Open Access
Add to BibSonomy
Abstract
We have epitaxially grown high-quality single-crystal rare-earth oxide thin films, including Gd2O3 and erbium-incorporated (ErGd)2O3, on silicon-on-insulator substrate, and investigated their optical properties when embedded in horizontal slot waveguides. (ErGd)2O3 with an erbium concentration in the mid-1021 cm−3 range shows well-resolved Stark-split photoluminescence emission peaks in the telecommunications band and a photoluminescence lifetime-concentration product as large as 2.67×1018 s·cm−3 at room-temperature. Using these materials, horizontal slot waveguides with strong optical confinement in low-refractive-index rare-earth oxide layers, have been fabricated for silicon-based integrated active photonic devices. Thanks to the strong light-matter interaction, a large waveguide modal absorption of 88 dB/cm related to erbium ions is achieved, leading to a large potential optical gain. Intense emissions from the waveguides are also observed, with a radiation efficiency on the order of 10−4. These results indicate that a combination of epitaxial rare-earth oxide thin films and horizontal slot waveguides provides a promising platform for light amplification and generation on silicon.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Silicon (Si) photonics is considered to be one of the most promising baseline technologies for high-speed and low-power-consumption data communications [1–3]. As well, it has recently attracted attention in variety of emerging application fields such as light detection and ranging (LiDAR) for vehicle automation [4,5], artificial intelligence [6,7], and quantum computing and information processing [8–11]. This is because complex large-scale photonic integrated circuits consisting several fundamental optical components can be fabricated on Si chips by using low-cost mature complementary metal-oxide-semiconductor (CMOS) compatible processes. So far, although most passive optical components have been demonstrated to have high performance, the indirect bandgap property of Si has made it challenging to make fundamental active devices such as optical amplifiers and lasers. Most efforts have thus focused on hybrid integration combining other active materials on Si chips. III-V compounds are most investigated active materials for this purpose. With them, high-performance laser sources can be fabricated by wafer bonding or direct growth on Si substrate [12,13]. However, the fabrication complexity and process incompatibility with Si prevent them from being a means of large-scale and low-cost integration. Most recently, direct bandgap Ge/GeSn grown on Si has attracted attention because of its process compatibility with Si and potential for monolithic integration, but the demonstrated laser devices suffered from a high threshold and emission wavelengths shifted away from the telecommunications band, making their usefulness controversial [14,15].
Another promising platform for Si-based optical amplifiers and lasers consists of rare-earth (specifically, erbium)-doped solid-state materials, as the optical transition wavelengths of erbium (Er) ions are in the telecommunications band. In particular, Er-doped fiber amplifiers are widely used in long-haul optical communication systems [16], and recent Er-doped waveguide amplifiers (EDWAs) and lasers (EDWLs) have shown great potential for on-chip applications [17]. Typically, they have rather low optical gain per unit length (on the order of several dB/cm) due to their low Er concentration. To further miniaturize their device footprint for Si photonics, a much higher optical gain (e.g., on the order of several tens to hundreds of dB/cm), is desirable. For most amorphous and polycrystalline host materials, it is rather difficult to increase the optical gain by simply increasing Er concentration since detrimental effects such as Er clustering, concentration quenching, and defects, are unavoidable [17]. On the other hand, high-quality single-crystal host materials provide a best choice to suppress these effects. Recently, single-crystal Er chloride silicate nanowires with Er concentrations as high as 1.62$\times$10$^{22}$ cm$^{-3}$ and giant net material gains over 100 dB/cm have been experimentally demonstrated [18]. Complex Er silicate thin films have been also investigated by several groups, and net optical gains over 50 dB/cm have been achieved [19–21]. Some Er-doped functional oxides, such as yttria-stabilized zirconia, have also been investigated, and strong light emissions have been observed from them [22]. Most of these materials, however, are difficult to grow directly on Si, thus, their monolithic integration with Si photonic devices is not straightforward. What is needed, therefore, is to find high-quality Si-compatible materials.
Single-crystal rare-earth oxides (REOs) are one of the most promising candidates for such purposes. Very high Er concentrations up to 2.7$\times$10$^{22}$ cm$^{-3}$ could possibly be achieved using Er$_2$O$_3$ and the Er concentration can be widely tuned through formation of oxide compounds by incorporating other rare-earth elements. Most importantly, most REOs have a cubic crystalline phase and nearly lattice-matched with Si (with respect to twice of the lattice constant of Si(111)), thus, they can be easily grown on Si epitaxially [23]. In fact, REOs have been investigated as high-$\kappa$ gate dielectrics for metal-oxide-semiconductor field-effect transistors, indicating their excellent process compatibility with Si [24–28]. Er$_2$O$_3$ has been regarded as a promising gain material for a long time [29] and amorphous thin films on Si substrate [30] or crystal thin films on sapphire substrate [31] have been formed, both of which show potential for high-gain optical amplifiers. An optical gain of 5.9 dB/cm has been demonstrated in epitaxially grown Er:(Gd,Lu)$_2$O$_2$ on Y$_2$O$_3$ substrate [32]. However, very few studies have been tried to fabricate high-quality single-crystal thin films on Si substrate and investigate their optical properties [33]. Recently, we have shown that a variety of single-crystal REOs can be directly grown on Si substrate with high quality, including Er$_2$O$_3$ [34], Sc$_2$O$_3$ [35], and CeO$_2$ [36], as well as their compounds with varied compositions.
