Giant enhancement of stimulated Brillouin scattering with engineered phoxonic crystal waveguides

Zejie Yu; Xiankai Sun

doi:10.1364/OE.26.001255

1. Introduction

Stimulated Brillouin scattering (SBS) is a nonlinear optical process in a bulk material where traveling light waves are scattered by traveling acoustic waves [1]. This effect is greatly enhanced in nanosized waveguides owing to the mutual optomechanical interaction: the interplay between the pump and the Stokes waves generates dynamic optical forces via optical radiation pressure and electrostrictive effect; the mechanical vibration of the waveguide excited by the optical forces in turn scatters light from the pump into the Stokes wave [2]. Exploiting this physical mechanism, Brillouin photonics has rapidly grown as a new branch of cavity optomechanics, producing various applications for wideband RF and photonic signal processing [3–6]. For example, traveling-wave Brillouin processes have enabled unique schemes for optical pulse compression, pulse and waveform synthesis [4, 7–9], coherent frequency comb generation [4, 7, 10], variable-bandwidth optical amplifiers [9, 11], reconfigurable filters [12], coherent beam combining [13], and nonreciprocal transmission and light storage [14, 15].

More recently, the creation of strong Brillouin nonlinearities has been achieved in nanoscale silicon waveguides on an integrated platform [16–18]. SBS was first achieved in a bare silicon waveguide with amplification of 0.5 dB (12%) by using suspended silicon nanowire structures [16]. Later, Kittlaus et al. reported large Brillouin amplification in a membrane-suspended silicon waveguide, which efficiently suppresses other nonlinear effects such as the two-photon absorption and the subsequent carriers-induced free-carrier absorption [18]. The devices achieved amplification levels greater than 5 dB and demonstrated a record low (5 mW) threshold for net amplification. However, all the realizations are based on waveguides extending several centimeters in length to obtain sufficiently strong interaction between light and sound. Such long waveguide lengths are incompatible with device miniaturization and thus undesirable for on-chip integration. Additionally, compared with silica, the relatively high stiffness of silicon makes guiding of acoustic waves a challenging task, thus motivating fabrication of a delicately suspended waveguide [16] or a complicated membrane waveguide [18] to reduce the phonon dissipation.

Photonic and phononic crystals are artificial periodic structures which exhibit frequency bandgaps for optical (light) and acoustic (sound) waves, respectively. Structures that are designed to have simultaneous photonic and phononic bandgaps are referred to as phoxonic crystals. They possess all the optical and acoustic properties of the individual photonic and phononic crystals. As such, phoxonic crystal waveguides can guide propagating optical and acoustic waves [19] and can even achieve simultaneous slow light and slow sound if properly designed [20]. Phoxonic crystal cavities can confine standing optical and acoustic waves, producing large modal overlap and enabling strong optomechanical interaction [21, 22]. These phoxonic crystal structures have been employed for a wide range of optomechanical applications.

In this paper, we engineer a two-dimensional phoxonic crystal waveguide structure [20, 23–30] which simultaneously confines and guides slow traveling optical and acoustic waves in a nanoscale line defect to achieve strong SBS and to obtain giant enhancement of SBS gain. The defect-guided optical and acoustic waves can both be slowed down due to their relatively flat dispersion, leading to substantially increased interaction time between the light and sound. In addition, the phoxonic crystal structure also reduces phonon dissipation from the defect waveguide due to the complete structural suspension and the absolute phononic bandgap. The proposed structure is also easy to fabricate by using one step of electron-beam lithography and dry etching followed by underlying oxide removal on a standard silicon-on-insulator wafer. Our theoretical and numerical results show that the slow light and the phase-matching condition among the pump light, Stokes light, and acoustic wave can both be achieved in the entire C band, covering the wavelength range of 1520–1565 nm. As a result, the SBS gain is enhanced significantly, with the calculated gain coefficient greater than 3 × 10⁴ W⁻¹ m⁻¹ in the entire C band and the highest value beyond 10⁶ W⁻¹ m⁻¹. This is at least an order of magnitude higher than those in the existing demonstrations, which are 3,055 W⁻¹ m⁻¹, 2,750 W⁻¹ m⁻¹, and 1,020 W⁻¹ m⁻¹ in [16], [17], and [18], respectively. This giant enhancement of SBS gain will lead to significant reduction of device footprint and thus can tremendously increase the integration density of the SBS-based optoelectronic devices such as Brillouin lasers, amplifiers, and signal processors on a chip.

2. Derivation of the SBS gain of a phoxonic crystal waveguide

We use the Bloch wavevector k (q) to characterize the propagating optical (acoustic) mode of a phoxonic crystal waveguide with periodicity along the x direction. In a typical SBS process, the pump (p) and Stokes (s) waves can be described by vector fields

E_{p} (x, y, z, t) = a_{p} (x, t) e^{p} (x, y, z, t) + c . c .,

E_{s} (x, y, z, t) = a_{s} (x, t) e^{s} (x, y, z, t) + c . c .,

where the pump and Stokes modal fields with respective wavevector k_p and k_s are expressed as

e^{p} (x, y, z, t) = {\tilde{e}}^{p} (x, y, z) \exp [- i (ω_{p} t - k_{p} x)],

e^{s} (x, y, z, t) = {\tilde{e}}^{s} (x, y, z) \exp [- i (ω_{s} t - k_{s} x)] .

