization, 50% duty cycle poling is achieved on MgO:LN thin films with a single 550 V pulse with 1 ms rise time, 1 ms duration, and 5 ms ramp-down applied to electrodes with a 45% duty cycle and round tips. Due to the short electrode length, L in Fig. 1(a), the small poling current is easily masked by the stray capacitive charging of the circuit. To isolate the poling current, we averaged the current without contacting the sample over 100 pulses and subtracted it from the measured current. A 25-point moving-average filter is applied to smooth some of the noise induced by this subtraction. The open-circuit subtracted current and associated voltage waveform in Fig. 2(a) show that the poling current begins abruptly during the voltage ramp-up, near 2.5 ms. We have observed that the magnitude of this current spike corresponds to the amount of poling, with larger current showing an increase in the number and/or size of poled regions. Integrating the entire current waveform in Fig. 2(a) shows 45 pC has been transferred, which is in agreement with the predicted charge of 46 pC given by Q = 2PsA [44] (spontaneous polarization: Ps = 80 µC/cm2 [45], poled area: A = 11 domains × 7.5 µm period × 50% duty cycle × 700 nm thick film = 29 µm2). We systematically extract the voltage at which the current spike begins, Vth, from 26 devices on 3 different samples, all with 15 µm electrode gap size, and find it to be 439.5 ± 16.5 V (values are mean and standard deviation). The threshold voltage is found to be the same on LNOI test devices with 4 µm thick BOX and SHG devices with 1 µm BOX. The PFM image in Fig. 2(b) shows uniform poling with 50% duty cycle. In this image, the dark regions have been inverted and Cr electrode tips are visible at the top and bottom.
Fig. 2. Poling waveforms and piezoresponse force microscopy amplitude image (a.u.) of domains on MgO:LN (a)-(b) thin film, (c)-(d) bulk with SiO2 insulation, and (e)-(f) bulk without SiO2 insulation. Poling waveforms show voltage (VLN) and current (Ipol) after open-circuit subtraction. The piezoresponse force microscopy images are taken on the +x surface, where the dark regions have been inverted. Chromium electrode tips are visible at the top and bottom.
The same poling waveform was also applied to 1 mm thick bulk MgO:LN samples with and without the 100 nm SiO2 insulation, Fig. 2(c)-2(d) and 2(e)-2(f) respectively. The electrodes on both bulk samples have a 10 µm period, while the sample with SiO2 has a 52% duty cycle and the one without SiO2 has a 65% duty cycle. First, comparing the bulk sample with SiO2 to LNOI, Fig. 2(c)-2(d) vs. 2(a)-2(b), shows comparable domain quality but a few differences in poling current. The bulk sample has an increased capacitive charging current during the voltage ramp-up, before the large poling spike. The increased capacitance of the electrodes on bulk is expected as they have a larger duty cycle and the permittivity of LN is higher than the buried oxide layer present in LNOI. The bulk sample also shows a larger peak current and total transferred charge (90 pC) that is 1.7 times larger than LNOI after adjusting for the difference in electrode duty cycle which could suggest deeper poling. Lastly, the threshold voltage of bulk with SiO2 was measured on 13 devices from 2 different samples and found to be 436.5 ± 13.5 V, which is equivalent to LNOI within the standard deviation of our measurements. In contrast, the threshold voltage reported in undoped z-cut LNOI was 60% larger than its bulk counterpart [37]. Furthermore, the field strength actually used for poling undoped x- and y-cut LNOI was significantly larger than the coercive field of bulk LN (∼21 V/µm), ranging from ∼40−48 V/µm [14,30]. It was thought that the LN / buried SiO2 interface [14], or possibly the LNOI fabrication process itself, specifically Li+ out-diffusion during annealing [37], could be responsible for the increased Ec.