Furthermore, to make integrated active photonic devices, a waveguide structure is necessary. The host materials of most of the demonstrated EDWAs and EDWLs have a rather small refractive index contrast with the substrate; thus, thick films have usually been needed to confine the waveguide modes, and it is difficult to reduce the device footprint further [17,37]. This is not the case for REOs, since they are directly grown on Si and the film thickness is typically limited to $\sim$100 nm. A straightforward choice would be a Si waveguide capped with REO. However, because of its low refractive index compared with Si, most of the electromagnetic field would be confined in Si, rather than in the oxide, which would significantly reduce the effective optical gain. Moreover, REOs are generally difficult to pattern by normal dry etching. Therefore, a novel waveguide structure that has strong optical confinement in low-index REOs and that can be easily fabricated is necessary.
In this study, we have grown high-quality single-crystal REO (Gd$_2$O$_3$ and (ErGd)$_2$O$_3$) thin films on Si substrate, and fabricated horizontal slot waveguides with strong optical confinement in REO in a simple process without REO etching. Light propagation through the waveguides has been verified by using Gd$_2$O$_3$ based waveguides, and passive propagation losses for both TE and TM modes have been characterized. We have also investigated the optical absorption and light emission properties of Er$^{3+}$ ions in (ErGd)$_2$O$_3$ based waveguides. The results show the potential of our material and waveguide structure for Si-based optical amplifiers and lasers.
2. Material growth and characterization
REO thin films Gd$_2$O$_3$ and (ErGd)$_2$O$_3$ were epitaxially grown on silicon-on-insulator (111) substrate with a 100-nm-thick Si device layer and 2-$\mu$m-thick buried oxide layer by using molecular beam epitaxy. Gd$_2$O$_3$ has a cubic bixybite crystal structure with a lattice constant of $a$ = 10.81 Å, which has very small mismatch (0.46$\%$) with respect to twice that of Si (111) ($a$ = 5.43 Å) [28,38]. Moreover, Gd$_2$O$_3$ is also transparent in the telecommunications band since both its bulk and Gd$^{3+}$ ion absorption bands are in the ultraviolet wavelength range [39,40]. First, a 20-nm-thick Si buffer layer was grown at a substrate temperature of 700 $^{\circ }$C. After that, a Gd$_2$O$_3$ layer with a nominal thickness of 50 nm was grown at 650$\sim$750 $^{\circ }$C by using a metal Gd source under a flow of O$_2$. To achieve stoichiometric conditions, an optimized Gd flux rate of 0.06$\sim$0.08 Å/s and O$_2$ gas flow rate of 0.25 sccm were used. During growth, reflection high-energy electron diffraction (RHEED) was used to monitor the surface condition. Streaky RHEED patterns were maintained during growth, as shown in Fig. 1(a), indicating layer-by-layer growth of oxide on Si. Atomic force microscopy (AFM) measurements revealed that the grown Gd$_2$O$_3$ thin film had a very flat surface, with a root mean squared (RMS) roughness of $\sim$0.287 nm, as shown in Fig. 1(b). A typical cross-section transmission electron microscopy (TEM) image of the thin film is shown in Fig. 1(c), and an enlarged view of the Gd$_2$O$_3$/Si interface is shown in Fig. 1(d). Single-crystal growth could be clearly confirmed, and the interface was sharp without any amorphous interfacial layers. For Er-incorporated (Er$_x$Gd$_{1-x}$)$_2$O$_3$, additional Er flux at a rate of 0.01 Å/s was supplied simultaneously with Gd. The surface morphology and crystal quality were found to be similar to those of the Gd$_2$O$_3$ layer. Energy-dispersive X-ray spectroscopy indicated an Er composition $x$ of $\sim$12.9$\%$, which is consistent with the flux rate ratio of Er and Gd used during growth. This corresponds to an Er concentration of $\sim$3.3$\times$10$^{21}$ cm$^{-3}$.
Fig. 1. (a) RHEED pattern captured during growth of Gd$_2$O$_3$. (b) AFM image of Gd$_2$O$_3$ layer surface. (c) and (d) Cross-section TEM images of Gd$_2$O$_3$ layer grown on SOI substrate.
To characterize the optical properties of (ErGd)$_2$O$_3$ thin film, photoluminescence (PL) and time-resolved photoluminescence (TRPL) measurements were performed at room-temperature. Figure 2(a) shows a typical PL spectrum from thin film excited by a 1480 nm laser with a power density of $\sim$6.6 kW/cm$^2$. Intense light emissions with several narrow peaks can be observed. On the other hand, there was completely no light emission in this wavelength range from Gd$_2$O$_3$ thin film, indicating that these peaks are solely attributed to Er$^{3+}$ ions. Besides, the excitation wavelength used here is not coincident with optical transitions of Gd$^{3+}$ ions and the energy of excited Er$^{3+}$ ions is much lower that that of excited Gd$^{3+}$ ions [40], therefore energy transfer between Gd$^{3+}$ and Er$^{3+}$ ions can also be excluded. These peaks can be identified as transitions between the first excited state ($^4$I$_{13/2}$ manifold with crystal field induced Stark splitting levels Y$_i^{(')}$, $i=1\sim 7$) and the ground state ($^4$I$_{15/2}$ manifold with Stark levels Z$_j^{(')}$, $j=1\sim 8$) of Er$^{3+}$ ions at different crystalline symmetry sites (C$_2$ with non-inversion and C$_{3i}$ with inversion symmetry) [34,41]. For clarity, energy diagram of Er$^{3+}$ ions is shown in Fig. 2(b). The most intense peak at 1535.6 nm, corresponding to the Y$_1$ $\rightarrow$ Z$_1$ transition of Er$^{3+}$ ions at C$_2$ site, had a linewidth of $\sim$3.5 nm. These well-separated, narrow emission peaks indicate the high crystal quality of our materials.