Here, a_p and a_s are the envelope functions of the electric fields. ẽ^p and ẽ^s are the basis functions of the electric fields which vary periodically in the x direction. ω_p and ω_s are the angular frequencies of the pump light and the Stokes light. The pump and Stokes light’s energy flux and energy stored within one period of a phoxonic crystal waveguide can be expressed as [31]

P_{p} = 2 \int \hat{x} \cdot {{[e^{p} (x, y, z, t)]}^{*} \times h^{p} (x, y, z, t)} d y d z,

P_{s} = 2 \int \hat{x} \cdot {{[e^{s} (x, y, z, t)]}^{*} \times h^{s} (x, y, z, t)} d y d z,

ξ_{p} = 2 \int_{0}^{a} d x \int ε (x, y, z) {[e^{p} (x, y, z, t)]}^{*} \cdot e^{p} (x, y, z, t) d y d z,

ξ_{s} = 2 \int_{0}^{a} d x \int ε (x, y, z) {[e^{s} (x, y, z, t)]}^{*} \cdot e^{s} (x, y, z, t) d y d z,

where a denotes the period of phoxonic crystal waveguide in the x direction. The energy flux is independent of position in the phoxonic crystal waveguide along the traveling direction [31]. The energy flow rates, or group velocities, of the pump and Stokes light can be expressed as v_p = aP_p/ξ_p and v_p = aP_s/ξ_s, respectively.

Similarly, an optically generated acoustic wave can be expressed as

U (x, y, z, t) = b (x, t) u (x, y, z, t) + c .c .,

where the acoustic modal field with wavevector q = k_p − k_s and angular frequency Ω = ω_p − ω_s is expressed as

u (x, y, z, t) = \tilde{u} (x, y, z) \exp [- i (Ω t - q x)] .

Here, b is the envelope function of the acoustic wave, and ũ is the basis function of the acoustic wave which varies periodically in the x direction. The acoustic energy flux and energy stored within one period of a phoxonic crystal waveguide can be expressed as

P_{b} = - 2 i Ω \int \sum_{j m l} c_{x j m l} u_{j}^{*} \frac{\partial u_{l}}{\partial m} d y d z,

ξ_{b} = 2 Ω^{2} \int_{0}^{a} d x \int ρ {| u |}^{2} d y d z,

where c is the stiffness tensor and ρ is the density of the medium. ∂/∂m denotes the spatial derivative with respect to the m direction (m = x, y, z). The acoustic energy flux is also independent of position in the phoxonic crystal waveguide along the traveling direction. The acoustic energy flow rate can be expressed as v_b = aP_b/ξ_b.

Next, we adopt the method in [32] and extend it to the case of a longitudinally varying waveguide [33] to derive the light–sound interaction equations. The acoustic field induces perturbations to both electric fields (∆E) and electric displacements (∆D), which should also satisfy the Maxwell equations, so we can write the wave equation as

\nabla \times \nabla \times (E + Δ E) + μ_{0} \frac{\partial (D + Δ D)}{\partial t} = 0.

Under the phase-matching condition, we can express ∆E and ∆D as

Δ E = Δ e^{s} a_{p} b^{*} + Δ e^{p} a_{s} b + c .c .,

Δ D = Δ d^{s} a_{p} b^{*} + Δ d^{p} a_{s} b + c .c ..

Then, we evaluate the contribution from the pump light to Eq. (13),

0 = \nabla \times \nabla \times (a_{s} e^{s} + Δ e^{s} a_{p} b^{*}) + μ_{0} \frac{\partial (a_{s} d^{s} + a_{p} b^{*} Δ d^{s})}{\partial t},

\begin{matrix} 0 = a_{s} (\nabla \times \nabla \times e^{s} + μ_{0} \frac{\partial d^{s}}{\partial t}) + [\hat{x} \times (\nabla \times e^{s}) + \nabla \times (\hat{x} \times e^{s})] \frac{\partial a_{s}}{\partial x} - 2 i ω_{s} μ_{0} d^{s} \frac{\partial a_{s}}{\partial t} \\ + a_{p} b^{*} (\nabla \times \nabla \times Δ e^{s} + μ_{0} \frac{\partial^{2} Δ d^{s}}{\partial t^{2}}) + h .o .t . + c .c ., \end{matrix}

where “h.o.t.” stands for the higher-order terms in the perturbations and higher-order derivatives of the envelope functions. We can neglect “h.o.t.” under the slowly-varying-envelope assumption. We can also drop all the complex conjugate terms after projecting onto the mode e^s and averaging over a time interval much longer than the optical time scale. Multiplying (e^s)^* on both sides of Eq. (17) and integrating it over one period in the x direction yields

- i ω_{s} a P_{s} \frac{\partial a_{s}}{\partial x} - i ω_{s} \frac{P_{s}}{v_{s}} \frac{\partial a_{s}}{\partial t} = a_{p} b^{*} \int_{x_{0}}^{x_{0} + a} d x \int [{(e^{s})}^{*} \cdot {(- i ω)}^{2} Δ d^{s} - {(- i ω)}^{2} d^{s} \cdot Δ e^{s}] d y d z .