Next, comparing the two bulk samples highlights the importance of reducing the leakage current with the SiO2 insulation in order to achieve high quality poling. Figure 2(e) shows that without the SiO2, a very large current flows resulting in a total transferred charge of 576 µC, which is over 3 orders of magnitude larger than with SiO2. The large current causes substantial voltage drops across RS and R3 which reduces the applied voltage to less than 200 V. The PFM image, Fig. 2(f), displays non-uniform domains with numerous filaments. Many different waveforms, including larger voltage, longer durations, and multiple pulses were used to try to fill in these domains, but all attempts resulted in similarly non-uniform domains. It is evident that domain broadening is not possible after domains form the low resistance path. In contrast, with SiO2 the leakage current is reduced, Fig. 2(c), and the individual filaments become uniform domains, Fig. 2(d). The threshold voltage of 5 devices without SiO2 is measured to be 271 ± 3 V, which is 38% less than those with SiO2. The difference in Vth between samples with and without SiO2 is in good agreement with static electric field modeling performed using the finite element method (COMSOL). The simulations showed that the SiO2 insulation reduces the voltage across the LN by 33%, where the voltage has been evaluated at the electrode tips, 10 nm below the LN surface. The threshold voltage has also been measured using pulses with longer rise times. MgO:LN is a slowly responding material, causing Ec to be a function of the voltage ramping rate [46]. By increasing the rise time from 1 ms to 400 ms we find Vth decreases to 147 V. All of the threshold voltage measurements are summarized in Table 1.
Table 1. Summary of measured threshold voltage (Vth) with 15 µm electrode gap size
3.2 Second harmonic generation
For further characterization of the poling, waveguide SHG is performed in MgO doped LNOI loaded with a SiN strip. The waveguide is designed to support the fundamental TE0 modes, which access LN’s nonlinear coefficient d33, at a pump wavelength of 1550 nm and second harmonic (SH) output at 775 nm. The optical modes are calculated using the semi-vectorial beam propagation method (BPM) and are used to find the waveguide cross-section with strong modal overlap between the pump and SH modes while also maintaining most of the optical power within the LN film. In the simulations, the refractive indices for MgO:LN [47] and SiO2 [48] are based on available Sellmeier equations while the SiN has been measured in-house with spectroscopic ellipsometry. Specifically, the indices of MgO:LN (extraordinary), SiO2 and SiN are set to 2.13, 1.44, and 2.16 at 1550 nm and 2.17, 1.45 and 2.20 at 775 nm respectively. The mode profiles of the final design are shown in Fig. 3(b) and 3(c) with a 700 nm thick LN film and 200 nm × 1.3 µm SiN strip. The pump has an effective index nω of 2.00, mode area (1/e field cutoff) of 2.1 µm2, and contains 88% of the power in the LN film. The SH mode index n2ω is 2.13, the mode area is 1.3 µm2, and has 90% power in the LN film. Based on the effective indices, the poling period for first-order QPM is 6.0 µm (Λ =λ2ω / (n2ω - nω)). Figure 3(d) shows the fabricated device, which has been coated with a Cr anti-charging layer for scanning electron microscopy (SEM) imaging.
Fig. 3. (a) Generated second harmonic power as a function of input pump power (a.u.) showing characteristic quadratic scaling. Insert shows PFM amplitude image (a.u.) on +x surface; the SiN waveguide can be seen between the electrodes. (b) and (c) Simulated TE0 modes profiles (Ez component) at pump (1550 nm) and second harmonic (775 nm) wavelengths. (d) SEM image of fabricated device coated with Cr anti-charging layer.
Poling is performed with the waveguide in place using the optimized electrodes and waveform design described in section 3.1. The poling current exhibits a similar shape as seen in Fig. 2(a), but with larger amplitude due to the 25 times longer electrode length. Furthermore, the PFM image seen in the insert of Fig. 3(a) shows that we achieve the same high-quality poling as the test devices, Fig. 2(b), even with the presence of the SiN waveguide, the increased electrode length, and reduced BOX thickness. The domains remained stable over a four month observation period during which the sample is stored at room temperature. The 15 µm electrode gap size has been designed large enough to avoid optical absorption and therefore they are not removed for optical measurements.