Fig. 2. (a) Room-temperature PL spectrum of (ErGd)$_2$O$_3$ layer. Optical transition wavelengths between different Stark-split levels in Er-doped polycrystalline Y$_2$O$_3$ are shown as dashed lines. (b) Energy diagram of the $^4$I$_{13/2}$ and $^4$I$_{15/2}$ manifolds of Er$^{3+}$ ions at different crystalline symmetry sites. (c) Time-resolved PL decay curve of (ErGd)$_2$O$_3$ layer.
For the TRPL measurement, the excitation laser was modulated by a square wave with a frequency of 100 Hz and duty cycle of 50$\%$ by using an acousto-optic modulator, and the emitted light was coupled to a single-mode fiber and detected by a superconducting nanowire single-photon detector. The peak laser power density was about 490 W/cm$^2$. Figure 2(b) shows the TRPL decay curve of the (ErGd)$_2$O$_3$ layer at 1535.6 nm (Y$_1$ $\rightarrow$ Z$_1$). It is obvious that the curve doesn’t follow a single-exponential decay. The most probable mechanism for this is cooperative up-conversion (an Auger-like process), in which the energy of an excited Er$^{3+}$ ion at the $^4$I$_{13/2}$ manifold is transferred to another excited Er$^{3+}$ ion, which is then promoted to a higher state. Mathematically, this will induce a nonlinear term in the rate equation [35], in which the decay rate of the population density is proportional to its square, thus leading to discrepancy from single-exponential decay. We determined the effective lifetime to be 0.81 ms by fitting the PL decay curve with a stretched-exponential function [42]. The PL lifetime-concentration product of our material is about 2.67$\times$10$^{18}$ s$\cdot$cm$^{-3}$, which is one order of magnitude higher than that of Er$_2$O$_3$ grown by atomic layer epitaxy [33] and on the same order as other Er-incorporated materials shown to have high optical gain [43,44].
3. Waveguide structure and fabrication
Aiming for integrated active photonic devices, we propose here to use a horizontal slot waveguide as a basic waveguide structure (Fig. 3(a)). The epitaxial REO layer is sandwiched between two Si layers. With this configuration, the electromagnetic field energy for the TM mode can be confined in the low-refractive-index REO layer [45]. Furthermore, a ridge configuration with only the top Si layer etched is used to provide lateral optical confinement, which will significantly simplify the fabrication process by eliminating the necessity of etching REO. A similar waveguide structure has been exploited in several other material systems. However, most of them were operated using TE modes with rather small confinement factors [46–49]. A TM-polarized horizontal slot waveguide has been realized by using Er/Yb silicate thin film [50], but the confinement factor was also much lower than those of our devices. Figure 3(a) shows the simulated electromagnetic field intensity distribution of the fundamental TM mode of a horizontal slot waveguide with a 123-nm-thick bottom Si layer, 44-nm-thick Gd$_2$O$_3$ layer, 148-nm-thick top Si layer, and a width of 1 $\mu$m. We can clearly see that most of the electromagnetic energy is well confined in the thin Gd$_2$O$_3$ layer. The mode confinement factor [51] in the REO layer is calculated to be as large as 50.8$\%$ at a wavelength of 1535 nm even with such an ultra-thin film.
Fig. 3. (a) Distribution of electromagnetic field intensity of the fundamental TM mode of the horizontal slot waveguide. (b) Cross-section TEM image of Gd$_2$O$_3$-based horizontal slot waveguide. (c) and (d) are SEM images of fabricated grating coupler and microring resonators, respectively.
To fabricate this waveguide, after MBE growth of REO, another Si cap layer was deposited on top in the same MBE chamber. It is known that Si does not wet REO surfaces because the surface energy of REO is much lower than that of Si, and thus, islands with high surface roughness are usually formed during high-temperature growth [52]. To achieve a flat surface which is important for reducing the scattering loss of TM mode of the waveguide, the top Si layer was grown at a rather low temperature of $\sim$110$^\circ$C, at which 3D island growth didn’t occur. We therefore expect the interface between REO and Si cap layer keeping smooth after Si deposition. The deposited Si layer appeared to be amorphous, but had a surface roughness as low as $\sim$0.4 nm. After that, standard electron beam lithography and inductively coupled plasma etching were performed on these samples to pattern waveguides. Waveguides with different width and lengths, together with fiber-to-waveguide grating couplers, microring resonators, which will be mentioned below, were fabricated on the sample chip. Subsequently, post thermal annealing at 500 $^\circ$C in an O$_2$ atmosphere was performed, and it was found to significantly reduce the propagation loss. Figure 3(b) shows a typical cross-section TEM image of the fabricated waveguide.
4. Characterization of passive Gd$_2$O$_3$ waveguides
4.1 Fiber-to-waveguide grating couplers
For the waveguide characterization, a coupling structure has to be placed between waveguides and single-mode fibers. Here we used grating couplers (GCs) for their simple fabrication process. The GCs were also fabricated on the top Si layer, using the same procedure for the waveguides. Figure 3(c) shows a typical SEM image of a fabricated GC. The grating period and duty cycle were optimized in order to obtain high coupling efficiency around 1536 nm. A simulation indicated the designed horizontal slot waveguides could also support TE modes. For comparison, GCs for both TE and TM polarizations were designed and fabricated. Note that since GC is typically highly polarization-dependent, TE and TM-polarized GCs were designed with different grating periods and duty cycles. Figure 4(a) and 4(b) show the coupling loss spectra of GCs for TE and TM polarizations, respectively. The optical image acquired by an InGaAs camera for TM polarization is shown in the inset of Fig. 4(b). It clearly shows that light propagated through the waveguide. The results of a two-dimensional finite-difference time-domain (FDTD) simulation indicated that the peak wavelengths and spectrum shapes are consistent with the design. On the other hand, the peak coupling efficiencies for TE and TM polarizations were 16$\%$ and 7$\%$, respectively, which are lower than the designed values. As will be explained later, we attribute this discrepancy to the large waveguide propagation loss that was not subtracted from these data. The 1-dB bandwidths of the GCs for TE and TM polarizations were 25 and 57 nm, respectively. We can see that the GCs for TM polarization had a much broader bandwidth, which is advantageous for optical amplifiers since it becomes possible to couple both pump and signal lasers with separated wavelengths into the waveguide with a single GC.
