Open Access
Add to BibSonomy
Abstract
A neodymium (Nd) doped lithium-niobate-on-insulator (LNOI) was fabricated from a lithium niobate (LN) substrate diffusion-doped with Nd prior to the LNOI fabrication process. Diffusion doping enables the fabrication of differently doped regions on a single LNOI substrate, a key advantage for future applications in highly integrated photonics. Fluorescence spectra, emission cross-section, fluorescence lifetime, and small-signal gain of Nd:LNOI were determined for the first time to our knowledge and compared with data published for diffusion-doped ridge or channel waveguides in LN substrates. For an 11 mm-long sample, we achieve a small-signal gain for π (σ) polarization of about 14 (10) dB when pumping with a Ti:Sapphire laser, which is promising for future active components in Nd:LNOI like amplifiers and lasers.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Successful wafer-scale fabrication of thin films of lithium-niobate-on-insulator (LNOI) and recent breakthroughs in nanofabrication techniques have made LNOI a rapidly advancing new material platform for a wide range of applications in integrated photonics [1]. Exploiting the advantages of this new platform has opened the door to the development of ultra-compact devices with superior properties, including electro-optic modulators, high-Q ring resonators, wavelength converters, or frequency combs [2–5]. On the other hand, active components in rare-earth-(RE)-doped conventional LiNbO3 (LN) substrates have attracted great attention not only for compact waveguide lasers [6–8] but also for their use as integrated-optical quantum memories [9]. The very low loss achievable in the LNOI platform, in combination with the high level of miniaturization, brings thus exciting new opportunities also for RE:LNOI. With the first availability of Er:LNOI fabricated using initially bulk-doped wafers, Er:LNOI ring lasers have been demonstrated very recently [10–12]. In another recent work, Tm:LNOI was fabricated, again based on the use of a volume-doped starting material [13].
Instead of the use of bulk-doped crystals, diffusion doping of wafers offers the full spectrum of spatially varying doping with the required (different) ions and individual concentrations for a particular device. For example only small areas of the substrate can be doped with ideal concentrations (e.g. the area of a ring resonator used as laser cavity), while other circuit sections (e.g. waveguides for feeding pump light, modulators, or splitters) may remain undoped. However, in-diffusion of RE ions, requiring temperatures of 1000 °C and beyond, into (undoped) LNOI substrates is hardly possible because of limited temperature stability of LNOI, where delamination of the thin films typically starts at temperatures around 600 °C. Recent attempts to fabricate RE-doped thin-film LN included the use of ion implantation with Er and Yb [14,15]. However in this approach, the limited temperature stability of LNOI severely restricts the required additional annealing treatments after ion implantation, in order to avoid clustering effects of implanted ions and to heal inherent lattice damage. As a result the spectroscopic properties of RE:LNOI doped by ion implantation are limited.
Here we report on a promising alternative to develop tailored RE-doped LNOI, that is in-diffusion into LN wafers prior to the implantation/bonding step. For the first time, to the best of our knowledge Nd:LNOI thin-film wafer pieces have been fabricated from diffusion-doped Nd:LN wafers, and the samples have been spectroscopically characterized for future use in active RE-doped LNOI devices.
2. Sample fabrication
To fabricate the Nd:LNOI the diffusion doping was done as follows: a 15 nm-thick Nd layer was deposited on a 3” z-cut wafer of undoped LN (YCC Yamaja Ceramics) by e-beam evaporation. The wafer was annealed for 200 h at 1393 K in a closed platinum vessel afterwards. Inside the vessel the sample was placed on a platinum web above a powder consisting of congruent LN to suppress Li out-diffusion. The diffusion-doped wafer was sent to NanoLN (Jinan Jingzheng Electronics Co., Ltd.) for LNOI fabrication by ion-slicing resulting in a 600 nm-thick Nd:LN film on a 2 µm-thick buried silica layer and a 0.5 mm-thick z-cut LN bottom substrate.