Measurements are performed using a continuous-wave tunable telecom laser as the pump source. The light is passed through a polarization controller to excite the TE mode and then enters an erbium-doped fiber amplifier. The light is coupled on and off the chip using tapered lensed fibers at the polished end faces. On the output, the pump and SH signals are separated with a fiber-based wavelength division multiplexer and detected with InGaAs and Si photodetectors, respectively. The signals are also measured with an optical spectrum analyzer. The maximum SH power occurs at a 1532 nm pump wavelength in a device with a 6.00 µm poling period. Figure 3(a) shows the SH power scales quadratically with pump power, in agreement with theory. The characteristic power scaling confirms quasi-phase matched SHG is enabled by our poling process.
3.3 Poling depth
The domain depth in the SHG device is determined by FIB milling a trench through the entire LN film thickness, exposing the ± z faces, followed by HF etching. FIB milling is performed using Ga+ ions. The trenches are 1.3 µm wide x 25 µm long and located near the center of the electrode gap such that it partially overlaps with the SiN strip as illustrated in the inset of Fig. 4. The sample was coated with a Cr anti-charging layer for milling and was subsequently removed with chromium etchant. Afterwards, the sample is etched in 49% HF for 5 min which attacks the -z surface faster than +z, revealing the poling domains. During the etching, the buried oxide is undercut and the film collapses down onto the LN substrate. Also, the SiN strip is attacked, leaving only a little residue visible in its place. SEM images show that the domains penetrate ∼65% through the film thickness with a 50% duty cycle, Fig. 4(a). An additional device poled with two identical pulses shows domains penetrating the entire film thickness with a 55% duty cycle, Fig. 4(b).
Fig. 4. Scanning electron microscopy images of FIB milled cross-sections through center of SiN strip after HF etching. The sample is tilted ∼45°. Inset shows location of milled trench on +x face. (a) Device poled with one pulse shows domains penetrate ∼65% through the LN film. (b) Poling with two pulses shows domains penetrate entire depth.
4. Conclusion
We have presented methods for the fabrication and periodic poling of ion-sliced x-cut MgO:LN thin films with a silicon dioxide insulation layer under co-planar poling electrodes to reduce leakage current. Poling is performed by applying high voltage pulses to the patterned electrodes on top of the thin film. Uniform domains are achieved with periods as small as 6 µm. Poling of bulk MgO:LN, with and without the SiO2 insulation, is also performed and compared to thin films in terms of threshold voltage, poling current, and domain quality as imaged by PFM. The SiO2 insulation layer is shown to reduce the poling current by over 3 orders of magnitude and allow for filaments to broaden and form uniform domains. The threshold voltage of 5 mol% MgO doped LN thin films are found to be the same as bulk companions. In bulk samples, the SiO2 insulation layer is found to increase the threshold voltage by 38%. The poling of thin films is also used to demonstrate waveguide second harmonic generation in the LN thin film loaded by a SiN strip with 1550 nm pump wavelength. Lastly, the poling depth is investigated by FIB milling and HF etching. The etching revealed that two 550 V pulses, with 1 ms rise time, 1 ms duration, and 5 ms ramp-down time, are required for the domains to penetrate the entire film thickness. These results pave a route to compact and efficient nonlinear optical devices in MgO doped LN.
Funding
National Science Foundation (NSF) (1809894).