Fig. 4. Measured and simulated coupling loss spectra of (a) TE-polarized GC, with grating period of 600 nm and duty cycle of 60$\%$ and (b) TM-polarized GC, with grating period of 900 nm and duty cycle of 85$\%$. The inset of (a) is an optical microscope image of the device used for coupling loss measurement, which consists of two identical GCs connected by two 100-$\mu$m-long tapers and a 100-$\mu$m-long narrow waveguide. The inset of (b) is an optical image of a TM-polarized waveguide under test acquired by an InGaAs camera.
4.2 Microring resonators
A microring resonator is a promising optical resonator structure for making lasers. It is also can be used to precisely characterize the waveguide propagation loss without using multiple devices of different lengths as are needed in the conventional cut-off method [53]. Figure 3(d) shows a typical SEM image of a fabricated microring resonator. The gap widths between the bus waveguide and microring were varied in order to determine under- or over-coupling between them, which is necessary for extracting the waveguide loss. Figure 5 shows TE- and TM-polarized transmission spectra of microring resonators, with a radius of 50 $\mu$m. Periodic resonance can be clearly seen in both cases, except that different optical extinctions appear because of the different waveguide-microring coupling strengths. We can determine the propagation loss of the ring waveguide by fitting the resonant dips to the theoretical transmission formula of a microring resonator [54]
where $a = \textrm {exp}(-\alpha L)$ is the single path amplitude transmission coefficient, $\alpha$ the power attenuation coefficient of the ring waveguide, $\phi = \beta L$ the single path phase shift, $\beta$ the propagation constant of the ring waveguide, and $r$ the self-coupling coefficient of the bus waveguide. Here, we focused on the resonance around 1535 nm. Zoomed views of the resonances and fitted curves are shown in the insets of Fig. 5. The propagation losses for the TE and TM polarizations were estimated to be 249 and 284 dB/cm, respectively, both values are rather large compared with those of state-or-the-art Si waveguides. Since the microring radius is rather large, we can ignore the radiative loss caused by bending, therefore the above values can also be regarded as the propagation losses of straight waveguides.Fig. 5. Transmission spectra of waveguide-coupled microring resonators for (a) TE and (b) TM polarizations. The widths of the bus waveguide and microring are both 0.8 $\mu$m and the ring radius is 50 $\mu$m. The resonances around 1535 nm, together with their fitted curves, are shown in the insets.
Although both modes had comparably large propagation losses, the losses are likely due to different mechanisms. First, the scattering losses caused by the roughness of the sidewalls and layer interfaces are not dominant. We have confirmed that normal Si waveguides (without REO and a-Si layers) fabricated using the same process show rather low loss on the order of several dB/cm. The sidewall scattering losses in horizontal slot waveguides are thus expected to be on the similar level, which are much smaller than the total measured losses above. Also, as mentioned before, the multilayers of Si/REO/Si had very smooth interfaces with an RMS roughness less than 0.4 nm; as such, they were also not likely able to induce such a large scattering loss for the TM polarization. This means that the large losses should mainly come from the optical absorption of the materials. Since TE and TM modes have different mode distributions, that is, the TE mode is mostly confined in Si and the TM mode in REO, the total propagation losses for TE and TM modes are most likely caused by the absorptions of top a-Si and REO layers, respectively. A-Si is known to have large absorption due to the existence of numerous surface dangling bonds. In fact, it has been reported that a-Si deposited by MBE under a similar condition as described here showed absorption coefficient larger than 100 cm$^{-1}$ [55,56]. On the other hand, there is also report that epitaxial oxide films show large optical absorptions due to defects such as oxygen vacancies [48]. Indeed, we have also found that post annealing at 500 $^{\circ }$C in an O$_2$ atmosphere significantly increases the transmission by over than 10 dB for both TE and TM polarizations. To reduce the propagation loss further, however, growth and annealing conditions must be optimized; this will be the subject of future investigations.
5. Optical properties of Er$^{3+}$ ions in (ErGd)$_2$O$_3$ waveguides
Horizontal slot waveguides based on (ErGd)$_2$O$_3$ thin films were fabricated using a similar procedure to the one described above. Figure 6(a) shows the transmission spectra of TE- and TM-polarized waveguides. For comparison, the PL spectrum from blank thin film is also shown (same with that in Fig. 2(a)). Compared with the transmission spectra of passive Gd$_2$O$_3$ waveguides, two pronounced dips around 1536 and 1547 nm appear in both spectra. The wavelengths are consistent with the emission peaks in the PL spectrum and, thus can be attributed to the Er$^{3+}$-related absorption. We examined the absorption dips around 1536 nm, which corresponds to the Y$_1$ $\rightarrow$ Z$_1$ optical transition of Er$^{3+}$ ions at C$_2$ site. The dip depths for TE and TM are about 1.2 and 2.9 dB, corresponding to waveguide modal absorptions of 36 and 88 dB/cm. The large difference comes from the different confinement factor of the TE and TM modes, as shown in their mode distributions (Fig. 6(b) and (c)). The TM mode has a confinement factor of 47.8$\%$, which is about 2.8 times that of the TE mode (16.8$\%$). From these results, we determined the material absorption coefficient of (ErGd)$_2$O$_3$ to be about 42 cm$^{-1}$ and the absorption cross-section of Er$^{3+}$ in (ErGd)$_2$O$_3$ to be 1.28$\times$10$^{-20}$ cm$^2$, which is reasonably consistent with the reported value of 1.54$\times$10$^{-20}$ cm$^2$ of Er$^{3+}$ in polycrystalline Y$_2$O$_3$ [41]. The large material and modal absorptions also indicates that a large potential gain for our material is available and that the TM-polarized horizontal slot waveguide is a promising structure for active waveguide devices.
