We demonstrate fine-pitch periodic poling of bulk Rb-doped KTiOPO4 using a coercive field grating induced by ion-exchange. These samples were used for quasi phase matched (QPM) second harmonic generation at 795 nm with a normalized conversion efficiency of 1.7% W−1cm−1. Additionally, the ion-exchange introduced a refractive index change over the sample thickness resulting in a phasematching wavelength shift of 0.21 nm, adding extra tunability to the QPM device.
© 2016 Optical Society of America
KTiOPO4 (KTP) has been an important crystal for nonlinear optics since more than 30 years . With the development of periodic poling the nonlinear response can be tailored for desired processes, which has further increased its usefulness [2,3]. However, despite a low coercive field, Ec ~2kV/mm, it is not straight forward to periodically pole the commonly available flux grown KTP as it has a high spatially varying ionic conductivity, σion, which screens the applied electric field during poling. The σion originates from the special crystallographic structure consisting of open chiral chains along the c-axis. In these chains the potassium ions hop over potassium vacancies, , originating from crystal growth  and this spatial variation in σion makes it difficult to pole larger pieces of crystals with good yield .
Recently Rb-doped KTiOPO4 (RKTP) appeared as an attractive alternative to KTP for periodic poling , as it presents a σion two order of magnitude lower than that of KTP. Furthermore, with a low Rb-doping (<0.3%) the advantageous linear and nonlinear optical properties of undoped KTP are preserved, at the same time as this material shows an improved grey-tracking resistance [7,8]. Fabrication of high aspect ratio domain gratings in RKTP has been achieved by conventional periodic poling techniques, demonstrating 5mm-thick highly efficient quasi-phase matching (QPM) devices , as well as dense domain structures for deep-blue light generation . Nevertheless, engineering of sub-µm periodicity QPM gratings for applications involving generation of counter-propagating photons [10,11], still remains challenging.
In the late 80’s and early 90’s, ion-exchange processes were developed for KTP resulting in high-quality waveguides. The exchange is typically performed by immersing the crystal in a molten monovalent nitrate melt where the larger ions (i.e. Rb+-, Cs+- or Tl+-) are exchanged with the smaller K+-ion, creating the desired refractive index increase and at the same time lowering the σion . The exchange is highly anisotropic, happening exclusively in the polar direction through the open channels resulting in a composition MxK1-xTiOPO4, where x is the fraction of sites in the crystal occupied by the dopant ion M+. The ion-exchange follows a normal one-dimensional diffusion resulting in a complementary error function distribution for the M+-ion with a linear relation between the Rb+-concentration and the refractive index change, Δns. The refractive index profile n(z) can hence be written:
The vacancy becomes an additional hopping site for the monovalent ions, which increase the rate of exchange as well as the σion . Interestingly, ion-exchange can be used to engineer not only the refractive index and the σion, but also the coercive field. Risk and Lau , exploited the difference in Ec between the exchanged and non-exchanged regions to achieve periodically poled KTP waveguides for QPM second harmonic generation (SHG). However, this approach was never tested for fabrication of bulk periodically poled samples.
To explore alternative routes for domain engineering it can be attractive to fabricate an Ec grating by ion-exchange, since it potentially allows for poling with planar electrodes, alleviating the domain-broadening problem associated with the fringing fields using periodic metal electrodes . Moreover, since the ion-exchange is not intended for waveguiding, its depth can be made relatively large, which will also have an impact on the refractive indices of the bulk crystal.
In this work, we investigated the use of ion-exchange for fabrication of Ec gratings in RKTP and studied the impact of the exchange on both the electrical and optical properties of the bulk devices. We investigated three different ion-exchange conditions to find one which both reduced the conductivity in the exchanged region and made a large difference in Ec between exchanged and un-exchanged regions to promote improved periodic poling. A 35% increase in Ec and a 50% decrease in σion was obtained by ion-exchange in a RbNO3/Ba(NO3)2/KNO3 (73%/ 7%/ 20%) mixture. This was exploited to obtain a 1 mm thick periodically poled crystal with a period of 3.16 µm without domain broadening, resulting in a high normalized conversion efficiency of 1.7%W−1cm−1 for QPM SHG. Furthermore, the process introduced an almost linear refractive index change of 10−4 across the sample thickness, which added extra tunability to the QPM device.
