研究目的
To present waveguide photonic crystal re?ectors on the InP-membrane-on-silicon (IMOS) platform, and a method to accurately measure the re?ectivity of those re?ectors.
研究成果
Photonic crystal re?ectors on the InP-membrane-on-Silicon (IMOS) platform are designed and fabricated, showing re?ection coef?cients of over 95% experimentally. We show simulation results and two methods for measuring the re?ection coef?cient. The small difference between the measured values is caused by systematic errors in both of the techniques. Using both techniques for a similar re?ector gives a reliable interval for the actual value of the re?ection coef?cient for that re?ector. Using the designed re?ectors, a 50 μm long Fabry-P′erot cavity with a quality factor of almost 16,000 is also demonstrated.
研究不足
The systematic error is made by neglecting the propagation loss in the cavity. Typically, the propagation loss in 400 nm wide waveguides is 5-10 dB/cm. For 700 nm wide waveguides, the loss was not measured, but is expected to be lower, due to lower scattering from sidewall roughness. Considering the length of the cavity that is used, it is reasonable to neglect the propagation loss. Consequently, the calculated re?ectivity from the resonance peaks will be lower than the actual re?ectivity of the re?ector.
1:Experimental Design and Method Selection:
The photonic crystal re?ectors are simulated using a three dimensional ?nite-difference time-domain (3D FDTD) software tool. For both designs shown in Fig. 2 the number of holes was varied in the simulations. The dimensions of the holes is kept similar to the values shown in Table I. The hole sizes for the two holes on both ends of the re?ector are reduced to 60% for the ?rst hole and 80% for the second hole, with respect to the standard hole size. This is done to create a smooth transition between the waveguide and the re?ector.
2:Sample Selection and Data Sources:
The photonic crystal re?ectors are designed on the IMOS platform. A brief summary of the fabrication process of the chip used to obtain the reported results is described in this section. Typically, we start by bonding an InP wafer to a silicon carrier wafer as shown in Fig. 5. Silicon dioxide is deposited on both wafers for improved adhesion, and benzocyclobutene (BCB) is spun on the silicon carrier wafer. In a wafer bonding tool, pressure and high temperature are applied, and the wafers are bonded together. After the bonding, the indium phosphide substrate and indium galium arsenide etch-stop layers are removed by selective wet-chemical etching. Remaining is the 300 nm thick InP membrane, as shown in Fig. 5(b).
3:Silicon dioxide is deposited on both wafers for improved adhesion, and benzocyclobutene (BCB) is spun on the silicon carrier wafer. In a wafer bonding tool, pressure and high temperature are applied, and the wafers are bonded together. After the bonding, the indium phosphide substrate and indium galium arsenide etch-stop layers are removed by selective wet-chemical etching. Remaining is the 300 nm thick InP membrane, as shown in Fig. 5(b).
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: After bonding, IMOS devices are fabricated in the membrane using electron beam lithography (EBL), to de?ne etched regions. A simpli?ed version of the process ?ow is shown in Fig. 7. In step (a), ?rst a layer of 50 nm silicon nitride is deposited on the indium phosphide membrane. ZEP photo-resist is spun on the wafer, and the exposure is done using electron beam lithography, as shown in Fig. 7(a). The pattern is then transferred to the silicon nitride mask using reactive ion etching, and the photoresist is removed, which is shown in Fig. 7(b). Finally, a different reactive ion etching process is used to transfer the patterns in the silicon nitride mask to the indium phosphide, after which the silicon nitride mask is removed. This results in patterned devices in the indium phosphide membrane, as shown in Fig. 7(c). A scanning electron microscope (SEM) image of the fabricated photonic crystal re?ector in a 700 nm wide waveguide is shown in Fig.
4:In step (a), ?rst a layer of 50 nm silicon nitride is deposited on the indium phosphide membrane. ZEP photo-resist is spun on the wafer, and the exposure is done using electron beam lithography, as shown in Fig. 7(a). The pattern is then transferred to the silicon nitride mask using reactive ion etching, and the photoresist is removed, which is shown in Fig. 7(b). Finally, a different reactive ion etching process is used to transfer the patterns in the silicon nitride mask to the indium phosphide, after which the silicon nitride mask is removed. This results in patterned devices in the indium phosphide membrane, as shown in Fig. 7(c). A scanning electron microscope (SEM) image of the fabricated photonic crystal re?ector in a 700 nm wide waveguide is shown in Fig. Experimental Procedures and Operational Workflow:
6.
4. Experimental Procedures and Operational Workflow: We present two methods for characterization of the photonic crystal re?ectors. For the ?rst method, the transmission over a wavelength range of around 100 nm through a Fabry-P′erot cavity is measured. From the spectral resonances, the re?ectivity can be derived, as well as quality factor, ?nesse and free spectral range (FSR) of the cavity as explained in Section V-A1. With the second method, the re?ection is measured directly, as explained in Section V-A
5:With the second method, the re?ection is measured directly, as explained in Section V-AData Analysis Methods:
2.
5. Data Analysis Methods: The transmission spectrum for this structure is measured, from which the FSR and full-width at half-maximum (FWHM) of the resonance peaks can be obtained [7]. The ?nesse F is then given by F = FSR/FWHM, and the re?ection coef?cient R is obtained by R = e^(-2π/F).
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容