Fig. 6. (a) TE- and TM-polarized transmission spectra of (ErGd)$_2$O$_3$-based horizontal slot waveguide with width of 0.8 $\mu$m and length of $\sim$330 $\mu$m, together with PL spectrum from (ErGd)$_2$O$_3$ blank thin film. (b) and (c) are electromagnetic field intensity distribution of fundamental TE and TM modes of the waveguide, respectively.
We also measured the light output from the waveguides at room-temperature. During the measurement, a tunable laser was launched from the input GC and the light from the waveguide was collected by a single-mode fiber through the output GC, and measured by an optical spectrum analyzer. Figure 7(a) shows the photoluminescence excitation (PLE) spectrum contour plot, together with PLE spectrum with an emission wavelength at 1536 nm (Y$_1$ $\rightarrow$ Z$_1$ transition) and PL spectrum with an excitation wavelength of 1517 nm (Y$_3$ $\rightarrow$ Z$_1$ transition) in Fig. 7(b) and 7(c), all for a TM-polarized waveguide. Similar to the PL spectrum from the blank thin film, well-resolved absorption/emission peaks, corresponding to Stark-split transitions, can be clearly seen in the PLE and PL spectra. The linewidths of the emission peaks are also similar to those from the blank film. These results indicate high crystal quality and a uniform Er concentration across a large area of our samples. Furthermore, despite the large fiber-to-waveguide coupling loss and waveguide propagation loss, light emissions with a peak power on the order of several tens of picowatt were obtained. The radiative efficiency for the emission coupled to the TM mode was roughly estimated to be on the order of 10$^{-4}$ by the following formula:
Fig. 7. (a) PLE contour plot of TM-polarized light emission output from (ErGd)$_2$O$_3$-based horizontal slot waveguide with width of 1.0 $\mu$m and length of $\sim$330 $\mu$m. (b) PLE spectrum with emission wavelength of 1536 nm. (c) PL spectrum with excitation wavelength of 1517 nm.
The combination of epitaxial single-crystal REO thin film and horizontal slot waveguide has shown very high optical absorption and light emission efficiency, suggesting that it is a very promising platform of light amplification and generation for Si photonic integrated circuits. Due to the rather large propagation loss induced by material absorption, a net optical gain couldn’t be achieved in the current stage of development. Through dedicated material and process optimization, for example, post thermal annealing to reduce the defect density in REOs [48], replacement of MBE-grown a-Si by low absorption hydrogenated a-Si [57] or single-crystal Si [49,52], we can expect a much lower propagation loss, and thus, envision a large optical gain. Moreover, the Er concentration of the thin films used in this study was not optimized. In our previous work, we found concentration quenching with increasing Er composition due to up-conversion [35]. This up-conversion not only reduces the population in the desired $^4$I$_{13/2}$ excited states; the produced visible luminescence may also be absorbed by the Si layers, leading to free carrier absorption in waveguides. To achieve high optical gain, therefore, Er concentration and waveguide structure should be systematically optimized.
6. Conclusions
A promising platform aiming for Si-based integrated active photonic devices, was realized by combining single-crystal REO thin films and horizontal slot waveguides. High-quality single-crystal REO thin films, including Gd$_2$O$_3$ and (ErGd)$_2$O$_3$, were epitaxially grown on silicon-on-insulator (111) substrate, and characterized as having flat surfaces, sharp Si/REO interfaces, and narrow and intense light emission peaks. TRPL measurements showed that a PL lifetime-concentration product of 2.67$\times$10$^{18}$ s$\cdot$cm$^{-3}$ could be achieved in (ErGd)$_2$O$_3$ with Er concentration of 3.3$\times$10$^{21}$ cm$^{-3}$, which is comparable to those of other Er-incorporated materials with large optical gain. To solve the problems of a low refractive index and process difficulties of REO thin films, ridge-type horizontal slot waveguides could be used as a basic waveguide structure for integrated photonic devices. By using such a structure, optical confinement factors as large as $\sim$50$\%$ could be implemented even with REO layer thicknesses far less than 100 nm. Due to the large residual absorption of REO and a-Si cap layers, the propagation losses of these waveguides were still large. Despite this, strong Er$^{3+}$ ion-related light absorption and highly efficient light emission were observed in the waveguides. A large optical gain can be expected if the waveguide loss can be reduced to the state-of-the-art levels of current Si waveguides. Our present results indicate that the materials and waveguide structure demonstrated here are promising for realizing high-performance CMOS-compatible optical amplifiers and lasers for Si photonic integrated circuits.
Funding
Japan Society for the Promotion of Science (KAKENHI 19H02207, KAKENHI 19H02636).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. H.-F. Liu, “Integrated silicon photonics links for high bandwidth data transportation,” in Optical Fiber Communication Conference, (Optical Society of America, 2014), p. Th1D.1.