Using diamond blade dicing with a wafer saw (DAD322 from Disco) samples were cut from the Nd:LNOI wafer for characterization. For some measurements narrow ridge waveguides were prepared by dicing, too, into the LNOI top layer [16]. In order to couple light in and out, additional polishing cuts were made on both endfacets of the sample.
The Nd doping concentration in the LNOI layer was estimated using literature data. So far only a few values for the diffusion coefficients and the solubility of Nd in LN have been published. Buchal et al. studied z-cut substrates at 1300 K and 1345 K [17] and Hempstead published results for x-cut crystals annealed at three different temperatures [18]. These values are summarized in Table 1.
Table 1. Diffusion coefficients D from literature, coefficients D’ calculated assuming an activation energy of ${{E}_{A}} = 5{\; }$eV and resulting diffusion coefficient at 1393 K for this work [*].
Assuming that the temperature dependence of the diffusion coefficient is given by $D(T )= {D_0}\;\textrm{ exp(} - {E_A}\textrm{/}{k_B}\textrm{T)}$ with diffusion constant ${D_0}$, activation energy ${E_A}$ and Boltzmann constant ${k_B}$ and that the diffusion coefficient at the temperature T is known, the coefficient at $T^{\prime}$ is given by $D^{\prime}({T^{\prime}} )= D(T )\textrm{exp}({ - ({E_A}/{k_B}} )({T - T^{\prime}} )/({TT^{\prime}} ))$. On the basis of the diffusion coefficients D for the lowest temperatures determined by Buchal and Hempstead [17,18] the value for ${E_A}$ can be estimated by comparing calculated values $\; D^{\prime}$ with data for higher temperatures. From the results given in Table 1 it turned out that a value of ${E_A} = 5\; $eV fits best. Thus the diffusion coefficient of Nd in z-cut LN at 1393 K can be estimated to be ${D_z} = 3.9 \times {10^{ - 13}}$ cm2/s. According to this, we expect a diffusion depth of 10 µm and a surface concentration of ${C_{max}} = 1.55 \times {10^{24}}\;\textrm{ }{\textrm{m}^{ - 3}}$ (0.08 mol%). Due to the high value for the diffusion depth, the concentration of Nd can be expected to be constant in the first μm of the sample and thus in the whole LNOI top layer.
3. Experimental methods
For the measurement of emission spectra in Section 4.1, a high-index rutile prism was used to couple pump light from a continuous-wave Ti:Sapphire laser (wavelength 751.2 nm) into the planar Nd:LNOI film. The light reaching the endfacet of the substrate after about 10 mm of propagation was collected using a 40× microscope objective, coupled into a fiber and sent to an optical spectrum analyzer (AQ 6370D from Yokogawa).
As the coupling efficiency was not high enough with the prism coupling setup to measure the fluorescence lifetime in Section 4.2, a second setup using endfacet coupling depicted in Fig. 1 was employed. To increase both the transmitted power in the sample and the light collection by a photodetector a 12 µm-wide and 11 mm-long ridge structure was prepared in the top LNOI layer using diamond blade dicing. The pump light (wavelength 810 nm) from the Ti:Sapphire laser was coupled into this multimode ridge waveguide using a 40× microscope objective. For collecting the light at the endfacet a 50× objective was used. The in-coupled power and polarization were controlled using a half-wave plate and a polarizer. A chopper was used to modulate the input intensity and to generate 6 ms-long pulses. The pump light was filtered from the transmitted signal by a long-pass filter with a cut-on wavelength of 1 µm. The amplitude of the fluorescence power was measured using a fast photodiode (PDB450C from Thorlabs) and recorded by a LeCroy 7300A digital storage scope.
Fig. 1. Setup used for fluorescent lifetime measurement with λ/2: half-wave plate, pol.: polarizer, bs: beam splitter, 40x/50x: objective lenses, and CCD: CCD camera.
For the measurement of small-signal gain in Section 4.3 the chopper was removed and light from a DBR1083PN laser diode with a tunable output wavelength of (1083-1084) nm and a maximum output power of 100 mW was additionally coupled into the ridge. Transmitted pump and signal light were measured with the optical spectrum analyzer.
