Graphene microcavity frequency combs and related methods of manufacturing

Abstract

Based on graphene heterostructure in chip-scale silicon nitride microresonators, optoelectronic control and modulation in frequency combs via group velocity dispersion modulation can be demonstrated. By tuning graphene Fermi level from 0.50 eV to 0.65 eV via electric-field gating, deterministic in-cavity group velocity dispersion control from anomalous (−62 fs2/mm) to normal (+9 fs2/mm) can be achieved with Q factor remaining high at 106. Consequently, both the primary comb lines and the full comb spectra can be controllable dynamically with the on/off switching of the Cherenkov radiation, the tuning of the primary comb lines from 2.3 THz to 7.2 THz, and the comb span control from zero comb lines to ˜781 phase-locked comb lines, directly via the DC voltage.

Description

FIELD OF THE INVENTION

The present invention generally relates to frequency combs and, more specifically, graphene microcavity frequency combs.

BACKGROUND

An optical frequency comb can be referred to as an optical spectrum that is made up of a series of discrete, equally spaced frequency lines. The frequency comb may have equidistant optical frequency components, while the intensity of the comb lines can vary substantially. The frequencies of these lines can be known to a high degree of accuracy, which allows the frequency comb to be used as an optical ruler. Optical frequency combs can provide equidistant frequency markers in various bands, allowing for substantial breakthroughs across a variety of different domains, such as but not limited to precision metrology and spectroscopy, astronomical observations, ultrafast optics, and quantum information. For example, if the comb frequencies are known, the frequency comb can be used to measure unknown frequencies by measuring beat notes, which reveal the difference in frequency between the unknown frequency and the comb frequencies.

Early attempts to produce broadband frequency combs were based on strongly driven electro-optic modulators, which can impose dozens of sidebands on a single-frequency input beam from a single-frequency continuous-wave laser. It was then found that this process could be made more efficient (for obtaining more comb lines) by placing the modulator in a resonant cavity, particularly when the intracavity dispersion was minimized. Further improvements were based on parametric amplification.

SUMMARY OF THE INVENTION

One embodiment includes a microresonator including a silicon nitride ring cladded with silicon dioxide, wherein the silicon oxide defines a window above a region of the silicon nitride ring, a bus waveguide coupled to the silicon nitride ring, a graphene ion-gel heterostructure disposed within the window above the region of the silicon nitride ring, wherein the graphene/ion-gel heterostructure includes a graphene monolayer, source-drain electrodes integrated on the graphene monolayer, and a layer of ionic liquid used as a gate dielectric, wherein the microresonator is configured to form a frequency comb tunable through the modulation of the dispersion of the microresonator.

In another embodiment, the modulation of the intracavity dispersion is achieved through the tuning of the Fermi level of the graphene monolayer using an applied gate voltage.

In a further embodiment, the Fermi level of the graphene monolayer can be tuned from about 0.45 eV to about 0.65 eV.

In still another embodiment, the dispersion of the microresonator can be tuned from about −62 fs²/mm anomalous dispersion to about +9 fs²/mm normal dispersion through tuning the gate voltage from about −2 V to about 0 V.

In a still further embodiment, the microresonator is configured to form a frequency comb having a relative spectral location that is tunable from about 2.3 THz to about 7.2 THz.

In yet another embodiment, the microresonator is configured to form a frequency comb having a relative spectral location that is tunable from about 2.3 THz to about 7.2 THz by tuning the applied gate voltage from about −1 V to about −1.5 V.

In a yet further embodiment, a fixed source-drain voltage of 10 mV is applied to the graphene ion-gel heterostructure.

In another additional embodiment, the microresonator is capable of achieving multi-soliton states.

In a further additional embodiment, the microresonator is capable of achieving soliton states with soliton numbers of 12, 11, 9, 8, 6, 5, and 4.

In another embodiment again, the microresonator has a loaded quality factor of about 1.6×10⁶.

In a further embodiment again, the source-drain electrodes comprise Ti/Au.

In still yet another embodiment, the ionic liquid comprises diethylmethyl(2-methoxyethyl)ammonium bis(trifluoromethylsulfonyl)imide.

In a still yet further embodiment, the layer of ionic liquid provides a capacitance of about 7.2 μF/cm².

In still another additional embodiment, the graphene monolayer is disposed about 100 nm away from the silicon nitride ring.