References
1. A. Boes, B. Corcoran, L. Chang, J. Bowers, and A. Mitchell, “Status and potential of lithium niobate on insulator (LNOI) for photonic integrated circuits,” Laser Photonics Rev. 12(4), 1700256 (2018). [CrossRef]
2. A. Rao and S. Fathpour, “Heterogeneous thin-film lithium niobate integrated photonics for electrooptics and nonlinear optics,” IEEE J. Sel. Top. Quantum Electron. 24(6), 1–12 (2018). [CrossRef]
3. S. Bogdanov, M. Y. Shalaginov, A. Boltasseva, and V. M. Shalaev, “Material platforms for integrated quantum photonics,” Opt. Mater. Express 7(1), 111–132 (2017). [CrossRef]
4. M. Vainio and L. Halonen, “Mid-infrared optical parametric oscillators and frequency combs for molecular spectroscopy,” Phys. Chem. Chem. Phys. 18(6), 4266–4294 (2016). [CrossRef]
5. A. E. Willner, S. Khaleghi, M. R. Chitgarha, and O. F. Yilmaz, “All-optical signal processing,” J. Lightwave Technol. 32(4), 660–680 (2014). [CrossRef]
6. M. Leidinger, S. Fieberg, N. Waasem, F. Kühnemann, K. Buse, and I. Breunig, “Comparative study on three highly sensitive absorption measurement techniques characterizing lithium niobate over its entire transparent spectral range,” Opt. Express 23(17), 21690–21705 (2015). [CrossRef]
7. E. H. Turner, D. Chen, and G. N. Otto, “High-frequency electro-optic coefficients of lithium niobate,” Appl. Phys. Lett. 8(11), 303–304 (1966). [CrossRef]
8. M. M. Choy and R. L. Byer, “Accurate second-order susceptibility measurements of visible and infrared nonlinear crystals,” Phys. Rev. B 14(4), 1693–1706 (1976). [CrossRef]
9. K. Mizuuchi, A. Morikawa, T. Sugita, and K. Yamamoto, “Electric-field poling in Mg-doped LiNbO3,” J. Appl. Phys. 96(11), 6585–6590 (2004). [CrossRef]
10. Y. Furukawa, K. Kitamura, A. Alexandrovski, R. K. Route, M. M. Fejer, and G. Foulon, “Green-induced infrared absorption in MgO doped LiNbO3,” Appl. Phys. Lett. 78(14), 1970–1972 (2001). [CrossRef]
11. Y. Furukawa, K. Kitamura, S. Takekawa, A. Miyamoto, M. Terao, and N. Suda, “Photorefraction in LiNbO3 as a function of [Li]/[Nb] and MgO concentrations,” Appl. Phys. Lett. 77(16), 2494–2496 (2000). [CrossRef]
12. P. Rabiei and P. Gunter, “Optical and electro-optical properties of submicrometer lithium niobate slab waveguides prepared by crystal ion slicing and wafer bonding,” Appl. Phys. Lett. 85(20), 4603–4605 (2004). [CrossRef]
13. R. Geiss, S. Saravi, A. Sergeyev, S. Diziain, F. Setzpfandt, F. Schrempel, R. Grange, E.-B. Kley, A. Tünnermann, and T. Pertsch, “Fabrication of nanoscale lithium niobate waveguides for second-harmonic generation,” Opt. Lett. 40(12), 2715–2718 (2015). [CrossRef]
14. L. Chang, Y. Li, N. Volet, L. Wang, J. Peters, and J. E. Bowers, “Thin film wavelength converters for photonic integrated circuits,” Optica 3(5), 531–535 (2016). [CrossRef]
15. L. Cai, Y. Wang, and H. Hu, “Efficient second harmonic generation in χ(2) profile reconfigured lithium niobate thin film,” Opt. Commun. 387, 405–408 (2017). [CrossRef]
16. A. Rao and S. Fathpour, “Second-harmonic generation in integrated photonics on silicon,” Phys. Status Solidi A1700684 (2017).