4. Spectroscopic results
4.1 Emission spectra and emission cross-sections
The pump wavelength 751.2 nm corresponds to the 4I9/2 → 2H9/2: 4F5/2 transition of Nd3+ ions. From the 2H9/2: 4F5/2 manifold the ions show a non-radiative decay to the 4F3/2 level. From this level radiative transitions to the 4I13/2, 4I11/2 and 4I9/2 levels take place. Figure 2 shows the observed infrared emission bands around 900, 1085 and 1374 nm, respectively. The Stark effect in the crystal field of LN causes terms with total angular momentum J to split into multiplets of (2J + 1)/2 doubly-degenerate lines. The emission spectra and position of lines are in good agreement with data for Nd-doped bulk crystals [19,20] considering a typical measurement uncertainty of ∼10%. No significant line broadening or shifting is observed, thus the thin film could be expected to be free of stress as well as defects.
Fig. 2. Emission spectra when pumping Nd3+ at λ = 751.2 nm in the Nd:LNOI layer. The red (solid) lines show the π-polarized, the blue (dashed) lines the σ-polarized spectra.
By adopting the Füchtbauer-Ladenburg equation
In principle, the denominator is the sum over all manifolds lying lower than the upper laser level (4I15/2, 4I13/2, 4I11/2 and 4I9/2). However, we neglected the 4I15/2 manifold here, because no signal was detected in the corresponding wavelength range and the branching ratio can be expected to be small [22,23].
The determined branching ratios are summarized in Table 2. Our values agree with results published by Lee et al. for diffusion doped MgO:LN [24] and with the results published by Sahar et al. for a congruent LN crystal doped with 8.46×10−19 cm-3 neodymium [25].
Table 2. Calculated branching ratios βj for Nd:LNOI for different transitions.
For calculating the emission cross-sections we used a value of the radiative lifetime of ${\tau _R}$ = 170 µs calculated by Loro et al. based on the Judd-Ofelt theory [22]. The peak emission cross-sections found are listed in Table 3 and the full spectra are presented in Fig. 3. For the transition 4F3/2 → 4I11/2 cross-sections for π polarization (at 1084 nm) of $\sigma _e^{\pi \; }$ = 15.2×10−20 cm2 and for σ polarization (at 1093 nm) of $\sigma _e^{\sigma \; }$ = 4.05×10−20 cm2 are obtained. These results are similar to the values for π polarization of $\sigma _e^{\pi \; }$(1084 nm) = 18×10−20 cm2 from Fan et al. [20] and $\sigma _e^{\pi \; }$(1084 nm) = 17.4×10−20 cm2 from Lee et al. [24]. For σ polarization they found $\sigma _e^{\sigma \; }$ (1093 nm) = 5.0×10−20 cm2 and $\sigma _e^{\sigma \; }$ (1093 nm) = 5.13×10−20 cm2, respectively. On the other hand, Sekita et al. published $\sigma _e^{\pi \; }$ (1084 nm) = 27×10−20 cm2 and $\sigma _e^{\sigma \; }$ (1093 nm) = 5.57×10−20 cm2 for Nd:MgO-doped near-stoichiometric LN [26]. These values are comparable to other Nd-doped host crystals, for example Nd:YAG with σe = (27 - 88)×10−20 cm2 [27].
Fig. 3. Calculated emission cross-section of Nd:LNOI using a radiative lifetime of
${\tau _R}$ = 170 µs from [22]. The red (solid) lines show the π-polarized, the blue (dashed) lines the σ-polarized spectra.
Table 3. Calculated peak emission cross-sections for Nd:LNOI.