In a still further additional embodiment, the graphene monolayer is disposed about 300 nm away from the silicon nitride ring.

In still another embodiment again, the graphene monolayer and the silicon nitride ring have a planar interaction arc length of about 80 μm.

In a still further embodiment again, the silicon nitride ring has a cross section of about 1200 nm×800 nm.

In yet another additional embodiment, the silicon nitride ring has a diameter of about 350 μm.

In a yet further additional embodiment, the bus waveguide has a cross section of about 1000 nm×800 nm.

In yet another embodiment again, the bus waveguide is disposed about 600 nm away from the silicon nitride ring.

Additional embodiments and features are set forth in part in the description that follows, and in part will become apparent to those skilled in the art upon examination of the specification or may be learned by the practice of the invention. A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and the drawings, which forms a part of this disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

The description will be more fully understood with reference to the following figures and data graphs, which are presented as exemplary embodiments of the invention and should not be construed as a complete recitation of the scope of the invention.

FIG. 1A conceptually illustrates a schematic architecture of a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 1B shows a cross section electrical field distribution of a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 1C shows a micrograph of a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 1D shows calculated group-velocity-dispersion and third-order-dispersion of graphene in accordance with an embodiment of the invention.

FIG. 1E shows simulated Kerr comb dynamics in a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 2A shows an electronic measurement of a graphene/ion-gel capacitor in accordance with an embodiment of the invention.

FIG. 2B shows theoretical predictions of the changes in the effective index of a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 2C shows measured transmissions and mode free-spectral range of a graphene microring resonator in accordance with an embodiment of the invention.

FIG. 2D shows tuned Q factor and dispersion a graphene microring resonator in accordance with an embodiment of the invention.

FIGS. 3A-3F show various observations of gate-tunable graphene Kerr frequency combs in accordance with many embodiments of the invention.

FIGS. 4A-4D conceptually illustrate four specific examples of the soliton crystal states under optimized gate voltages in accordance with various embodiments of the invention.

DETAILED DESCRIPTION

Turning now to the drawings, gate-tunable frequency combs and related methods of manufacturing in accordance with embodiments of the invention are illustrated. Optical frequency combs are cornerstones in frequency metrology, precision spectroscopy, astronomical observations, ultrafast optics, and quantum information. Based on the Kerr and Raman nonlinearities in monolithic ultrahigh-Q microresonators, chip-scale frequency combs have recently been examined, advancing studies in optical clockwork and observations of temporal cavity solitons. However, the intracavity dispersion, which can determine the comb formation, is hardly amenable to electric-field tenability-whether in microcavities or fiber cavities. Arising from its exceptional Fermi-Dirac tunability and ultrafast carrier mobility, graphene has a complex dispersion (n_g=n_g,r+in_g,i), which can be determined by its gate-tunable optical conductivity. This characteristic has brought about a variety of optoelectronic advances, ranging from modulators, photodetectors to controllable plasmonics. Many embodiments of the invention are directed at combining the cross-disciplinary field of Dirac fermions in two-dimensional graphene to show the gated intracavity-tunability of graphene based optical frequency combs. The 2D surface material is also logically extensible to other 2D materials other than graphene, such as but not limited to other thin-film semiconductors, semimetals, transition metal dichalcogenides. With the unique linear and massless band structure of graphene, the gate-tunable optical conductivity can be coupled onto a photonic microresonator, modulating its second- and higher-order dispersions through the Fermi level. In many embodiments, a dual-layer ion-gel-gated transistor can be implemented to tune the electron-hole carrier population for the dynamical and reversible frequency comb control while also preserving cavity quality factors up to 10⁶ in the graphene-based comb. The Fermi level of the graphene can be tuned up to 0.65 eV under single-volt-level control. The formation of charge-tunable primary comb lines, coherent Kerr frequency combs, controllable Cherenkov radiation, and controllable multi-soliton states can all be achieved in a single microcavity. The voltage-tunable transitions from periodic soliton crystals to crystals with defects can also be demonstrated, mapped by an ultrafast second harmonic optical autocorrelation. This heterogeneous graphene-microcavity can provide a new fundamental platform for the understanding of dynamical frequency combs and ultrafast optics at the interface of single atomic layer nanoscience and ultrafast optoelectronics.