Since the envelope functions a_s, a_p, and b vary slowly and can be treated as constants over a period, they can be taken out of the integral in the above equation. Following the same steps, we can also obtain the equation for the envelope of the Stokes light which is similar to Eq. (18), so we arrive at a set of coupled equations for the envelope functions

v_{s} \frac{\partial a_{s}}{\partial x} + \frac{\partial a_{s}}{\partial t} = - \frac{i ω_{s} Q_{OM}^{s}}{ξ_{s}} a_{p} b^{*},

v_{p} \frac{\partial a_{p}}{\partial x} + \frac{\partial a_{p}}{\partial t} = - \frac{i ω_{p} Q_{OM}^{p}}{ξ_{p}} a_{s} b,

where

Q_{OM}^{p} = \int_{x_{0}}^{x_{0} + a} d x \int [{(e^{p})}^{*} \cdot Δ d^{p} - d^{p} \cdot Δ e^{p}] d y d z,

Q_{OM}^{s} = \int_{x_{0}}^{x_{0} + a} d x \int [{(e^{s})}^{*} \cdot Δ d^{s} - d^{s} \cdot Δ e^{s}] d y d z .

On the other hand, the acoustic wave in the phoxonic crystal waveguide is governed by

ρ \frac{\partial^{2} U_{i}}{\partial t^{2}} + \sum_{j m l} \frac{\partial}{\partial j} [c_{i j m l} + η_{i j m l} \frac{\partial}{\partial t}] \frac{\partial}{\partial m} U_{l} = - F_{i},

where η is the viscosity tensor. The source term F on the right-hand side is the external driving force field per unit volume, through which the coupling to the electromagnetic fields is introduced. The driving force density F can be expressed as

F (r, t) = f (r, t) a_{s}^{*} (x, t) a_{p} (x, t) + c .c ..

Substituting this ansatz into Eq. (23) and dropping higher-order terms eventually yields

- i Ω \sum_{j m l} [(c_{i x m l} \frac{\partial}{\partial m} + \frac{\partial}{\partial j} c_{i j x l}) u_{l} \frac{\partial b}{\partial x} - 2 i Ω ρ u_{i} \frac{\partial b}{\partial t} + \frac{\partial}{\partial j} η_{i j m l} \frac{\partial u_{l}}{\partial m} b + f_{i} a_{s}^{*} a_{p}] + c .c . = 0.

After projecting onto the mode u under the phase-matching condition, we arrive at the equation for the acoustic envelope function

v_{b} \frac{\partial b}{\partial x} + α_{b} b + \frac{\partial b}{\partial t} = - \frac{i Ω Q_{OM}^{b}}{ξ_{b}} a_{s}^{*} a_{p},

where α_b is the loss rate of the propagating acoustic wave and

Q_{OM}^{b} = \int_{x_{0}}^{x_{0} + a} d x \int u^{*} \cdot f d y d z .

As the acoustic frequency is orders of magnitude smaller than the optical frequency, we can treat the frequencies of the pump (ω_p) and Stokes (ω_s) light to be the same (ω), then we get the relationship if we ignore the irreversible optomechanical coupling in the phoxonic crystal waveguide [32]

Q_{OM}^{s} = {(Q_{OM}^{p})}^{*} = Q_{OM}^{b},

so that we can use a single parameter Q_OM to characterize the optomechanical coupling [34]. In the steady state, all the time derivatives (∂/∂t) become zero. Equations (19), (20), and (26) reduce to a set of coupled equations for the time-independent envelope functions of the pump, Stokes, and acoustic waves in the phoxonic crystal waveguide:

v_{s} \frac{\partial a_{s} (x)}{\partial x} = - \frac{i ω_{s} Q_{OM}^{*}}{ξ_{s}} a_{p} (x) b^{*} (x),

v_{p} \frac{\partial a_{p} (x)}{\partial x} = - \frac{i ω_{p} Q_{OM}}{ξ_{p}} a_{s} (x) b (x),

v_{b} \frac{\partial b (x)}{\partial x} + α_{b} v_{b} b (x) = - \frac{i Ω Q_{OM}^{*}}{ξ_{b}} a_{s}^{*} (x) a_{p} (x) .

The optomechanical coupling is the sum of the coupling coefficients from the photoelastic effect Q_PE [35] and from the moving-boundary effect Q_MB [36]. The photoelastic effect refers to the variation of a material’s electric susceptibility (or dielectric constant) due to the strain. The photoelastic contribution can be derived from a first-order perturbation theory [37]

Q_{PE} = ε_{0} \int_{V} d^{3} r \sum_{i j m l} ε_{r}^{2} {(E_{i}^{p})}^{*} E_{j}^{s} p_{i j m l} S_{m l},

where ε_r is the relative permittivity (dielectric constant) of the periodic waveguide and ε₀ is the vacuum permittivity, p is the rank-four photoelastic tensor, and S is the strain tensor. The moving-boundary effect refers to the frequency variation of an optical mode due to the displacement of structural surfaces. As an acoustic wave propagates along a waveguide, it induces a strong change to the optical fields over a very small area at the waveguide surfaces. The moving-boundary contribution can be expressed as [19]