17. C. Wang, C. Langrock, A. Marandi, M. Jankowski, M. Zhang, B. Desiatov, M. M. Fejer, and M. Lončar, “Ultrahigh-efficiency wavelength conversion in nanophotonic periodically poled lithium niobate waveguides,” Optica 5(11), 1438–1441 (2018). [CrossRef]
18. J.-Y. Chen, Y. M. Sua, H. Fan, and Y.-P. Huang, “Modal phase matched lithium niobate nanocircuits for integrated nonlinear photonics,” OSA Continuum 1(1), 229–242 (2018). [CrossRef]
19. R. Luo, Y. He, H. Liang, M. Li, and Q. Lin, “Highly tunable efficient second-harmonic generation in a lithium niobate nanophotonic waveguide,” Optica 5(8), 1006–1011 (2018). [CrossRef]
20. C. Wang, X. Xiong, N. Andrade, V. Venkataraman, X.-F. Ren, G.-C. Guo, and M. Lončar, “Second harmonic generation in nano-structured thin-film lithium niobate waveguides,” Opt. Express 25(6), 6963–6973 (2017). [CrossRef]
21. G. Li, Y. Chen, H. Jiang, and X. Chen, “Broadband sum-frequency generation using d33 in periodically poled LiNbO3 thin film in the telecommunications band,” Opt. Lett. 42(5), 939–942 (2017). [CrossRef]
22. A. Rao, N. Nader, M. J. Stevens, T. Gerrits, O. S. Magaña-Loaiza, F. Camacho-González, J. Chiles, A. Honardoost, M. Malinowski, R. Mirin, and S. Fathpour, “Photon pair generation on a silicon chip using nanophotonic periodically-poled lithium niobate waveguides,” in Conference on Lasers and Electro-Optics (Optical Society of America, 2018), p. JTh3C.2.
23. J. Jiang, X. J. Meng, D. Q. Geng, and A. Q. Jiang, “Accelerated domain switching speed in single-crystal LiNbO3 thin films,” J. Appl. Phys. 117(10), 104101 (2015). [CrossRef]
24. K. Tanaka and T. Suhara, “Fabrication of domain inverted ridge waveguide in ion-sliced LiNbO3 for wavelength conversion devices,” in Microoptics Conference (MOC), 2015 20th (IEEE, 2015), pp. 1–2.
25. G. Shao, Y. Bai, G. Cui, C. Li, X. Qiu, D. Geng, D. Wu, and Y. Lu, “Ferroelectric domain inversion and its stability in lithium niobate thin film on insulator with different thicknesses,” AIP Adv. 6(7), 075011 (2016). [CrossRef]
26. R. V. Gainutdinov, T. R. Volk, and H. H. Zhang, “Domain formation and polarization reversal under atomic force microscopy-tip voltages in ion-sliced LiNbO3 films on SiO2/LiNbO3 substrates,” Appl. Phys. Lett. 107(16), 162903 (2015). [CrossRef]
27. T. Volk, R. Gainutdinov, and H. Zhang, “Domain patterning in ion-sliced LiNbO3 films by atomic force microscopy,” Crystals 7(5), 137 (2017). [CrossRef]
28. D. Djukic, G. Cerda-Pons, R. M. Roth, R. M. Osgood Jr, S. Bakhru, H. Bakhru, et al., “E/O tunable second-harmonic-generation gratings in ion-exfoliated thin films of periodically poled LiNbO3,” in Conference on Lasers and Electro-Optics (Optical Society of America, 2007), p. CFN3.