Temporal cavity solitons in active nonlinear microresonators have recently been observed. These femtosecond soliton pulses can span an octave using well-established nonlinear spectral broadening techniques, providing a concrete path towards self-referenced optical frequency microcombs. Different from the canonical soliton in a conservative system, these temporal cavity solitons can be formed in dissipative nonlinear resonators, and their existence can rely on the composite balance between parametric gain and cavity loss as well as between Kerr nonlinearity and cavity dispersion. Dispersion, which determines the temporal redistribution of pulse energy, can be important in defining the pulse shape, stability zone and the excitation pathway of the temporal cavity solitons. Nevertheless, while the parametric gain (via Kerr nonlinearity) can be controlled with the change of pump power, the cavity dispersion is generally predefined only by the waveguide design in the current state of the art. Furthermore, since light is tightly confined in the wavelength-scale waveguide, its dispersion can inevitable be prone to intrinsic nanometer-scale fabrication errors, which can result in device-to-device deviations from the designed dispersion values.

Gate-Tunable Frequency Combs

The gated intracavity-tunability of graphene-based optical frequency combs can be achieved in many different ways in accordance with various embodiments of the invention. In many embodiments, the gate-tunable optical conductivity of graphene can be coupled onto a photonic resonator, allowing for the modulation of its second- and higher-order dispersions through the Fermi level. In some embodiments, the Fermi level of graphene can be tuned up to 0.65 eV while incorporated on an optical ring cavity by using a graphene-based dual-layer transistor with ionic liquid as the gate electric. In some embodiments, a microresonator can be enabled with widely gate-tunable group velocity dispersion (“GVD”), from anomalous to normal region (e.g., −62 fs²/mm to 9 fs²/mm), and a notable Q factor (up to 10⁶), which can be orders-of-magnitude higher than prior graphene-based studies. Consequently, the generation and dynamic operation of Kerr combs in a graphene chip can be achieved. In various embodiments, the formation of controllable primary comb lines from around 2.3 THz up to around 7.2 THz at −1.5 V gate voltages, the switchable formation of coherent Kerr frequency combs, and Cherenkov dispersive radiation with gate-controlled phase matching can all be achieved in a single microresonator. These can be measured with correlated optical spectra, RF spectra, and ultrafast second-harmonic optical autocorrelation. At a Fermi level of 0.59 eV, a phase-locked multi-soliton state with eight solitons per cavity round-trip can be achieved, deterministically formed and controlled by the gate voltage.

FIG. 1A conceptually illustrates the concept and fabrication of a graphene gate-tunable Kerr frequency comb with source-drain and top gating in accordance with an embodiment of the invention. As shown in the illustrative embodiment, a graphene/ion-gel heterostructure 100 can be incorporated in a nitride microresonator 102. In many embodiments, a high-Q silicon nitride microresonator is nanofabricated in a silicon foundry in a 350 μm diameter ring structure with measured loaded Q˜1.6×10⁶ (intrinsic Q of ˜2×10⁶) and free-spectral range (“FSR”) of ˜90 GHz. In some embodiments, the nitride core 102 can be fabricated to have a 1200×800 nm² cross-section, a 600 nm gap to the input-output coupling waveguide of 1000×800 nm² cross-section, and a top oxide cladding. In various embodiments, single atomic layer graphene 104 can be grown via chemical vapor deposition and transferred onto an exposed region of the nitride ring 102 (with etched SiO₂ window). A cross section electrical field distribution of a device in accordance with an embodiment of the invention is shown in FIG. 1B. In the illustrative embodiment, the device is fabricated to have a 1.2×0.8 μm²Si₃N₄ core. In some embodiments, the device can be fabricated such that graphene and the top gate probe are separated by ˜1 μm with the interlayer ion-gel capacitor. The distance between the monolayer graphene and nitride core can be optimized to ˜400 nm with a controlled top oxide cladding etch, enabling 8% of the TE₀₁ mode intensity evanescently overlapping the graphene electron-hole carriers and dispersion.

To ensure transparency and minimal effect on the resonator Q for coherent comb generation, the interacting graphene can be top-gated to pull the Fermi level up to 0.6 eV for reduced photon absorption in the nearly-massless Dirac cone. An ion-gel capacitor can be implemented on top of the graphene monolayer where the ionic liquid electric double layer provides a capacitance as high as ˜7.2 μF/cm². This value can enable high doping control and comb tunability with a few-volt level gating. This can be important for sharp modulation on cavity and comb dispersion while keeping the cavity loss low. In several embodiments, the graphene-nitride gap can be optimized to 300 nm. In various embodiments, the planar graphene-nitride interaction arc length can be optimized to ˜80 μm for a large frequency comb tunability with minimal graphene absorption losses. Further details on the nanofabrication and baseline characterization are described below in further detail.