2. T. Shi, T.-I. Su, N. Zhang, C.-Y. Hong, and D. Pan, “Silicon photonics platform for 400G data center applications,” in Optical Fiber Communication Conference, (Optical Society of America, 2018), p. M3F.4.
3. M. Glick, L. C. Kimmerling, and R. C. Pfahl, “A roadmap for integrated photonics,” Opt. Photonics News 29(3), 36–41 (2018). [CrossRef]
4. J. Sun, E. Timurdogan, A. Yaacobi, E. S. Hosseini, and M. R. Watts, “Large-scale nanophotonic phased array,” Nature 493(7431), 195–199 (2013). [CrossRef]
5. W. Xie, T. Komljenovic, J. Huang, M. Tran, M. Davenport, A. Torres, P. Pintus, and J. Bowers, “Heterogeneous silicon photonics sensing for autonomous cars,” Opt. Express 27(3), 3642–3663 (2019). [CrossRef]
6. Y. Shen, N. C. Harris, S. Skirlo, M. Prabhu, T. Baehr-Jones, M. Hochberg, X. Sun, S. Zhao, H. Larochelle, D. Englund, and M. Soljacic, “Deep learning with coherent nanophotonic circuits,” Nat. Photonics 11(7), 441–446 (2017). [CrossRef]
7. A. N. Tait, T. F. De Lima, E. Zhou, A. X. Wu, M. A. Nahmias, B. J. Shastri, and P. R. Prucnal, “Neuromorphic photonic networks using silicon photonic weight banks,” Sci. Rep. 7(1), 7430 (2017). [CrossRef]
8. N. C. Harris, G. R. Steinbrecher, M. Prabhu, Y. Lahini, J. Mower, D. Bunandar, C. Chen, F. N. Wong, T. Baehr-Jones, M. Hochberg, S. Lloyd, and D. Englund, “Quantum transport simulations in a programmable nanophotonic processor,” Nat. Photonics 11(7), 447–452 (2017). [CrossRef]
9. N. C. Harris, J. Carolan, D. Bunandar, M. Prabhu, M. Hochberg, T. Baehr-Jones, M. L. Fanto, A. M. Smith, C. C. Tison, P. M. Alsing, and D. Englund, “Linear programmable nanophotonic processors,” Optica 5(12), 1623–1631 (2018). [CrossRef]
10. X. Qiang, X. Zhou, J. Wang, C. M. Wilkes, T. Loke, S. O’Gara, L. Kling, G. D. Marshall, R. Santagati, T. C. Ralph, J. B. Wang, J. L. O’Brien, M. G. Thompson, and J. C. F. Matthews, “Large-scale silicon quantum photonics implementing arbitrary two-qubit processing,” Nat. Photonics 12(9), 534–539 (2018). [CrossRef]
11. J. Wang, S. Paesani, Y. Ding, R. Santagati, P. Skrzypczyk, A. Salavrakos, J. Tura, R. Augusiak, L. Mančinska, D. Bacco, D. Bonneau, J. W. Silverstone, Q. Gong, A. Acìn, K. Rottwitt, L. K. Oxenløwe, J. L. O’Brien, A. Laing, and M. G. Thompson, “Multidimensional quantum entanglement with large-scale integrated optics,” Science 360(6386), 285–291 (2018). [CrossRef]
12. T. Komljenovic, M. Davenport, J. Hulme, A. Y. Liu, C. T. Santis, A. Spott, S. Srinivasan, E. J. Stanton, C. Zhang, and J. E. Bowers, “Heterogeneous silicon photonic integrated circuits,” J. Lightwave Technol. 34(1), 20–35 (2016). [CrossRef]
13. A. Y. Liu and J. Bowers, “Photonic integration with epitaxial III–V on silicon,” IEEE J. Sel. Top. Quantum Electron. 24(6), 1–12 (2018). [CrossRef]
14. S. Wirths, R. Geiger, N. von den Driesch, G. Mussler, T. Stoica, S. Mantl, Z. Ikonic, M. Luysberg, S. Chiussi, J. M. Hartmann, H. Sigg, H. Faist, D. Buca, and D. Grutzmacher, “Lasing in direct-bandgap GeSn alloy grown on Si,” Nat. Photonics 9(2), 88–92 (2015). [CrossRef]
15. S. Bao, D. Kim, C. Onwukaeme, S. Gupta, K. Saraswat, K. H. Lee, Y. Kim, D. Min, Y. Jung, H. Qiu, H. Wang, E. A. Fitzgerald, C. S. Tan, and D. Nam, “Low-threshold optically pumped lasing in highly strained germanium nanowires,” Nat. Commun. 8(1), 1845 (2017). [CrossRef]
16. P. M. Becker, A. A. Olsson, and J. R. Simpson, Erbium-Doped Fiber Amplifiers: Fundamentals and Technology (Elsevier, 1999), chap. 1.
17. J. D. Bradley and M. Pollnau, “Erbium-doped integrated waveguide amplifiers and lasers,” Laser Photonics Rev. 5(3), 368–403 (2011). [CrossRef]
18. H. Sun, L. Yin, Z. Liu, Y. Zheng, F. Fan, S. Zhao, X. Feng, Y. Li, and C.-Z. Ning, “Giant optical gain in a single-crystal erbium chloride silicate nanowire,” Nat. Photonics 11(9), 589–593 (2017). [CrossRef]
19. H. Isshiki, F. Jing, T. Sato, T. Nakajima, and T. Kimura, “Rare earth silicates as gain media for silicon photonics,” Photonics Res. 2(3), A45–A55 (2014). [CrossRef]
20. X. Wang, P. Zhou, Y. He, and Z. Zhou, “Erbium silicate compound optical waveguide amplifier and laser,” Opt. Mater. Express 8(10), 2970–2990 (2018). [CrossRef]
21. A. Minowa, H. Isshiki, F. Kondow, K. Miyagi, and G. Nakamura, “Low threshold and extreme high optical gain in ErxYbyY2−x−ySiO5 compound crystalline thin films,” in International Conference on Solid State Devices and Materials, (2018), pp. 523–524.