Deep ion-exchange in RKTP and the formation of Ec gratings
First, we investigated the impact of the melt composition on the Ec and the σion of the ion-exchanged crystals. We compared three different recipes with varying ratios of Rb and K, whereas the concentration of Ba(NO3)2 was kept constant: A “Rb-rich” composition (73 mol% RbNO3/ 7 mol% Ba(NO3)2/ 20 mol% KNO3);a “half-Rb-K” composition (47 mol % RbNO3/7 mol % Ba(NO3)2/46 mol % KNO3) and a “Rb-poor” composition (20 mol % RbNO3/7 mol % Ba(NO3)2/73 mol % KNO3).
For our experiments, we used c-cut RKTP crystals of dimensions 5 × 5 × 1 mm (a × b × c crystallographic axes), respectively, with 0.3 atomic % Rb in the as-grown crystals. First, we created an ion-exchange stop layer on the c+ faces of the crystals since we wanted to promote ion-exchange exclusively from the c- faces. The reason being that the c- face is the preferred surface for patterning -and thus for ion-exchange- owing to the fact that domain nucleation in KTP isomorphs occurs preferentially on this polar face . The stop-layer was created by exposing the c+ faces to oxygen plasma in a reactive ion etching machine (20 sccm O2, 100W at 100 mTorr for 4 × 7 s). Although the physical mechanism behind this process has not been fully investigated, it is assumed that the plasma ions damage the sub-surface structure and block in- and out- diffusion of ions. In fact, it has been previously reported that an argon plasma can be used for this purpose . Afterwards, the ion-exchange was performed as follows: the samples were slowly heated to 375 °C, then immersed in the melt for 4 h, then lifted out of the melt and slowly cooled down to room temperature. Roelofs et al. have shown that the amount of vacancies formed during exchange in KTP depends on the Ba2+ concentration in the melt . As it was the same, 7%, for the three melts, we can assume that all the samples obtained approximately the same amount of vacancies after exchange. They also showed that the surface concentration of Rb-ions in the crystal will be approximately eight times lower than the Rb-concentration in the melt, at exchange equilibrium, for the Rb/K-ratios used here. This would mean that the surface Rb-concentration would be, approximately 47%, 13% and 3.5%, respectively, and the corresponding refractive index increase, Δnz ≈0.0062, 0,0025 and 0,001, respectively . To check this we measured the Rb- and K-content in the y-surface layer with energy-dispersive X-ray spectroscopy (EDX) on the Rb-rich planar sample, see Fig. 1. As it can be seen, approximately 43% of the K-ions are replaced by Rb at the surface and the exchange profile follows the expected erfc (z/d) with a depth, d = 14 µm. This is in reasonable agreement with the estimation from .
The Ec of the crystals was determined by measuring the current through the sample when the applied electric field (E-field) across the sample was increased with a linear ramp rate of 550V/ms. Figure 2 shows the applied field (black) and current traces of virgin and ion-exchanged samples using the three different exchange melts (virgin (green), “Rb-poor” (blue), “half-Rb-K” (purple) and “Rb-rich” (red)). The ionic-current increased quadratically with the applied field in accordance with the Mott-Gurney law , while the polarization-switching appeared as a current peak at different fields for the three samples that poled. This is a well-known response for KTP isomorphs . Ec is defined here as the field at which the maximum of the switching-current peak occurs.
Crystals exchanged in the “Rb-rich” composition, obtained the largest increase in Ec (1.7 kV/mm) compare to the virgin samples, while the σion at the same time decreased by almost 50%. Crystals exchanged in the “half-Rb-K” melt showed an increase of 1.06 kV/mm in Ec, but also an 85% increase in σion, whereas the samples exchanged in the Rb-poor melt got a so large increase in σion that no domain switching could be obtained. In fact, we tried several different pulse shapes, pulse length and magnitudes, but no polarization reversal could be achieved in any of these crystals before reaching catastrophic breakdown. Table 1 summarizes the results.