Q_{MB} = \int_{S} d^{2} r (u^{*} \cdot \hat{n}) [(ε_{r} - ε_{air}) ε_{0} {(\hat{n} \times E^{p})}^{*} (\hat{n} \times E^{s}) - (ε_{r}^{- 1} - ε_{air}^{- 1}) ε_{0}^{- 1} {(\hat{n} \cdot D^{p})}^{*} (\hat{n} \cdot D^{s})],

where

\hat{n}

is the unit vector of surface normal pointing outward, ε_air is the relative permittivity of air, and d^p and d^s are the basis electric displacement fields of the pump and Stokes light. The optomechanical coupling parameter Q_OM depends on the magnitudes of the optical fields and the mechanical displacement of the acoustic field. In the community of cavity optomechanics, the optomechanical coupling coefficient g_OM is defined as the frequency shift of the optical mode per unit length of mechanical displacement. Therefore, the g_OM in our case can be defined as

g_{OM} = ω \sqrt{Q_{OM}^{s} Q_{OM}^{p} / (ξ_{s} ξ_{p})},

where the normalized acoustic modal field [max(|u|) = 1] should be used.

Lastly, the SBS gain coefficient in units of W⁻¹ m⁻¹ under the phase-matching condition can be expressed as [32]

G = \frac{Ω {| g_{OM} |}^{2}}{ω v_{p} v_{s} v_{b} α_{b} (ξ_{b} / a)},

where ξ_b should also be evaluated by using the normalized acoustic modal field. According to Eq. (35), the SBS gain is inversely proportional to the group velocities of both the pump (v_p) and Stokes (v_s) waves. Therefore, the SBS gain will be enhanced by slowing down both the pump and Stokes light’s speeds in the waveguide. Additionally, acoustic waves of a slower group velocity will induce a larger strain in the core of the waveguide than that in the slab, leading to stronger modal confinement to the waveguide core with a higher modal amplitude [24]. As a result, the SBS gain will also be enhanced by slowing down the group velocity of the acoustic wave.

3. Engineering the phoxonic crystal waveguide for obtaining simultaneous slow light and slow sound

A two-dimensional phoxonic crystal structure can be formed by periodic arrangement of cross-shaped openings in a silicon slab, with the unit cell depicted in Fig. 1(a). Among various lattice structures, we choose the square lattice because the triangular lattice cannot achieve a sizable phononic bandgap and the hexagonal lattice’s photonic bandgap is unsuitable for practical use [27]. The geometric parameters of the unit cell are the thickness h = 340 nm, the array pitch a = 490 nm, and the cross-shaped slot size 0.2a × 0.8a. We use the following material parameters for silicon: refractive index n = 3.47, Young’s modulus E = 170 GPa, Poisson’s ratio ν = 0.28, and density ρ = 2329 kg/m³. We also assume that the [100], [010], and [001] crystalline directions are aligned with the x, y, and z axis respectively, where the photoelastic tensor p_ijkl in the contracted notation is (p₁₁, p₁₂, p₄₄) = (−0.09, 0.017, −0.051) [15]. Figures 1(b) and 1(c) show the quasi-TE photonic and phononic band diagrams calculated from MPB [38] and COMSOL Multiphysics, respectively, where the bandgaps are marked in red and the light cone is marked in gray.

Fig. 1 (a) Unit cell of a two-dimensional silicon phoxonic crystal structure. The structural parameters are h = 340 nm and a = 490 nm. (b, c) Calculated quasi-TE photonic (b) and phononic (c) band diagrams of the phoxonic crystal structure with the unit cell shown in (a). The bandgaps are marked in red and the light cone is marked in gray.

Download Full Size | PDF

By removing a row of cross-shaped slots along the Γ–X direction from the above two-dimensional phoxonic crystal structure we create a W1 phoxonic crystal waveguide as shown in Fig. 2(a). This phoxonic crystal waveguide has three symmetry planes x = 0, y = 0, and z = 0, which should be aligned with the respective crystalline planes of the material so that its anisotropic optical, elastic, and photoelastic constants have simple expressions. We calculated the photonic [Fig. 2(b)] and phononic [Fig. 2(c)] band diagrams along the x direction, where the defect-guided optical and acoustic modes (marked in red) are clearly identified in the respective bandgaps. The right panel of Fig. 2(a) shows the modal profiles of the defect-guided optical (upper two) and acoustic (lower two) modes. Figure 2(d) shows the calculated group velocities of the defect-guided optical mode (in units of the speed of light in vacuum, c) and acoustic mode (in units of m/s). Both the light and sound speeds are slowed down remarkably in the defect phoxonic crystal waveguide band and diminish to zero at the edge of the Brillouin zone where the propagating waves reduce to standing waves.

Fig. 2 (a) Left: Schematic of the W1 phoxonic crystal waveguide structure. Right: Profiles of the defect-guided optical mode (|E_y|²) and acoustic mode (U_y). (b) Photonic band diagram of the W1 phoxonic crystal waveguide structure. The cyan-shaded regions denote the slab-guided band continuum, where the modes in the dielectric (air) band reside mostly in the material (air holes). The black and red lines in the bandgap denote the defect-guided optical modes which do not cross each other in the entire k_x range. (c) Phononic band diagram of the W1 phoxonic crystal waveguide structure. The light blue regions denote the phononic bandgaps. The red line denotes the defect-guided acoustic mode. (d) Group velocities of the defect-guided optical (v_l,g, in red) and acoustic (v_s,g, in blue) modes propagating in the phoxonic crystal waveguide.