FIG. 1C shows an optical micrograph depicting a bus waveguide 106, ring resonator 108, and Au/Ti metallized patterns 110 in accordance with an embodiment of the invention. An etched window 112 can be designed to ensure both the graphene-light interaction and a reduced propagation loss. FIG. 1D shows plots of the computed group velocity dispersion (β₂) and the computed third order dispersion (β₃) for tuned Fermi levels from 0.2 eV to 0.8 eV of a graphene monolayer in accordance with an embodiment of the invention. For each Fermi level, the wavelength-oscillations are noted in both β₂ and β₃, which can arise from the graphene carrier relaxation oscillation lifetime captured in the resonance of the monolayer sheet conductivity. As a result, the graphene β₂ can be tuned from anomalous to normal dispersion and then back to anomalous via gate voltage, which can be important for nonlinear phase-matching tunability. This can enable a wide and tunable frequency comb generation in a graphene microring resonator (“GMR”). Based on the modeled overall graphene β₂ and β₃, a heterogeneous microresonator can be modeled for Kerr frequency comb generation. FIG. 1E shows simulated Kerr comb dynamics as a temporal map in a GMR via the Lugiato-Lefever equation modeling, which can be determined by the Fermi level of the graphene, in accordance with an embodiment of the invention. As shown in the exemplary embodiment, at E_F of 0.2 eV, the Q factor is low and hence there is no comb generation. At E_F of 0.5 eV, the GMR has a Q˜8×10⁵, β₂˜−50 fs²/mm and β₃˜0, resulting in a slow comb generation. At E_F of 0.8 eV, a fast full comb generation is observed numerically under a Q higher than 1×10⁶, β₂˜−30 fs²/mm and β₃˜−400 fs³/mm.

Although the discussion above with respect to FIGS. 1A-1E describes graphene gate-tunable frequency combs with specific dimensions and observations, a person having ordinary skill in the art would appreciate that a graphene gate-tunable frequency comb can be fabricated with varying degrees in dimensions and characteristics.

Observations and Measurements of GMRs

FIG. 2A shows the electrical tuning performance of graphene in a GMR in accordance with an embodiment of the invention. With a fixed source-drain voltage V_SD of 10 mV, the source-drain current I_SD can be tuned with gate voltage V_G. In the exemplary embodiment, when V_G reaches 2.4 V, I_SD has a minimum of 6.5 μA. Here the carrier density of the graphene monolayer reaches the Dirac point. When V_G is less than 2.4 V, graphene can be p-doped. In a cyclic V_G tuning, a clear hysteresis loop can be observed due to the electronic trapping. Correspondingly, the gate-tunable Fermi energy |E_F|, |E_F|=n|v_F|(πN)^−1/2 is plotted in the bottom panel of FIG. 2A. In the exemplary embodiment, |E_F| is shown to be ˜(V_G)^1/2. V_G can be tuned in range of −2 V to 0 V, thereby controlling the graphene |E_F between 0.65 eV to 0.45 eV. For V_G=0 V, the graphene monolayer in the GMR can already be heavily doped, which can be significant for the dispersion tuning with a low loss.

FIG. 2B maps the calculated real and imaginary parts of a GMR varying with |E_F| and wavelength λ in accordance with an embodiment of the invention. For the GMR, n_eff can be determined by graphene permittivity ε_G={−−lm(σ_G)+iRe(σ_G)}/{2πfΔ}, where σ_G is the graphene conductivity, f is the optical frequency, and Δ=0.4 nm is the thickness of the monolayer graphene. ∂ⁿRe(n_eff)/∂λⁿ refers to the nth order dispersion, while Im(n_eff) refers the waveguide loss. In the two maps, the curves in the top panel denote the boundary where dispersion dramatically changes, and the curve in the bottom panel denotes the low loss region. In the exemplary embodiment, the measurements are taken by applying a high-power continuous-wave (“CW”) pump at 1600 nm. At this wavelength, when the |E_F| is tuned from 0.45 eV to 0.65 eV, n_eff can be controlled from 1.789+0.058i to 1.781+0.001i.