From these results we see that Ba increased, while Rb decreased, the σion, as expected. The desired condition for poling, a large Ec increase and a simultaneous reduction in conductivity was obtained with the “Rb-rich” melt only. Next, we studied the effect of the exchange time for the “Rb-rich” recipe on the coercive field and the σion, see Fig. 3. An increase in Ec of 60% (ΔEc~3 kV/mm) was seen for exchange times of 8 hours and longer. At the same time the σion was reduced by ~80%. This is also consistent with the saturation in σion observed by Roelofs et al.  for KTP.
Periodic poling with Ec gratings
In order to perform periodic poling, RKTP crystals of dimensions 11 × 6 × 1 mm were patterned with 8x4 mm2 photoresist gratings with a period Λ = 3.16 µm and a 33/67 duty cycle on the c--surface using standard photolithography. A stop layer was created in the photoresist openings as well as in the c+-faces by oxygen plasma bombardment. Subsequently, the photoresist was removed. Long ion-exchange times with a periodic grating could create excessive stress in the crystal lattice due to the approximately 1% larger cell unit for RTP relative to KTP that might lead to breakdown during domain reversal. Therefore, we limited, the ion-exchange in the Rb-rich melt to 4 h. Figure 4 displays the resulting refractive-index grating, which had a visual depth of ~70 µm and was uniform over the whole sample. Afterwards, the crystals were connected to the external electrical circuit using planar liquid electrodes, and periodically poled at room temperature by applying single 5 ms long triangular pulses of magnitudes varying between 7 and 8 kV/mm.
After poling the samples were selectively etched in a KNO3:KOH solution in order to reveal the domain structure. This solution attacks the domains c--surface but leaves the c+-surface relatively untouched .It also results in contrast between the exchanged and the non-exchanged regions since it etches the non-exchanged regions slightly faster than the exchanged regions. Figure 5 shows optical micrographs of the former (a) c--, and (b) c+-faces. As expected, domain reversal occurred in the non-exchanged regions, as can be clearly observed on the border between the poled and non-poled area in Fig. 5(a). No lateral broadening of the reversed domains into the ion-exchanged regions can be observed. Note that domain duty-cycle is maintained over the whole crystal thickness, as can be seen from the corresponding domain pattern on the former c+-face in Fig. 5(b).
Finally, we evaluated the performance of a representative periodically poled crystal as a QPM device in a single-pass SHG experiment. For that purpose, a z-polarized continuous wave (cw) Ti:sapphire laser beam was launched along the x-direction of the crystal and confocally focused with a 50 mm lens to a beam waist of 25 µm (1/e2 intensity). A waveplate-polarizer arrangement was used to adjust the pump power. The uncoated crystal was kept at a constant temperature of 40°C using a Peltier element. Close to the exchanged surface, i.e. the waveguide region, we obtained limited SHG over a broad wavelength range. This was because the fundamental power was distributed over very many fundamental modes, with little power in each one of them, and every mode has its own effective index and corresponding phasematching wavelength altogether resulting in low conversion efficiency. However, below the exchanged region (i.e., ~70 µm below the patterned face) the conversion was uniform and efficient, and 3.4 mW of blue light was generated with 520 mW of fundamental power at 795 nm. This corresponds to a normalized conversion efficiency of 1.7% W−1 cm−1 and an estimate of deff = 10 pm/V, close to the optimal deff for KTP . This result is similar to that obtained by Zukauskas et al for RKTP  in which periodic poling was achieved using conventional patterned electrodes.