4.2 Fluorescence lifetime
Figure 4 shows the data obtained and a fitted decay curve assuming a single-exponential decay $exp ({ - t/{\tau_F}} )+ {P_b}$ where ${P_b}$ is background light power. A fluorescence lifetime of ${\tau _F}$ = (120 ± 3) µs was found for the investigated Nd:LNOI sample. Sekita et al. observed an increase of the fluorescence lifetime with decreasing Nd concentration in near-stoichiometric Nd:LN crystals [26]. For a concentration of 0.02 wt.% (corresponding to 3.3×1024 m-3) they found ${\tau _F}$ = 120 µs in agreement with our findings. For diffusion-doped crystals Hempstead also observed concentration quenching measuring a maximum of ${\tau _F}$ = (111 ± 5) µs for a Nd3+ concentration of 1.83×1025 m-3 [20]. With ${\tau _F}$ = 120 µs and ${\tau _R}$ = 170 µs the fluorescence quantum efficiency from the 4F3/2 state is estimated to be η = ${\tau _F}/{\tau _R}$Z ≈ 0.7 showing that no additional relaxation mechanism is present or that the multi-phonon relaxation rate is not increased due to the implantation/ion-slicing during LNOI fabrication. Together with the reasonable narrow linewidth observed in Fig. 2 these results are promising towards future laser emission of Nd:LNOI using the 4F3/2 → 4I11/2 transition.
Fig. 4. Measured decay of fluorescence light power P with time t. The red (solid) line shows the fit of a single exponential function (with offset) with τ F = 120 µs.
4.3 Small-signal gain
In order to measure the small-signal gain, first the input power of the signal at 1083 nm was kept constant at low µW level for both polarizations, and the transmitted signal power was measured for increasing incident pump power. From the power ratio of transmitted signal light with/without pump light the gain can be calculated. Figure 5 shows the measured small-signal gain at 1083 nm as a function of the coupled pump at 815 nm. A maximum of 9.7 dB at 20.4 mW was found for σ polarization. For values of the coupled pump power higher than ∼25 mW optical damage could be observed in form of an unstable intensity distribution on the waveguides’ endfacet. For π polarization, a small signal gain of 14 dB was found for the 11 mm-long sample. Here a higher gain was achieved but the instabilities appeared already for pump powers higher than ∼15 mW.
Fig. 5. Small-signal gain at 1083 nm wavelength as a function of the incident pump power at λ = 815 nm for π (red solid line) and σ polarization (blue dashed line). The amplified spontaneous emission has been subtracted. The dashed lines are guides to the eye only.
In a follow-up experiment, the pump power at 815 nm was kept at 8 mW and the input power of the signal beam was varied. The result for π polarization is given in Fig. 6. As can be seen, the initial small signal gain of 9.5 dB (for input signal power of 6 μW) drops to values around 1 dB for coupled signal powers above 100 μW.
Fig. 6. Signal gain at 1084 nm as a function of the incident signal light power at constant pump power of 8 mW at λ = 815 nm for π polarization.
5. Conclusion
Nd:LNOI has been fabricated from a diffusion doped wafer for the first time, to the best of our knowledge. The optical quality has been proven by the spectroscopic analysis of a Nd:LNOI substrate being comparable to bulk samples. Emission cross-sections of 4.05×10−20 cm-2 for σ polarization (at 1093 nm) and 15.2×10−20 cm-2 for π polarization (at 1084 nm), respectively, have been obtained from the measurement of the fluorescence spectra. Additionally, a maximum small-signal gain of 14 dB for π-polarized light has been obtained in diced ridges pumped with 15 mW at 815 nm. Here optical damage occurred for pump power larger than 15 mW, while this threshold is slightly higher at 25 mW for σ polarization. Similar observations have been made for Er-LNOI ring lasers when pumped at 980 nm at comparable pump power levels [11,12]. To avoid such effects, it may be possible in the future to use, for example, MgO-doped LN substrates or the additional diffusion of Zn into LNOI.
To summarize, these results proof the method of diffusion doping prior to LNOI fabrication a suitable strategy for the fabrication of RE-doped LNOI for future use in active devices. As diffusion doping works also with structured layers, this approach provides the opportunity to realize active gain areas with locally varying doping concentrations and the use of different dopants on the same LNOI sample. These properties are crucial for applications like quantum memories as well as to achieve a high efficiency of devices such as ring lasers.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the corresponding author upon reasonable request.