FIG. 2C shows the measured transmissions (top panel) and wavelength-FSR dependences (bottom panel) of a GMR under different gate voltages in accordance with an embodiment of the invention. In the exemplary embodiment, a broadband tunable laser serves as the light source with less than 10 mW, which is below the comb generation threshold. For a selected resonance around 1600 nm, when V_G is tuned from 0 V to −2 V, the extinction ratio can increase from 63% to 84%, with resonance linewidth decreasing from 3.1 μm to 1.6 μm. Across the whole spectrum, the mode non-equidistance D_FSR refers to GVD as D_FSR=−β₂c(2πf_FSR)²/n_eff, where c is the light velocity in a vacuum. As shown, it is 320 kHz/mode under VG −1 V (anomalous dispersion) and −45 kHz/mode under V_G−1.8 V (normal dispersion).

FIG. 2D shows the gate tuning performance of a GMR in accordance with an embodiment of the invention. Under gate voltages from 0 V to −2 V, Q factor of the GMR can increase from ˜5×10⁵ to 10⁶, which can enable comb generation under a 1 W pump. Simultaneously, the dispersion of the resonator can be dynamically tuned, varying continuously from −62 fs²/mm anomalous dispersion to +9 fs²/mm normal dispersion.

Although FIGS. 2A-2D show measurements of specific GMRs, a person having ordinary skill in the art would appreciate that these measurements are taken from specific embodiments of the invention and are not representative of every embodiment of the invention.

FIG. 3A shows primary comb lines at controlled gate voltages and Fermi levels of graphene in accordance with an embodiment of the invention. In the exemplary embodiment, for applied V_G=−1 V, −1.2 V, and −1.5 V, the measured frequency offset between the primary comb line and the pump Δf_pri, proportional to (1/β₂)^1/2 is observed at 2.36 THz, 3.25 THz, and 7.17 THz respectively. When V_G=−1.8 V, GVD of the GMR becomes positive and can therefore be harder for phase-matching without local mode-crossing-induced dispersion. FIG. 3B shows the optical spectra under carefully controlled laser-cavity detuning in accordance with an embodiment of the invention. Full frequency combs generated under gate voltages of −1 V, −1.2 V, −1.5 V, and −1.8 V are shown. In the exemplary embodiment, the launched pump power is fixed at 34.5 dBm, and Kerr combs are generated via fine-adjustment of the pump wavelength. In each controlled laser-cavity detuning conditions, the comb envelope and the Cherenkov peak are highlighted by using dashed curves. In the exemplary embodiment, at V_G=−1 V, GVD β₂˜−62 fs²/mm while third-order-dispersion (“TOD”) β₃˜−9 fs³/mm, the Kerr comb has a span of ˜350 nm, with a highly symmetrical shape. With V_G=−1.2 V, β₂˜−33 fs²/mm while β₃˜−630 fs³/mm, a frequency comb spectra spanning over 600 nm is observed, which is consistent with the general route of a smaller GVD bringing about a broader comb spectrum. The comb spectrum is highly asymmetric, with intensity contributions from Cherenkov radiation. The central spectral position of the Cherenkov radiation can be determined by β₂/β₃. In this case, the energy transfer can also be helpful to stabilize the frequency comb and generate Cherenkov solitons. Thirdly, when V_G=−1.5 V, β₂˜−8 fs²/mm while β₃˜−213 fs³/mm. Since β₂ here is quite small (less than 10 fs²/mm), it can be difficult to support a stable Kerr comb. Furthermore, in the exemplary embodiment, the observed comb lines are not even, and the Cherenkov peak in the spectral is also indistinguishable.