Nevertheless, the ion-exchange process has an impact in the crystal bulk that manifests itself as a refractive index gradient throughout the crystal. We evaluated this effect by translating the sample in the z-direction in steps of 50 µm well below the exchanged region (~70 µm) while measuring the SHG wavelength, see Fig. 6(a). The phase matching wavelength changed approximately linearly with distance in the polar direction. The step-like shape of the curve is an artifact of limited resolution (0.05 nm) of the optical spectrum analyzer. The total change in the fundamental QPM wavelength was Δλ = 0.21 nm. The refractive index of the ion-exchanged regions can be calculated from the QPM condition, taking into account the Ec-grating duty-cycle and using the KTP Sellmeier equations by Fan et al.  with temperature corrections by Weichmann et al. . The change in the fundamental QPM wavelength then corresponds to a total change of the difference in refractive index as Δn(2ω)z -Δn(ω)z = 4·10−4 over the scanned region. The change in SH wavelength was further verified by measuring the corresponding temperature-tuning curves at the same positions along the polar direction while keeping the pump wavelength fixed at 784.98 nm. The tuning curves measured at 100 µm (red) and 600 µm (black) below the patterned face are displayed in Fig. 6(b). Note that both curves have a full width half maximum of ΔTFWHM = 1.9 °C, confirming that a high domain-grating homogeneity is maintained over the crystal depth. The QPM condition shifts towards higher temperatures the further away it gets from the ion-exchanged surface. The total shift in temperature is ΔT = 2.3 °C, which corresponds to a wavelength shift of Δλ = 0.13 nm, in good agreement with the data in Fig. 6(a).
From these measurements it is clear that the ion-exchange process induces a shift in the QPM condition that can be exploited to further expand the tuning capabilities of bulk PPRKTP and possibly other crystals with deep ion-exchange. Careful tailoring of the ion-exchange parameters allows for engineering of the refractive index gradient. Moreover, when both the linear and the nonlinear properties are modulated in the bulk crystal, providing simultaneously a χ(1) and a χ(2) grating, novel features could be exploited. As an example, KTP waveguides segmented χ(1) gratings has been used to Bragg-lock diode lasers on the QPM peak for stable SHG [22,23]. Similar technique could be utilized with the type of structure obtained here.
In summary, we have demonstrated deep ion-exchange in RKTP and used it to create Ec and σion gratings. With proper exchange conditions we obtained gratings with a period of 3.16 µm in 1 mm thick RKTP samples with Δ Ec and σion differences between the exchanged and non-exchanged regions of 35% and 50%, respectively. These samples were periodically poled and used for efficient, single-pass, cw QPM blue light SHG with a normalized conversion efficiency of 1.7%W−1cm−1. Moreover, the ion-exchange process introduced a linear stress-induced refractive index change below the exchanged region, which translated into a SHG wavelength shift of 0.21 nm, adding tunability to the nonlinear process. We believe this new technique can be used to fabricate high quality periodically poled devices with high grating aspect ratios and additionally built in tunability.
This work was supported by the Swedish Foundation for Strategic Research and the Swedish Research Council (VR) through its Linnæus Center of Excellence ADOPT.