22. A. Ruiz-Caridad, G. Marcaud, J. M. Ramirez, L. Largeau, T. Maroutian, S. Matzen, C. Alonso-Ramos, G. Agnus, E. Cassan, D. Marris-Morini, P. Lecoeur, and L. Vivien, “Towards optical amplification in complex functional oxides: exploring optical gain in erbium-doped yttria-stabilized zirconia waveguides,” Proc. SPIE 10921, 1092112 (2019). [CrossRef]
23. R. Dargis, A. Clark, F. Erdem Arkun, T. Grinys, R. Tomasiunas, A. O’Hara, and A. A. Demkov, “Monolithic integration of rare-earth oxides and semiconductors for on-silicon technology,” J. Vac. Sci. Technol., A 32(4), 041506 (2014). [CrossRef]
24. H. Osten, J. Liu, P. Gaworzewski, E. Bugiel, and P. Zaumseil, “High-k gate dielectrics with ultra-low leakage current based on praseodymium oxide,” in International Electron Devices Meeting, (IEEE, 2000), pp. 653–656.
25. H. Osten, A. Laha, M. Czernohorsky, E. Bugiel, R. Dargis, and A. Fissel, “Introducing crystalline rare-earth oxides into Si technologies,” Phys. Status Solidi A 205(4), 695–707 (2008). [CrossRef]
26. G. Niu, B. Vilquin, N. Baboux, C. Plossu, L. Becerra, G. Saint-Grions, and G. Hollinger, “Growth temperature dependence of epitaxial Gd2O3 films on Si(111),” Microelectron. Eng. 86(7-9), 1700–1702 (2009). [CrossRef]
27. R. Xu, Y. Zhu, S. Chen, F. Xue, Y. Fan, X. Yang, and Z. Jiang, “Epitaxial growth of Er2O3 films on Si(001),” J. Cryst. Growth 277(1-4), 496–501 (2005). [CrossRef]
28. R. Dargis, D. Williams, R. Smith, E. Arkun, R. Roucka, A. Clark, and M. Lebby, “Structural and thermal properties of single crystalline epitaxial Gd2O3 and Er2O3 grown on Si(111),” ECS J. Solid State Sci. Technol. 1(2), N24–N28 (2012). [CrossRef]
29. S. Saini, K. Chen, X. Duan, J. Michel, L. C. Kimerling, and M. Lipson, “Er2O3 for high-gain waveguide amplifiers,” J. Electron. Mater. 33(7), 809–814 (2004). [CrossRef]
30. A. M. Grishin, E. Vanin, O. Tarasenko, S. Khartsev, and P. Johansson, “Strong broad C-band room-temperature photoluminescence in amorphous Er2O3 film,” Appl. Phys. Lett. 89(2), 021114 (2006). [CrossRef]
31. A. Kuznetsov, S. Sadofev, P. Schäfer, S. Kalusniak, and F. Henneberger, “Single crystalline Er2O3: sapphire films as potentially high-gain amplifiers at telecommunication wavelength,” Appl. Phys. Lett. 105(19), 191111 (2014). [CrossRef]
32. A. Kahn, H. Kühn, S. Heinrich, K. Petermann, J. D. Bradley, K. Wörhoff, M. Pollnau, Y. Kuzminykh, and G. Huber, “Amplification in epitaxially grown Er:(Gd, Lu)2O3 waveguides for active integrated optical devices,” J. Opt. Soc. Am. B 25(11), 1850–1853 (2008). [CrossRef]
33. C. Michael, H. Yuen, V. Sabnis, T. Johnson, R. Sewell, R. Smith, A. Jamora, A. Clark, S. Semans, P. Atanackovic, and O. Painter, “Growth, processing, and optical properties of epitaxial Er2O3 on silicon,” Opt. Express 16(24), 19649–19666 (2008). [CrossRef]
34. T. Tawara, H. Omi, T. Hozumi, R. Kaji, S. Adachi, H. Gotoh, and T. Sogawa, “Population dynamics in epitaxial Er2O3 thin films grown on Si(111),” Appl. Phys. Lett. 102(24), 241918 (2013). [CrossRef]
35. T. Tawara, Y. Kawakami, H. Omi, R. Kaji, S. Adachi, and H. Gotoh, “Mechanism of concentration quenching in epitaxial (ErxSc1−x)2O3 thin layers,” Opt. Mater. Express 7(3), 1097–1104 (2017). [CrossRef]
36. T. Inaba, T. Tawara, H. Omi, H. Yamamoto, and H. Gotoh, “Epitaxial growth and optical properties of Er-doped CeO2 on Si (111),” Opt. Mater. Express 8(9), 2843–2849 (2018). [CrossRef]
37. M. Pollnau and J. D. Bradley, “Optically pumped rare-earth-doped Al2O3 distributed-feedback lasers on silicon,” Opt. Express 26(18), 24164–24189 (2018). [CrossRef]
38. N. Hirosaki, S. Ogata, and C. Kocer, “Ab initio calculation of the crystal structure of the lanthanide Ln2O3 sesquioxides,” J. Alloys Compd. 351(1-2), 31–34 (2003). [CrossRef]
39. Y. A. Kuznetsova and A. F. Zatsepin, “Optical properties and energy parameters of Gd2O3 and Gd2O3: Er nanoparticles,” J. Phys.: Conf. Ser. 917, 062001 (2017). [CrossRef]