FIG. 3C summarizes the gate tunability of the graphene Kerr combs in accordance with an embodiment of the invention. V_G ranging from 0 V to −2 V are shown. For primary comb lines, their relative spectral location Δf_pri=|f_pri−f_pump| can be dramatically controlled (from 2.3 THz to 7.2 THz) with V_G only from −1.0 V to −1.5 V respectively. This modulation matches the GVD engineering well and can be influenced by the slight nonlinearity enhancement brought by the graphene. For full-generated combs, FIG. 3C also demonstrates the electric-field control on their spectral span, across 38 THz to 82 THz with V_G from −1.0 V to −1.3 V. Moreover, the gate-tuning can change the FSRs of the combs, from 89.6 GHz at −1.0 V to 89.9 GHz at −1.5 V. Such an optoelectronic tunability can enable different Kerr frequency combs with a variety of properties existing in the same device. FIG. 3D illustrates measured locations of Cherenkov peaks in comparison with computed designs in accordance with an embodiment of the invention. In contrast to the primary comb lines, the TOD can play an important role in the Cherenkov radiation. The spectral location of the Cherenkov peak can directly reflect the value of TOD. In the exemplary embodiment, three Cherenkov peaks are observed in the 1400 nm the 2000 nm window, with spectral locations Δf_c=|f_c−f_pump| spanning from 26.3 THz (V_G=−1.2 V), 49.2 THz (V_G=−1.3 V), and 17.7 THz (V_G=−1.5 V). In the exemplary embodiments in FIGS. 3C and 3D, results are collected in the region of −0.4 V to −1.6 V since, when V_G is more than −0.4 V, the Q factor of the GMR can be too low for comb generation and, when V_G is less than −1.6 V, the GVD can be too small to ensure a stable comb.

FIG. 3E shows an estimation of the modulation speed of a GMR in accordance with an embodiment of the invention. With V_G tuning, the output comb line intensity within the filter window can be modulated temporally. The modulation speed can be bounded by ion diffusion in the heterostructure, large ion-gel capacitance on the graphene, and the optical filter bandwidth. In the illustrative embodiment, an ion-gel based capacitor was used to ensure |E_F| is sufficiently high. The ion-gel based capacitor can have large capacitance (7.2 μF/cm2) and slow ion diffusion 10−10 m2/s) to limit the charge-discharge operation speed to less than hundreds of kHz. Optical filter bandwidth can be narrowed to improve the detection rate of the modulation by almost 7.5×, as shown in FIG. 3E. The modulated signal-to-noise ratio (“SNR”) can be shown with an RF spectrum analyzer by using optical filters with passband width of 50 nm, 9 nm, and 2 nm. Corresponding bandwidths were 80 kHz, 200 kHz and 600 kHz, respectively. Although sub-MHz modulation for the primary comb can be successfully demonstrated, fast modulation while preserving the full-grown Kerr comb across the entire modulation cycle could be much more challenging: with V_G tuning, not only the GVD, but also the FSR of the GMR is tuned. Compared to the primary combs shown in FIG. 3A, phase-matching of the full combs in FIG. 3B is more sensitive. A slight variation in the FSR from the gate modulation can cause the Kerr comb to collapse. To achieve reliable fast on-off switching in full-generated Kerr combs, inverse FSR compensation (e.g., via temperature feedback) can be be applied. Such a sub-MHz tunability for a Kerr comb could be potential in diverse applications such as tunable THz applications and precision measurements.

Dispersion is one of the most critical cavity parameter that defines the Kerr frequency comb dynamics. The significant broadband dispersion modulation, controlled by the gate voltage of the graphene-nitride heterogeneous microresonator, opens up the possibility of dynamically selecting the formation path of dissipative Kerr solitons and frequency combs. By using the gate tunable GMRs, we can engineer the dispersion dynamically for diverse soliton states formation via electrical control. With a fixed pump power of 2 W, FIG. 3F shows achieved soliton states in the measurements with the same experimental condition, when the gate voltage is in range of −1.6 V to −1.1 V. In total, soliton states with soliton number of 12, 11, 9, 8, 6, 5 and 4 can be found.