References
1. D. Zhu, L. Shao, M. Yu, R. Cheng, B. Desiatov, C. J. Xin, Y. Hu, J. Holzgrafe, S. Ghosh, A. Shams-Ansari, E. Puma, N. Sinclair, C. Reimer, M. Zhang, and M. Lončar, “Integrated photonics on thin-film lithium niobate,” Adv. Opt. Photonics 13(2), 242–352 (2021). [CrossRef]
2. M. He, M. Xu, Y. Ren, J. Jian, Z. Ruan, Y. Xu, S. Gao, S. Sun, X. Wen, L. Zhou, L. Liu, C. Guo, H. Chen, S. Yu, L. Liu, and X. Cai, “High-performance hybrid silicon and lithium niobate Mach–Zehnder modulators for 100 Gbit s−1 and beyond,” Nat. Photonics 13(5), 359–364 (2019). [CrossRef]
3. C. Wang, M. J. Burek, Z. Lin, H. A. Atikian, V. Venkataraman, I.-C. Huang, P. Stark, and M. Lončar, “Integrated high quality factor lithium niobate microdisk resonators,” Opt. Express 22(25), 30924–30933 (2014). [CrossRef]
4. J. Lu, J. B. Surya, X. Liu, A. W. Bruch, Z. Gong, Y. Xu, and H. X. Tang, “Periodically poled thin-film lithium niobate microring resonators with a second-harmonic generation efficiency of 250,000%/W,” Optica 6(12), 1455–1460 (2019). [CrossRef]
5. M. Zhang, B. Buscaino, C. Wang, A. Shams-Ansari, C. Reimer, R. Zhu, J. M. Kahn, and M. Lončar, “Broadband electro-optic frequency comb generation in a lithium niobate microring resonator,” Nature 568(7752), 373–377 (2019). [CrossRef]
6. C. Becker, T. Oesselke, J. Pandavenes, R. Ricken, K. Rochhausen, G. Schreiber, W. Sohler, H. Suche, R. Wessel, S. Balsamo, I. Montrosset, and D. Sciancalepore, “Advanced Ti:Er:LiNbO3 waveguide lasers,” IEEE J. Sel. Top. Quantum Electron. 6(1), 101–113 (2000). [CrossRef]
7. D. Brüske, S. Suntsov, C. E. Rüter, and D. Kip, “Efficient ridge waveguide amplifiers and lasers in Er-doped lithium niobate by optical grade dicing and three-side Er and Ti in-diffusion,” Opt. Express 25(23), 29374–29379 (2017). [CrossRef]
8. D. Brüske, S. Suntsov, C. E. Rüter, and D. Kip, “Efficient Nd:Ti:LiNbO3 ridge waveguide lasers emitting around 1085 nm,” Opt. Express 27(6), 8884–8889 (2019). [CrossRef]
9. E. Saglamyurek, N. Sinclair, J. Jin, J. A. Slater, D. Oblak, F. Bussières, M. George, R. Ricken, W. Sohler, and W. Tittel, “Broadband waveguide quantum memory for entangled photons,” Nature 469(7331), 512–515 (2011). [CrossRef]
10. Q. Luo, Z. Hao, C. Yang, and R. Zhang, “Microdisk lasers on an erbium-doped lithium-niobate chip,” Sci. China: Phys., Mech. Astron. 64(3), 234263 (2021). [CrossRef]
11. Y. A. Liu, X. S. Yan, J. W. Wu, P. Zhu, Y. P. Chen, and X. F. Chen, “On-chip erbium-doped lithium niobate microcavity laser,” Sci. China: Phys., Mech. Astron. 64(3), 234262 (2021). [CrossRef]
12. Z. Wang, Z. Fang, Z. Liu, W. Chu, Y. Zhou, J. Zhang, R. Wu, M. Wang, T. Lu, and D. Cheng, “An on-chip tunable micro-disk laser fabricated on Er3+ doped lithium niobate on insulator (LNOI),” arXiv:2009.08953 [physics.optics] (2021).