References and links
1. J. D. Bierlein and H. Vanherzeele, “Potassium titanyl phosphate: properties and new applications,” J. Opt. Soc. Am. B 6(4), 622–633 (1989). [CrossRef]
2. H. Karlsson and F. Laurell, “Electric field poling of flux grown KTiOPO4,” Appl. Phys. Lett. 71(24), 3474–3476 (1997). [CrossRef]
3. Q. Chen and W. P. Risk, “Periodic poling of KTiOPO4 using an applied electric field,” Electron. Lett. 30(18), 1516–1517 (1994). [CrossRef]
4. P. A. Morris, A. Ferretti, J. D. Bierlein, and G. M. Loiacono, “Reduction of the ionic conductivity of flux grown KTiOPO4 crystals,” J. Cryst. Growth 109(1–4), 361–366 (1991). [CrossRef]
5. C. Canalias, J. Hirohashi, V. Pasiskevicius, and F. Laurell, “Polarization-switching characteristics of flux-grown KTiOPO4 and RbTiOPO4 at room temperature,” J. Appl. Phys. 97(12), 124105 (2005). [CrossRef]
6. Q. Jiang, P. A. Thomas, K. B. Hutton, and R. C. C. Ward, “Rb-doped potassium titanyl phosphate for periodic ferroelectric domain inversion,” J. Appl. Phys. 92(5), 2717–2723 (2002). [CrossRef]
7. S. Wang, V. Pasiskevicius, and F. Laurell, “High-efficiency frequency converters with periodically-poled Rb-doped KTiOPO4,” Opt. Mater. 30(4), 594–599 (2007). [CrossRef]
8. A. Zukauskas, V. Pasiskevicius, and C. Canalias, “Second-harmonic generation in periodically poled bulk Rb-doped KTiOPO₄ below 400 nm at high peak-intensities,” Opt. Express 21(2), 1395–1403 (2013). [CrossRef] [PubMed]
9. A. Zukauskas, N. Thilmann, V. Pasiskevicius, F. Laurell, and C. Canalias, “5 mm thick periodically poled Rb-doped KTP for high energy optical parametric frequency conversion,” Opt. Mater. Express 1(2), 201–206 (2011). [CrossRef]
10. C. Canalias and V. Pasiskevicius, “Mirrorless optical parametric oscillator,” Nat. Photonics 1(8), 459–462 (2007). [CrossRef]
11. A. Zukauskas, G. Strömqvist, V. Pasiskevicius, F. Laurell, M. Fokine, and C. Canalias, “Fabrication of submicrometer quasi-phase- matched devices in KTP and RKTP [ Invited ],” Opt. Mater. Express 1(7), 1319–1325 (2011). [CrossRef]
12. J. D. Bierlein, A. Ferretti, L. H. Brixner, and W. Y. Hsu, “Fabrication and characterization of optical waveguides in KTiOPO4,” Appl. Phys. Lett. 50(18), 1216–1218 (1987). [CrossRef]
13. W. P. Risk and S. D. Lau, “Periodic electric field poling of KTiOPO 4 using chemical patterning,” Appl. Phys. Lett. 69(26), 3999–4001 (1996). [CrossRef]
14. G. Rosenman, K. Garb, A. Skliar, M. Oron, D. Eger, and M. Katz, “Domain broadening in quasi-phase-matched nonlinear optical devices,” Appl. Phys. Lett. 73(7), 865–867 (1998). [CrossRef]
15. J. Gehler, W. Karthe, C. Wachter, A. Brauer, A. Rasch, and M. Rottschalk, “ARROW’s in KTiOPO4,” IEEE Photonics Technol. Lett. 9(4), 470–472 (1997). [CrossRef]
16. N. F. Mott and R. W. Gurney, Electronic Processes in Ionic Crystals (Oxford University Press, 1940).
17. M. G. Roelofs, P. A. Morris, and J. D. Bierlein, “Ion exchange of Rb, Ba, and Sr in KTiOPO4,” J. Appl. Phys. 70(2), 720–728 (1991). [CrossRef]
18. F. Laurell, M. G. Roelofs, W. Bindloss, H. Hsiung, A. Suna, and J. D. Bierlein, “Detection of ferroelectric domain reversal in KTiOPO4 waveguides,” J. Appl. Phys. 71(10), 4664–4670 (1992). [CrossRef]
20. T. Y. Fan, C. E. Huang, B. Q. Hu, R. C. Eckardt, Y. X. Fan, R. L. Byer, and R. S. Feigelson, “Second harmonic generation and accurate index of refraction measurements in flux-grown KTiOPO(4),” Appl. Opt. 26(12), 2390–2394 (1987). [CrossRef] [PubMed]
22. M. G. Roelofs, A. Suna, W. Bindloss, and J. D. Bierlein, “Characterization of optical waveguides in KTiOPO4 by second harmonic spectroscopy,” J. Appl. Phys. 76(9), 4999–5006 (1994). [CrossRef]
23. J. D. Bierlein, W. Bindloss, F. Laurell, and J. D. Lee, “Segmented Waveguiding having Selected Bragg Reflection Characteristics,” U.S. patent 5243676 (1993) and U.S. Patent 5393561 (1995).