FIGS. 4A-4D conceptually illustrate four specific examples of the soliton crystal states under optimized gate voltages in accordance with various embodiments of the invention. In FIGS. 4A-4D, the left panels show the measured intensity transmission, the middle panels demonstrate the optical spectra, and the right panels illustrate the frame-by-frame frequency-resolved second-harmonic autocorrelation maps. These soliton states with low RF noise can be achieved following Turing patterns and chaotic states before transition into the soliton states. This can be characterized by transmission step, by tuning the pump laser gradually into the cavity resonance. FIG. 4A shows two examples of the soliton state with missing pulses, at a gate voltage V_G of −1.2 V. The corresponding pump laser wavelength is around 1600.2 nm. The optical spectra of these states can be characterized by the apparent existence of groups of comb lines that are separated by multiple cavity FSRs. Within each comb group, weaker single-FSR comb lines are present and they can effectively connect all comb groups without any spectral gaps. For the examples shown on FIGS. 4A and 4D, the comb groups are separated by 8 FSR, 5 FSR and 12 FSR, respectively. In the time domain, the AC traces reveal the common features of missing pulses in the otherwise equally spaced soliton states with higher effective repetition rate. The self-organization of multiple soliton pulses into equally spaced soliton pulse train resembles the crystallization process, or termed a soliton crystal, and the missing pulse structure is analogous to defects in crystal lattices. Graphene-nitride heterogeneous microresonators in accordance with various embodiments of the invention thus can provide a platform for gate-voltage and Femi-level tunable study of the soliton physics. When the soliton crystals are formed, the emitted soliton Cherenkov radiations are sharp and narrow, marked by the arrows 400.

Soliton crystals can be formed due to the strong mode interaction and intra-cavity interferences, and thus their evolution dynamics can depend critically on the exact dispersion profile of the microresonator. By further optimizing the GVD and TOD via gate tuning, two periodic soliton crystal states can be demonstrated. FIG. 4B shows a 4-soliton state with V_G=−1.3 V and pump laser at 1584.2 nm, while FIG. 4C shows an 11-soliton state with V_G=−1.4 V and pump laser at 1600.1 nm. These soliton crystal states show remarkable stability and can robustly survive a significant pump power fluctuation up to ±2 dB, or wavelength offset up to ±300 μm. The soliton crystal formation is also akin to harmonic mode-locking where stable high-repetition-rate pulse train can be attained even in longer cavities, and it attracts interests in applications such as high-speed communication, comb spectroscopy, and data storage. This first realization of the charge-tunable graphene heterostructure for controllable frequency combs and soliton dynamics opens a new architecture at the interface of single atomic layer nanoscience and ultrafast optoelectronics.

Device Design and Fabrication

A microresonator can be fabricated in various dimensions using any of a variety of embodiments. In many embodiments, the microresonator can be fabricated to have a 1000×800 nm² cross-section bus waveguide and a 1200×800 nm² cross-section core for the ring. By using photolithography followed by buffered oxide etching, a SiO₂ window can be created above the ring and exposed Si₃N₄ waveguide, resulting in a distance between the core and air around 300 nm. A monolayer graphene can then be transferred onto the etched window, followed by photolithography patterning and oxygen plasma etching. The graphene layer can be grown by chemical vapor deposition method on copper substrate and transferred using wet transfer technique. The monolayer graphene can be lithographically cut into an 80×100 μm² sheet. Next, source-drain electrodes (Ti/Au, 20/50 nm) can be deposited and patterned using electron-beam evaporation and electron-beam lithography. In some embodiments, the pad size is 80 μm×60 μm. Subsequently, ionic liquid [DEME-TFSI (N,N-diethyl-Nmethyl-N-(2-methoxyethyl)ammonium bis(trifluoromethylsulfonyl)imide, from Sigma-Aldrich] can be integrated as the gate dielectric, resulting in an electric dual-layer graphene transistor.

Q Factor and Dispersion Measurement

The microring resonator transmission can be measured using a tunable laser swept through its full wavelength tuning range at a speed of 40 nm/s. Accordingly, dispersion and Q factors can be measured. For accurate wavelength calibration, 1% of the laser output can be directed into a fiber coupled hydrogen cyanide gas cell (HCN-13-100, Wavelength References Inc.). The microring resonator and gas cell transmission can be recorded during the laser sweep by a data acquisition system through an unbalanced fiber Mach-Zehnder Interferometer (MZI). The MZI has an approximately 40 m path length difference, ensuring the measurement optical frequency sampling resolution of 5 MHz. Each resonance can be fitted with a Lorentzian lineshape unless a cluster of resonances are deemed too close to achieve a conclusive fit with a single Lorentzian. Then, an N-Lorentzian fit can be utilized, where N is the number of resonances being fitted. The dispersion of the ring resonator can be determined by analyzing the wavelength dependence of the FSR. In some embodiments, the microring resonator chip is gated by a probe.