13. S. Dutta, E. A. Goldschmidt, S. Barik, U. Saha, and E. Waks, “Integrated photonic platform for rare-earth ions in thin film lithium niobate,” Nano Lett. 20(1), 741–747 (2020). [CrossRef]
14. X. D. Jiang, D. Pak, A. Nandi, Y. Xuan, and M. Hosseini, “Rare earth-implanted lithium niobate: Properties and on-chip integration,” Appl. Phys. Lett. 115(7), 071104 (2019). [CrossRef]
15. D. Pak, H. An, A. Nandi, X. Jiang, Y. Xuan, and M. Hosseini, “Ytterbium-implanted photonic resonators on thin film lithium niobate,” J. Appl. Phys. 128(8), 084302 (2020). [CrossRef]
16. M. F. Volk, S. Suntsov, C. E. Rüter, and D. Kip, “Low-loss ridge waveguides in lithium niobate thin films by optical grade diamond blade dicing,” Opt. Express 24(2), 1386–1391 (2016). [CrossRef]
17. C. Buchal and S. Mohr, “Ion implantation, diffusion, and solubility of Nd and Er in LiNbO3,” J. Mater. Res. 6(1), 134–137 (1991). [CrossRef]
18. M. Hempsteadt, “Determination of diffusion profiles of neodymium in lithium niobate by means of spatially resolved fluorescence measurements,” J. Appl. Phys. 74(9), 5483–5492 (1993). [CrossRef]
19. V. T. Gabrielyan, A. A. Kaminskii, and L. Li, “Absorption and luminescence spectra and energy levels of Nd3+ and Er3+ ions in LiNbO3 crystals,” Phys. Status Solidi A 3(1), K37–K42 (1970). [CrossRef]
20. T. Y. Fan, A. Cordova-Plaza, M. J. F. Digonnet, R. L. Byer, and H. J. Shaw, “Nd:MgO:LiNbO3 spectroscopy and laser devices,” J. Opt. Soc. Am. B 3(1), 140–148 (1986). [CrossRef]
21. S. Singh, R. G. Smith, and L. G. Van Uitert, “Stimulated-emission cross section and fluorescent quantum efficiency of Nd3+ in yttrium aluminum garnet at room temperature,” Phys. Rev. B 10(6), 2566–2572 (1974). [CrossRef]
22. H. Loro, M. Voda, F. Jaque, J. García Solé, and J. E. Munoz Santiuste, “Polarized absorption of Nd3+ in LiNbO3: Effect of MgO codoping,” J. Appl. Phys. 77(11), 5929–5935 (1995). [CrossRef]
23. J. Wang, Y. Zhang, H. Xia, and J. Sheng, “Growth of Nd3+ doped LiNbO3 crystals using Bridgman method and its spectral properties,” Bull. Mater. Sci. 32(2), 183–186 (2009). [CrossRef]
24. C.-T. Lee and L.-G. Sheu, “Analysis of Nd:MgO:Ti:LiNbO3 waveguide lasers with nonuniform concentration distributions,” J. Lightwave Technol. 14(10), 2268–2276 (1996). [CrossRef]
25. M. R. Sahar, P. NgurahMadeD, and M. Radzi Sudin, “Optical transition of trivalent neodymium (Nd3+) in the anisotropic crystal lithium niobate,” Sol. State Sci. Technol. 14, 56–61 (2006).
26. M. Sekita, M. Nakamura, A. Watanabe, S. Takekawa, and K. Kitamura, “Induced emission cross sections of near-stoichiometric LiNbO3:Mg,Nd,” J. Appl. Phys. 100(10), 103501 (2006). [CrossRef]
27. M. Sekita, H. Haneda, T. Yanagitani, and S. Shirasaki, “Induced emission cross section of Nd:Y3Al5O12 ceramics,” J. Appl. Phys. 67(1), 453–458 (1990). [CrossRef]