29. G. Poberaj, H. Hu, W. Sohler, and P. Günter, “Lithium niobate on insulator (LNOI) for micro-photonic devices,” Laser Photonics Rev. 6(4), 488–503 (2012). [CrossRef]
30. A. Rao, M. Malinowski, A. Honardoost, J. R. Talukder, P. Rabiei, P. Delfyett, and S. Fathpour, “Second-harmonic generation in periodically-poled thin film lithium niobate wafer-bonded on silicon,” Opt. Express 24(26), 29941–29947 (2016). [CrossRef]
31. P. Mackwitz, M. Rüsing, G. Berth, A. Widhalm, K. Müller, and A. Zrenner, “Periodic domain inversion in x-cut single-crystal lithium niobate thin film,” Appl. Phys. Lett. 108(15), 152902 (2016). [CrossRef]
32. E. A. Eliseev, A. N. Morozovska, G. S. Svechnikov, V. Gopalan, and V. Ya. Shur, “Static conductivity of charged domain walls in uniaxial ferroelectric semiconductors,” Phys. Rev. B 83(23), 235313 (2011). [CrossRef]
33. C. S. Werner, S. J. Herr, K. Buse, B. Sturman, E. Soergel, C. Razzaghi, and I. Breunig, “Large and accessible conductivity of charged domain walls in lithium niobate,” Sci. Rep. 7(1), 9862 (2017). [CrossRef]
34. M. Schröder, A. Haußmann, A. Thiessen, E. Soergel, T. Woike, and L. M. Eng, “Conducting Domain Walls in Lithium Niobate Single Crystals,” Adv. Funct. Mater. 22(18), 3936–3944 (2012). [CrossRef]
35. N. Horikawa, T. Tsubouchi, M. Fujimura, and T. Suhara, “Formation of domain-inverted grating in MgO:LiNbO3 by voltage application with insulation layer cladding,” Jpn. J. Appl. Phys. 46(8A), 5178–5180 (2007). [CrossRef]
36. L. Liang, F. Wang, Y. Sang, F. Zhou, X. Xie, D. Sun, M. Zheng, and H. Liu, “Facile approach for the periodic poling of MgO-doped lithium niobate with liquid electrodes,” CrystEngComm 21(6), 941–947 (2019). [CrossRef]
37. L. Gui, “Periodically poled ridge waveguides and photonic wires in LiNbO3 for efficient nonlinear interactions,” Ph.D. thesis, University of Paderborn (2010).
38. T. R. Volk, R. V. Gainutdinov, and H. H. Zhang, “Domain-wall conduction in AFM-written domain patterns in ion-sliced LiNbO3 films,” Appl. Phys. Lett. 110(13), 132905 (2017). [CrossRef]
39. J. F. Ziegler, M. D. Ziegler, and J. P. Biersack, “SRIM – The stopping and range of ions in matter (2010),” Nucl. Instrum. Methods Phys. Res., Sect. B 268(11-12), 1818–1823 (2010). [CrossRef]
40. H. Han, L. Cai, and H. Hu, “Optical and structural properties of single-crystal lithium niobate thin film,” Opt. Mater. 42, 47–51 (2015). [CrossRef]
41. T. Jungk, Á. Hoffmann, and E. Soergel, “Contrast mechanisms for the detection of ferroelectric domains with scanning force microscopy,” New J. Phys. 11(3), 033029 (2009). [CrossRef]
42. T. Sugita, K. Mizuuchi, Y. Kitaoka, and K. Yamamoto, “Ultraviolet light generation in a periodically poled MgO: LiNbO3 waveguide,” Jpn. J. Appl. Phys. 40(Part 1), 1751–1753 (2001). [CrossRef]
43. R. G. Batchko, V. Y. Shur, M. M. Fejer, and R. L. Byer, “Backswitch poling in lithium niobate for high-fidelity domain patterning and efficient blue light generation,” Appl. Phys. Lett. 75(12), 1673–1675 (1999). [CrossRef]
44. T. Suhara and M. Fujimuura, Waveguide Nonlinear-Optic Devices (Springer, 2003).
45. M.-L. Hu, L.-J. Hu, and J.-Y. Chang, “Polarization switching of pure and MgO-doped lithium niobate crystals,” Jpn. J. Appl. Phys. 42(Part 1), 7414–7417 (2003). [CrossRef]
46. H. Ishizuki and T. Taira, “Study on the field-poling dynamics in Mg-doped LiNbO3 and LiTaO3,” in Nonlinear Optics: Materials, Fundamentals and Applications (Optical Society of America, 2007), p. WE35.
47. O. Gayer, Z. Sacks, E. Galun, and A. Arie, “Temperature and wavelength dependent refractive index equations for MgO-doped congruent and stoichiometric LiNbO3,” Appl. Phys. B: Lasers Opt. 91(2), 343–348 (2008). [CrossRef]
48. I. H. Malitson, “Interspecimen comparison of the refractive index of fused silica,” J. Opt. Soc. Am. B 55(10), 1205–1209 (1965). [CrossRef]
