6.3. III-V A LL -D IELECTRIC A CTIVE M ETASURFACES
6.3.1. Characterization of MQW Wafers
the fixed operating wavelength of λ1Tran ≈ 1563 nm as functions of the applied bias are plotted in Figs. 6.3g, h, respectively.
It can be observed that when changing the applied bias, modulation of the transmission amplitude, accompanied by a phase modulation coverage of ~270° is realized at the operating wavelength of λ1Tran ≈ 1563 nm.
As seen in this section, using Si-based active metasurfaces, we can achieve dielectric metasurface elements operating both in transmission and reflection modes (dual-mode operation) at two distinct operating wavelengths. By altering the bias applied between Si nanoantennas and an ITO layer, one can tune the amplitude and phase of the reflected/transmitted light, and accordingly, engineer the wavefront of the scattered light at will.
In order to characterize the DBR layer, we measure the reflectance spectrum of the planar MQW/DBR/GaAs structure as shown in Fig. 6.4a. As can be seen in Fig. 6.4a, a reflectance of ~ 100% is obtained at the wavelengths ranging from 915 nm to 990 nm, indicating that the DBR can act as a high-quality mirror in this wavelength range.
Moreover, a sharp reflectance dip can be observed at the wavelength of ~915 nm which originates from near-bandgap absorption in the MQW layer.
As a next step, we try to investigate the tunable optical response of the MQW under an applied bias. When applying a DC electric field across the quantum wells, one can expect enabling electrical modulation of the MQW complex refractive index [242] due to the shift of the interband transition energy by the QCSE. For our quantum well heterostructures, the expected modulation of the real part of the refractive index is on the order of Δn = 0.01 [255]. In order to be able to experimentally observe this small variation of the real part of the refractive index, we integrate a Fabry-Pérot (F-P) resonant cavity around the MQWs. The structure supporting the F-P resonance is obtained by depositing a 35 nm-thick semitransparent Au film as the top mirror on top of the MQWs. A 2 nm-thick Ti film is used to improve the adhesion of the Au to the top p-doped GaAs layer. Figure 6.4b plots the measured reflectance spectrum of the fabricated DBR/MQW/Au F-P cavity. As can be seen, the structure can exhibit a narrow resonance at a wavelength of 932.7 nm. This narrow resonance will be later used to enhance the optical modulation caused by the variation of the complex refractive index of the MQWs under applied bias.
In order to design the metasurface using the MQW heterostructure, we first need to identify the tunable optical response of the MQWs at different wavelengths by shifting the position of the F-P resonance to the desired spectral position. To this end, a number of planar DBR/MQW/polymethyl methacrylate (PMMA)/Au heterostructures (see the inset of Fig. 6.4c) that could support high-Q-factor F-P resonances are fabricated. By changing the thickness of the PMMA layer, and as a result, changing the cavity length, we could alter the spectral position of the high-Q resonances supported by these planar heterostructures. This would provide us a database of the change in the real (∆n) and imaginary (∆k) parts of the refractive index of the MQW at different wavelengths.
At a fixed PMMA thickness, and hence, a fixed unbiased resonant wavelength, the ∆n and ∆k values are evaluated by examining the shift of the resonant wavelength, and the change of the full width at half maximum (FWHM) of the resonance, respectively under
applied bias. Then, utilizing the same resonant mode (i.e. the first F-P cavity resonant mode), the ∆n and ∆k values are obtained at different wavelengths by changing the thickness of the PMMA layer in the DBR/MQW/PMMA/Au heterostructure. Figure 6.4c shows the measured reflectance spectra for different thicknesses of the PMMA layer.
Figure 6.4: Characterization of MQM wafers. Measured reflectance spectra of (a) a bare DBR/MQW and (b) a DBR/MQW/Ti/Au F-P cavity, and (c) DBR/MQW/PMMA/Au heterostructures for different thicknesses of the PMMA layer (t) under no applied bias. The insets show the schematics of corresponding structures. The shadowed region in (a) indicates the wavelength range shown in (b). (d) The measured reflectance spectrum of the DBR/MQW/Ti/Au F-P resonant cavity as a function of applied bias. (e) Measured wavelength shifts (black dots) and FWHM difference (blue dots) at the first F-P resonant mode of the DBR/MQW/PMMA/Au planar layers with different PMMA thicknesses under applied bias [117].
As can be seen, increasing the thickness of the PMMA layer would redshift the first F- P resonant mode. For PMMA thicknesses greater than 210 nm, the second and third F- P resonant modes start showing up.
Once we measured the reflectance spectra of the DBR/MQW/Au and the DBR/MQW/PMMA/Au planar heterostructures in the absence of the applied bias, we then measure their reflectance modulations under applied bias. To facilitate bias application, Ohmic contacts made of Ti (20 nm)/Pt (30 nm)/Au (300 nm) and Ge (43 nm)/Ni (30 nm)/Au (87 nm) are deposited on the topmost p-doped GaAs layer and at the back of the n-doped GaAs substrate, respectively. Then by applying a bias between the GaAs substrate (low potential) and the top Ohmic contact (high potential), the reflectance from our F-P resonant MQW samples is measured.
Figure 6.4d plots the measured reflectance spectrum of the DBR/MQW/Au planar layer as a function of applied bias. As can be seen, when applying an external bias, a shift of the resonant wavelength accompanied by a significant reflectance modulation is observed. It confirms the modulation of both the real and imaginary parts of the MQW refractive index by the applied bias. It should also be noted that a larger optical modulation is observed at shorter wavelengths, near the semiconductor band edge.
These measurements also show that the optimal wavelength for large reflectance modulation is expected to be between 915 nm and 920 nm.
The same measurements are performed to obtain the reflectance spectra of the planar DBR/MQW/PMMA/Au heterostructures with different PMMA thicknesses, for different applied biases. These measurements also show stronger amplitude modulation and larger wavelength shifts at shorter wavelengths. Figure 6.4e presents the measured bias-induced wavelength shift and variation of the FWHM of the first F-P resonance supported by the planar DBR/MQW/PMMA/Au heterostructure. Here, the wavelength shift ∆λ = λ (V = –10 V) – λ (V = 0), and the FWHM difference FWHM (V = –10 V) – FWHM (V = 0) are defined as an amount of the spectral shift in the resonance position and the change in FWHM when the bias is changed from 0 V to –10 V.
It can also be seen in Fig. 6.4e that larger FWHM changes are obtained at shorter wavelengths. These results are consistent with the analysis described in prior work [255], which reported the III-V compound MQW design used in our work. Based on the trend shown in Fig. 6.4e, we can conclude that the strongest refractive index
modulation can be obtained at wavelengths very close to the absorption band edge of our quantum wells, which is expected to be at the wavelength of ~915 nm.