40. W. T. Carnall, P. R. Fields, and K. Rajnak, “Electronic energy levels of the trivalent lanthanide aquo ions. ii. Gd3+,” J. Chem. Phys. 49(10), 4443–4446 (1968). [CrossRef]
41. J. B. Gruber, K. L. Nash, D. K. Sardar, U. V. Valiev, N. Ter-Gabrielyan, and L. D. Merkle, “Modeling optical transitions of Er3+(4f11) in C2 and C3i sites in polycrystalline Y2O3,” J. Appl. Phys. 104(2), 023101 (2008). [CrossRef]
42. R. Chen, “Apparent stretched-exponential luminescence decay in crystalline solids,” J. Lumin. 102-103, 510–518 (2003). [CrossRef]
43. L. Yin, H. Ning, S. Turkdogan, Z. Liu, P. L. Nichols, and C.-Z. Ning, “Long lifetime, high density single-crystal erbium compound nanowires as a high optical gain material,” Appl. Phys. Lett. 100(24), 241905 (2012). [CrossRef]
44. J. Rönn, W. Zhang, A. Autere, X. Leroux, L. Pakarinen, C. Alonso-Ramos, A. Säynätjoki, H. Lipsanen, L. Vivien, E. Cassan, and Z. Sun, “Ultra-high on-chip optical gain in erbium-based hybrid slot waveguides,” Nat. Commun. 10(1), 432 (2019). [CrossRef]
45. V. R. Almeida, Q. Xu, C. A. Barrios, and M. Lipson, “Guiding and confining light in void nanostructure,” Opt. Lett. 29(11), 1209–1211 (2004). [CrossRef]
46. C. Xiong, W. H. Pernice, M. Li, and H. X. Tang, “High performance nanophotonic circuits based on partially buried horizontal slot waveguides,” Opt. Express 18(20), 20690–20698 (2010). [CrossRef]
47. C. Xiong, W. H. Pernice, J. H. Ngai, J. W. Reiner, D. Kumah, F. J. Walker, C. H. Ahn, and H. X. Tang, “Active silicon integrated nanophotonics: ferroelectric BaTiO3 devices,” Nano Lett. 14(3), 1419–1425 (2014). [CrossRef]
48. F. Eltes, D. Caimi, F. Fallegger, M. Sousa, E. O’Connor, M. D. Rossell, B. Offrein, J. Fompeyrine, and S. Abel, “Low-loss BaTiO3–Si waveguides for nonlinear integrated photonics,” ACS Photonics 3(9), 1698–1703 (2016). [CrossRef]
49. S. Abel, F. Eltes, J. E. Ortmann, A. Messner, P. Castera, T. Wagner, D. Urbonas, A. Rosa, A. M. Gutierrez, D. Tulli, P. Ma, B. Baeuerle, A. Josten, W. Heni, D. Caimi, L. Czornomaz, A. A. Demkov, J. Leuthold, S. Pablo, and J. Fompeyrine, “Large Pockels effect in micro-and nanostructured barium titanate integrated on silicon,” Nat. Mater. 18(1), 42–47 (2019). [CrossRef]
50. R. Guo, B. Wang, X. Wang, L. Wang, L. Jiang, and Z. Zhou, “Optical amplification in Er/Yb silicate slot waveguide,” Opt. Lett. 37(9), 1427–1429 (2012). [CrossRef]
51. J. T. Robinson, K. Preston, O. Painter, and M. Lipson, “First-principle derivation of gain in high-index-contrast waveguides,” Opt. Express 16(21), 16659–16669 (2008). [CrossRef]
52. A. Fissel, R. Dargis, E. Bugiel, D. Schwendt, T. Wietler, J. Krügener, A. Laha, and H. Osten, “Single-crystalline Si grown on single-crystalline Gd2O3 by modified solid-phase epitaxy,” Thin Solid Films 518(9), 2546–2550 (2010). [CrossRef]
53. S. Xiao, M. H. Khan, H. Shen, and M. Qi, “Modeling and measurement of losses in silicon-on-insulator resonators and bends,” Opt. Express 15(17), 10553–10561 (2007). [CrossRef]
54. W. Bogaerts, P. De Heyn, T. Van Vaerenbergh, K. De Vos, S. Kumar Selvaraja, T. Claes, P. Dumon, P. Bienstman, D. Van Thourhout, and R. Baets, “Silicon microring resonators,” Laser Photonics Rev. 6(1), 47–73 (2012). [CrossRef]
55. B. Fogal, S. O’Leary, D. Lockwood, J.-M. Baribeau, M. Noël, and J. Zwinkels, “Disorder and the optical properties of amorphous silicon grown by molecular beam epitaxy,” Solid State Commun. 120(11), 429–434 (2001). [CrossRef]
56. S. O’Leary, B. Fogal, D. Lockwood, J.-M. Baribeau, M. Noël, and J. Zwinkels, “Optical dispersion relationships in amorphous silicon grown by molecular beam epitaxy,” J. Non-Cryst. Solids 290(1), 57–63 (2001). [CrossRef]
57. S. Zhu, G. Lo, and D. Kwong, “Low-loss amorphous silicon wire waveguide for integrated photonics: effect of fabrication process and the thermal stability,” Opt. Express 18(24), 25283–25291 (2010). [CrossRef]