Download Full Size | PDF

The SBS gain in a phoxonic crystal waveguide can be greatly enhanced if the three requirements are satisfied: (i) simultaneous slow optical and acoustic propagating waves, (ii) phase matching among the pump light, the Stokes light, and the acoustic wave, and (iii) strong optomechanical coupling between the slow optical and acoustic modes. First, let us consider a forward SBS process where a defect-guided pump light wave (angular frequency ω_p, wavevector k_p) interacts with a defect-guided acoustic wave (Ω, –q), producing a scattered Stokes light wave (ω_p – Ω, k_p + q). Since the acoustic frequency Ω is typically orders of magnitude smaller than ω_p, the scattered Stokes light mode is in the same defect-guided band as the pump light. In addition, due to the substantially different slopes of the dispersion curves, the acoustic wavevector q which matches the difference between the wavevectors of the pump light and Stokes light should approach zero. Therefore, all the three waves, the pump light, the Stokes light, and the acoustic wave, are in the respective slow-light and slow-sound regions, simultaneously.

Next, we explain how the phase-matching condition can be achieved among the pump light, the Stokes light, and the acoustic wave. As shown in Fig. 3, the pump light and Stokes light are both in the slow-light region of the defect-guided optical mode. In the right zoomed-in diagram, the green (red) line denotes the dispersion curve of the defect-guided optical (acoustic) mode traveling in the x direction of the phoxonic crystal waveguide. We superimpose the phononic band diagram onto the photonic band diagram by aligning the origin of the former to the operating point of the pump light in the latter. Since the frequency variation of an optical mode is orders of magnitude larger than that of the acoustic mode with the same wavevector, the dispersion curve of the optical mode should be much steeper than that of the acoustic mode, as shown in the right zoomed-in diagram of Fig. 3. Therefore, an intersection point between the photonic and phononic dispersion curves should always exist, which sets the phase-matching condition for the Stokes light.

Fig. 3 Illustration showing the phase matching among the pump light, the Stokes light, and the acoustic wave in the engineered phoxonic crystal waveguide. The pump light and Stokes light are both in the slow-light region of the defect-guided mode. In the right zoomed-in diagram, the green (red) line denotes the dispersion curve of the defect-guided optical (acoustic) mode traveling along the x direction in the phoxonic crystal waveguide. The phononic band diagram is superimposed onto the photonic band diagram with the origin of the former aligned to the operating point of the pump light in the latter.

Download Full Size | PDF

Third, a large optomechanical coupling coefficient g_OM can be obtained with all the three waves confined laterally and propagating slowly in the phoxonic crystal defect waveguide because of a large overlap between their modal fields. Actually, those defect-guided light modes can be categorized into “y_even/y_odd (z_even/z_odd)” according to the symmetry of their fields with respect to the y = 0 (z = 0) plane. As seen from the right panel of Fig. 2(a), the dominant component of the electric field |E_y|² is both y_even and z_even. Since the pump light and the Stokes light share similar modal field patterns, the resulting optical force distribution is also symmetric with respect to planes y = 0 and z = 0. In addition, the right panel of Fig. 2(a) also shows the acoustic modal profile U_y (mechanical displacement relative to the structural surface) of the phoxonic crystal waveguide. It is easy to find that the acoustic mode shares the same field symmetry with the optical force distribution, which renders a strong optomechanical coupling coefficient in the phoxonic crystal waveguide.

4. Calculation of the SBS gain in the engineered phoxonic crystal waveguide

In order to calculate the SBS gain coefficient in Eq. (35), we need all the six parameters g_OM, ξ_b, v_p, v_s, v_b, and α_b. Owing to the intrinsic material loss [39], the acoustic wave is attenuated as it propagates along the waveguide. According to the measurement data from bulk-mode acoustic resonators, it is reasonable to assume a dissipation rate α_bv_b of 1.63 × 10⁷ rad/s [40]. In typical cases of optical fibers and waveguides, the propagation loss of the pump and Stokes waves is negligible over the acoustic attenuation length α_b⁻¹ [33, 34]. In the case of phoxonic crystal waveguide considered here, the per-unit-length loss rates of optical and acoustic waves are both enhanced due to the slow-light and slow-sound effects. We calculated the group velocities v_p, v_s, and v_b in the slow-light region. Since the frequency difference between the pump and Stokes light is quite small compared with their optical frequencies, the pump light and Stokes light can be regarded to share the same group velocity. Figure 4(a) shows the group velocities of the pump and Stokes light v_l,g and acoustic wave v_s,g with the pump light wavelength varying in 1520–1565 nm under the phase-matching condition. It is clear that the three waves, the pump and Stokes light as well as the acoustic wave, all exhibit a great amount of slowing effect in the entire C band. It should be noted that the group velocity of light maintains larger than 0.005c from 1520 to 1565 nm, indicating that the optical slow-down factor is less than 100. In contrast, the acoustic wave under the phase-matching condition is slowed down by more than 1000 times. In this regard, it is still valid to assume negligible propagation loss of the pump and Stokes waves over the acoustic attenuation length α_b⁻¹.