Heterodyne and Autocorrelation

To measure the stability and soliton states of a frequency comb, a wavelength-division multiplexer (“WDM”) can be used to divide the C-band comb lines (1530 nm to 1570 nm) and the L-band comb lines (1570 nm to 1630 nm) with the pump laser. The 1570 nm to 1630 nm window can be monitored by an optical spectrum analyzer (“OSA”) (Advantest AQ8384). A stable CW laser with narrow linewidth (300 kHz, New Focus) can be applied as the heterodyne reference or the 1530 nm to 1570 nm beatnotes. A polarization controller can be used to optimize the input polarization. The beatnotes can be measured by a 3 GHz RF electric spectrum analyzer (“ESA”) (Agilent CXA 9000A). The comb signal can also be measured in the time-domain, by using sensitive autocorrelator. For the autocorrelation, a section of ‘7 m DCF+15 m SMF’ can be used to compensate the GVD, avoiding the pulse broadening.

Although specific gate-tunable frequency combs are discussed above, many different optical frequency combs can be implemented in accordance with many different embodiments of the invention. It is therefore to be understood that the present invention may be practiced in ways other than specifically described, without departing from the scope and spirit of the present invention. Thus, embodiments of the present invention should be considered in all respects as illustrative and not restrictive. Accordingly, the scope of the invention should be determined not by the embodiments illustrated, but by the appended claims and their equivalents.

Claims

1. A microresonator comprising: a silicon nitride ring cladded with silicon dioxide, wherein the silicon oxide defines a window above a region of the silicon nitride ring;a bus waveguide coupled to the silicon nitride ring;a graphene ion-gel heterostructure disposed within the window above the region of the silicon nitride ring, wherein the graphene/ion-gel heterostructure comprises: a graphene monolayer;source-drain electrodes integrated on the graphene monolayer; anda layer of ionic liquid used as a gate dielectric;wherein the microresonator is configured to form a frequency comb tunable through the modulation of the dispersion of the microresonator.

2. The microresonator of claim 1, wherein the modulation of the intracavity dispersion is achieved through the tuning of the Fermi level of the graphene monolayer using an applied gate voltage.

3. The microresonator of claim 2, wherein the Fermi level of the graphene monolayer can be tuned from about 0.45 eV to about 0.65 eV.

4. The microresonator of claim 2, wherein the dispersion of the microresonator can be tuned from about −62 fs2/mm anomalous dispersion to about +9 fs2/mm normal dispersion through tuning the gate voltage from about −2 V to about 0 V.

5. The microresonator of claim 2, wherein the microresonator is configured to form a frequency comb having a relative spectral location that is tunable from about 2.3 THz to about 7.2 THz.

6. The microresonator of claim 2, wherein the microresonator is configured to form a frequency comb having a relative spectral location that is tunable from about 2.3 THz to about 7.2 THz by tuning the applied gate voltage from about −1 V to about −1.5 V.

7. The microresonator of claim 1, wherein a fixed source-drain voltage of 10 mV is applied to the graphene ion-gel heterostructure.

8. The microresonator of claim 1, wherein the microresonator is capable of achieving multi-soliton states.

9. The microresonator of claim 1, wherein the microresonator is capable of achieving soliton states with soliton numbers of 12, 11, 9, 8, 6, 5, and 4.

10. The microresonator of claim 1, wherein the microresonator has a loaded quality factor of about 1.6×106.

11. The microresonator of claim 1, wherein the source-drain electrodes comprise Ti/Au.

12. The microresonator of claim 1, wherein the ionic liquid comprises diethylmethyl(2-methoxyethyl)ammonium bis(trifluoromethylsulfonyl)imide.

13. The microresonator of claim 1, wherein the layer of ionic liquid provides a capacitance of about 7.2 μF/cm2.

14. The microresonator of claim 1, wherein the graphene monolayer is disposed about 100 nm away from the silicon nitride ring.

15. The microresonator of claim 1, wherein the graphene monolayer is disposed about 300 nm away from the silicon nitride ring.

16. The microresonator of claim 1, wherein the graphene monolayer and the silicon nitride ring have a planar interaction arc length of about 80 μm.

17. The microresonator of claim 1, wherein the silicon nitride ring has a cross section of about 1200 nm×800 nm.

18. The microresonator of claim 1, wherein the silicon nitride ring has a diameter of about 350 μm.

19. The microresonator of claim 1, wherein the bus waveguide has a cross section of about 1000 nm×800 nm.

20. The microresonator of claim 1, wherein the bus waveguide is disposed about 600 nm away from the silicon nitride ring.