Fig. 4 (a) Calculated group velocities of the pump and Stokes light v_l,g (red) and acoustic wave v_s,g (blue) at the phase-matching condition with the pump light wavelength in the C band. (b) Calculated SBS gain coefficient (red) and optomechanical coupling coefficient g_OM (blue) at the phase-matching condition with the pump light wavelength in the C band.

Download Full Size | PDF

Due to their small frequency difference, the pump light and Stokes light can also be regarded to share the same modal fields. The blue line in Fig. 4(b) shows the calculated g_OM with the pump light wavelength varying in 1520–1565 nm under the phase-matching condition, where the red line shows the forward SBS gain coefficient. It is easy to find that as the pump wavelength moves from 1520 to 1565 nm, the g_OM increases by a factor of two while the SBS gain coefficient is enhanced by near two orders of magnitude, which confirms that the giant enhancement of the SBS gain results predominantly from the simultaneous slow-light and slow-sound effects.

It should be noted that the advantage of engineering simultaneous slow light and slow sound in a phoxonic crystal waveguide is to enhance the spatial interaction of light and sound, so as to increase the Brillouin scattering rate per unit length. The scattering rate is not enhanced in the time domain. Assuming a forward evolution direction (+x) of the acoustic wave, we solve Eqs. (29)–(31) by the Green’s function method [32],

b (x) = - i Ω g_{OM} \int_{0}^{+ \infty} d x^{'} [a_{s}^{*} (x - x^{'}) a_{p} (x - x^{'}) \exp (- α_{b} x^{'})] = \frac{i g_{OM} a_{s}^{*} (x) a_{p} (x)}{α_{b} v_{b}} .

Actually, the factor α_bv_b in Eq. (36) which characterizes the acoustic loss per unit time is a constant for a certain acoustic mode. Therefore, |b(x)| in the final steady state depends only on the optical fields a_s^*a_pand the optomechanical coupling strength g_OM, while the slow group velocity v_b does not play a role. Here, the advantage of the slow sound lies in the reduction of the device size that is required for the acoustic wave to reach its final steady state. The acoustic wave needs a length to build up from zero to its steady-state amplitude which is set by the optical fields a_s^*a_p. We solved b(x) numerically based on Eqs. (29)–(31) with initial a_p and a_s (|a_p| >> |a_s|). Figure 5 plots the evolution amplitude |b(x)| for acoustic waves of slow (v_g = 1.08 m/s) and normal (v_g = 5,330 m/s) group velocities. It is clear that they increase at the same rate in the end, but the length required to reach the final steady rate depends on the group velocity. Therefore, it is in this way that the engineered slow sound can effectively enhance the SBS gain and reduce the device size.

Fig. 5 Evolution of the amplitude of the acoustic wave |b(x)| along the propagation direction, where red solid and blue dash-dotted lines correspond to acoustic waves of slow (v_g = 1.08 m/s) and normal (v_g = 5,330 m/s) group velocities, respectively.

Download Full Size | PDF

5. Conclusion

We have engineered a suspended phoxonic crystal waveguide structure for giant enhancement of the forward stimulated Brillouin scattering. The designed structure can provide tight lateral confinement for both optical and acoustic modes to copropagate within a guided line defect thus enabling strong optomechanical coupling between them. We have also engineered the photonic and phononic band structures to achieve simultaneous slow light and slow sound. The combination of these two effects results in significantly enhanced SBS nonlinearity over that in conventional nonlinear fibers, suspended silicon waveguides, and hybrid Si/Si₃N₄ waveguides, etc. The calculated SBS gain coefficient is greater than 3 × 10⁴ W⁻¹ m⁻¹ in the entire C band with the highest value beyond 10⁶ W⁻¹ m⁻¹. This is at least an order of magnitude higher than those in the existing demonstrations in silicon. The giant SBS gain will enable high-performance Brillouin photonic devices with substantially reduced footprint and open up applications of photonics-assisted RF and microwave signal generation, amplification, and processing.

Funding

Hong Kong Research Grants Council Early Career Scheme (24208915); National Natural Science Foundation of China (NSFC) / Research Grants Council (RGC) of Hong Kong Joint Research Scheme (N_CUHK415/15).

References and links

1. R. W. Boyd, Nonlinear Optics (Academic, 2003).

2. P. T. Rakich, P. Davids, and Z. Wang, “Tailoring optical forces in waveguides through radiation pressure and electrostrictive forces,” Opt. Express 18(14), 14439–14453 (2010). [CrossRef] [PubMed]

3. P. Dainese, P. S. J. Russell, N. Joly, J. C. Knight, G. S. Wiederhecker, H. L. Fragnito, V. Laude, and A. Khelif, “Stimulated Brillouin scattering from multi-GHz-guided acoustic phonons in nanostructured photonic crystal fibres,” Nat. Phys. 2(6), 388–392 (2006). [CrossRef]

4. M. S. Kang, A. Nazarkin, A. Brenn, and P. S. J. Russell, “Tightly trapped acoustic phonons in photonic crystal fibres as highly nonlinear artificial Raman oscillators,” Nat. Phys. 5(4), 276–280 (2009). [CrossRef]

5. P. T. Rakich, C. Reinke, R. Camacho, P. Davids, and Z. Wang, “Giant enhancement of stimulated Brillouin scattering in the subwavelength limit,” Phys. Rev. X 2(1), 011008 (2012). [CrossRef]

6. J. Wang, Y. Zhu, R. Zhang, and D. J. Gauthier, “FSBS resonances observed in a standard highly nonlinear fiber,” Opt. Express 19(6), 5339–5349 (2011). [CrossRef] [PubMed]

7. D. Braje, L. Hollberg, and S. Diddams, “Brillouin-enhanced hyperparametric generation of an optical frequency comb in a monolithic highly nonlinear fiber cavity pumped by a cw laser,” Phys. Rev. Lett. 102(19), 193902 (2009). [CrossRef] [PubMed]

8. T. Schneider, M. Junker, and D. Hannover, “Generation of millimetre-wave signals by stimulated Brillouin scattering for radio over fibre systems,” Electron. Lett. 40(23), 1500–1502 (2004). [CrossRef]

9. T. Sakamoto, T. Yamamoto, K. Shiraki, and T. Kurashima, “Low distortion slow light in flat Brillouin gain spectrum by using optical frequency comb,” Opt. Express 16(11), 8026–8032 (2008). [CrossRef] [PubMed]

10. M. J. Damzen and M. H. R. Hutchinson, “High-efficiency laser-pulse compression by stimulated Brillouin scattering,” Opt. Lett. 8(6), 313–315 (1983). [CrossRef] [PubMed]

11. N. Olsson and J. V. D. Ziel, “Characteristics of a semiconductor laser pumped brillouin amplifier with electronically controlled bandwidth,” J. Lightwave Technol. 5(1), 147–153 (1987). [CrossRef]

12. T. Tanemura, Y. Takushima, and K. Kikuchi, “Narrowband optical filter, with a variable transmission spectrum, using stimulated Brillouin scattering in optical fiber,” Opt. Lett. 27(17), 1552–1554 (2002). [CrossRef] [PubMed]

13. B. C. Rodgers, T. H. Russell, and W. B. Roh, “Laser beam combining and cleanup by stimulated Brillouin scattering in a multimode optical fiber,” Opt. Lett. 24(16), 1124–1126 (1999). [CrossRef] [PubMed]

14. C.-H. Dong, Z. Shen, C.-L. Zou, Y.-L. Zhang, W. Fu, and G.-C. Guo, “Brillouin-scattering-induced transparency and non-reciprocal light storage,” Nat. Commun. 6, 6193 (2015). [CrossRef] [PubMed]

15. J. Kim, M. C. Kuzyk, K. Han, H. Wang, and G. Bahl, “Non-reciprocal Brillouin scattering induced transparency,” Nat. Phys. 11(3), 275–280 (2015). [CrossRef]

16. R. Van Laer, B. Kuyken, D. Van Thourhout, and R. Baets, “Interaction between light and highly confined hypersound in a silicon photonic nanowire,” Nat. Photonics 9(3), 199–203 (2015). [CrossRef]

17. H. Shin, W. Qiu, R. Jarecki, J. A. Cox, R. H. Olsson III, A. Starbuck, Z. Wang, and P. T. Rakich, “Tailorable stimulated Brillouin scattering in nanoscale silicon waveguides,” Nat. Commun. 4, 1944 (2013). [CrossRef] [PubMed]

18. E. A. Kittlaus, H. Shin, and P. T. Rakich, “Large Brillouin amplification in silicon,” Nat. Photonics 10(7), 463–467 (2016). [CrossRef]

19. A. H. Safavi-Naeini and O. Painter, “Design of optomechanical cavities and waveguides on a simultaneous bandgap phononic-photonic crystal slab,” Opt. Express 18(14), 14926–14943 (2010). [CrossRef] [PubMed]

20. V. Laude, J.-C. Beugnot, S. Benchabane, Y. Pennec, B. Djafari-Rouhani, N. Papanikolaou, J. M. Escalante, and A. Martinez, “Simultaneous guidance of slow photons and slow acoustic phonons in silicon phoxonic crystal slabs,” Opt. Express 19(10), 9690–9698 (2011). [CrossRef] [PubMed]

21. M. Eichenfield, J. Chan, R. M. Camacho, K. J. Vahala, and O. Painter, “Optomechanical crystals,” Nature 462(7269), 78–82 (2009). [CrossRef] [PubMed]

22. A. H. Safavi-Naeini, J. T. Hill, S. Meenehan, J. Chan, S. Gröblacher, and O. Painter, “Two-dimensional phononic-photonic band gap optomechanical crystal cavity,” Phys. Rev. Lett. 112(15), 153603 (2014). [CrossRef] [PubMed]

23. R. Zhang, G. Chen, and J. Sun, “Analysis of acousto-optic interaction based on forward stimulated Brillouin scattering in hybrid phononic-photonic waveguides,” Opt. Express 24(12), 13051–13059 (2016). [CrossRef] [PubMed]

24. J. M. Escalante, A. Martínez, and V. Laude, “Design of single-mode waveguides for enhanced light-sound interaction in honeycomb-lattice silicon slabs,” J. Appl. Phys. 115(6), 064302 (2014). [CrossRef]

25. F.-L. Hsiao, C.-Y. Hsieh, H.-Y. Hsieh, and C.-C. Chiu, “High-efficiency acousto-optical interaction in phoxonic nanobeam waveguide,” Appl. Phys. Lett. 100(17), 171103 (2012). [CrossRef]

26. Y. Pennec, J. O. Vasseur, B. Djafari-Rouhani, L. Dobrzyński, and P. A. Deymier, “Two-dimensional phononic crystals: Examples and applications,” Surf. Sci. Rep. 65(8), 229–291 (2010). [CrossRef]

27. S. Mohammadi, A. A. Eftekhar, A. Khelif, and A. Adibi, “Simultaneous two-dimensional phononic and photonic band gaps in opto-mechanical crystal slabs,” Opt. Express 18(9), 9164–9172 (2010). [CrossRef] [PubMed]

28. V. Laude, J. C. Beugnot, S. Benchabane, Y. Pennec, B. Djafari-Rouhani, N. Papanikolaou, and A. Martinez, “Design of waveguides in silicon phoxonic crystal slabs,” in 2010 IEEE International Ultrasonics Symposium(2010), pp. 527–530. [CrossRef]

29. Q. Rolland, M. Oudich, S. El-Jallal, S. Dupont, Y. Pennec, J. Gazalet, J. C. Kastelik, G. Lévêque, and B. Djafari-Rouhani, “Acousto-optic couplings in two-dimensional phoxonic crystal cavities,” Appl. Phys. Lett. 101(6), 061109 (2012). [CrossRef]

30. S. El-Jallal, M. Oudich, Y. Pennec, B. Djafari-Rouhani, V. Laude, J.-C. Beugnot, A. Martínez, J. M. Escalante, and A. Makhoute, “Analysis of optomechanical coupling in two-dimensional square lattice phoxonic crystal slab cavities,” Phys. Rev. B 88(20), 205410 (2013). [CrossRef]

31. T. Søndergaard and K. H. Dridi, “Energy flow in photonic crystal waveguides,” Phys. Rev. B 61(23), 15688–15696 (2000). [CrossRef]

32. C. Wolff, M. J. Steel, B. J. Eggleton, and C. G. Poulton, “Stimulated Brillouin scattering in integrated photonic waveguides: Forces, scattering mechanisms, and coupled-mode analysis,” Phys. Rev. A 92(1), 013836 (2015). [CrossRef]

33. W. Qiu, P. T. Rakich, M. Soljacic, and Z. Wang, “Stimulated brillouin scattering in slow light waveguides,” arXiv:1210.0738 (2012).

34. C. J. Sarabalis, J. T. Hill, and A. H. Safavi-Naeini, “Guided acoustic and optical waves in silicon-on-insulator for Brillouin scattering and optomechanics,” APL Photonics 1(7), 071301 (2016). [CrossRef]

35. D. K. Biegelsen, “Photoelastic tensor of silicon and the volume dependence of the average gap,” Phys. Rev. Lett. 32(21), 1196–1199 (1974). [CrossRef]

36. S. G. Johnson, M. Ibanescu, M. A. Skorobogatiy, O. Weisberg, J. D. Joannopoulos, and Y. Fink, “Perturbation theory for Maxwell’s equations with shifting material boundaries,” Phys. Rev. E 65(6), 066611 (2002). [CrossRef] [PubMed]

37. J. Chan, A. H. Safavi-Naeini, J. T. Hill, S. Meenehan, and O. Painter, “Optimized optomechanical crystal cavity with acoustic radiation shield,” Appl. Phys. Lett. 101(8), 081115 (2012). [CrossRef]

38. S. Johnson and J. Joannopoulos, “Block-iterative frequency-domain methods for Maxwell’s equations in a planewave basis,” Opt. Express 8(3), 173–190 (2001). [CrossRef] [PubMed]

39. S. A. Chandorkar, M. Agarwal, R. Melamud, R. N. Candler, K. E. Goodson, and T. W. Kenny, “Limits of quality factor in bulk-mode micromechanical resonators,” in 2008 IEEE 21st International Conference on Micro Electro Mechanical Systems(2008), pp. 74–77. [CrossRef]

40. G. Chen, R. Zhang, J. Sun, H. Xie, Y. Gao, D. Feng, and H. Xiong, “Mode conversion based on forward stimulated Brillouin scattering in a hybrid phononic-photonic waveguide,” Opt. Express 22(26), 32060–32070 (2014). [CrossRef] [PubMed]

Giant enhancement of stimulated Brillouin scattering with engineered phoxonic crystal waveguides

Abstract

1. Introduction

2. Derivation of the SBS gain of a phoxonic crystal waveguide

3. Engineering the phoxonic crystal waveguide for obtaining simultaneous slow light and slow sound

4. Calculation of the SBS gain in the engineered phoxonic crystal waveguide

5. Conclusion

Funding

References and links

Cited By

Figures (5)

Equations (36)

Optics Express