Design of Graphene Electro-optic Modulator in Mid-infrared Band

Research Background

The graphene material known as one of two-dimensional materials has brilliant electrical and optical properties. The graphene mainly interacts with in-plane electromagnetic field and shows great electro absorption ability, which makes it potential for an optoelectronic modulator. For monolayer graphene, we can control electro absorption ability by applying different voltages on electrodes. Typically, for mid-infrared band, the device can be built on a silicon wafer with 300nm silicon dioxide on top. To further decrease the insertion loss, chalcogenide material is used as the waveguide. However, due to large confinement of the vertical direction of 2D material, graphene will have different absorption for TE mode and TM mode as they have different polarization, which rises up the importance of the polarization-independent modulator. Two types of structures have proposed in recent paper, which are tilted-shape structure and polarization diversity scheme respectively.

For the tilted-shape method, one advantage is it’s simple geometry structure. Small size devices are more favored in integrated chips. The above structure has been reported recently, in which tilted graphene layers contribute to the polarization independence. White part is the silicon dioxide and red part is silicon as the waveguide. The green layer is h-BN (hexagonal boron nitride), isolating two graphene on the top and on the bottom. Partially tilted structure provide a chance of absorption for both TE and TM polarization.  However, in that paper, no experimental methods are given and it will be very hard to fabricate tilted layers. Therefore, this method just gives some clues of simulation and is not easy to fabricate from my point of view.  

The second method, known as “polarization diversity scheme” provide an idea of transferring TM mode into TE mode and modulate TE mode in the upper branch, then transfer TE back to the TM mode to tune TM.  For TE mode, it can be directly tuned in the lower branch. This method has more practical meaning than the above method. Therefore, in our program, we use this structure to design our own polarization-independent modulator in the mid-infrared band to get a deeper understanding in the field of integrated optics.

Design of A Splitter

A curved bridge can be used as a splitter. The principal comes from the basic idea that when two adjacent waveguides are close to each other, two hybrid modes will exist with different effective index, and thus cause the phase change of the light. When the change is 180 degrees, the light will go from one waveguide to the other. Therefore, we called this type of structure a splitter. As this structure is previously used in splitting 1.55um electromagnetic waves, in order to make it suitable in 3um band (mid-infrared band), a chalcogenide material replaces the conventional silicon material as a waveguide. The structure shows below. The middle curved waveguide is like a bridge to direct waves which satisfy phase matching conditions to another waveguide. However, based on the phase change principle, many other splitters can also work.

Considering that TE mode has more in-plane components than TM, the splitter should make TM mode match the phase condition, and split from the incident port to the upper branch. The left picture and the right picture show energy transmission of  the “TE” wave and the “TM” wave respectively.

Design of A Rotator

A cross section of normal waveguides are rectangular, which is symmetric. However, if we use waveguides with an asymmetric cross section, two mixed modes will exist (a cross section is shown in the following pictures). These two fundamental modes can be regarded as two basic vectors with the same magnitude if TM/TE fraction is 1.  Again with two mixed modes. different mode refractive index will cause a phase change and control the polarization. 

The following picture shows the shape of a rotator and white part represents an etched area. When a “TM” electromagnetic wave comes from the left side, the output will be a “TE” electromagnetic wave when the distance is properly designed. 

Design of A Modulator

We use a chalcogenide material as a waveguide, with aluminum oxide isolation layers on it (hBN also works as an isolation layer). Then monolayer graphene is on the top as an electro-optic material. Its cross section shows on the left picture below. Also the performance of this modulator shows in the right picture below. “TM1 Image” represents the original ” TM” electric field attenuation, while ” TM2 Image” represents the “TM” electric field attenuation which comes from the “TE” electromagnetic wave. These two functions are nearly the same! This indicates that a polarization-independent modulator is almost done, and the remaining task is just to regenerate the “TE” electromagnetic wave from the TM.

To make the modulator easier to understand, we present simulation results about its electro absorption ability. The left shows the attenuation under 0.05eV chemical potential; the right shows the attenuation under 0.2eV chemical potential. These results are consistent with the right picture above.

The Overall Model and Simulation

According to the polarization diverse scheme, the overall model shows in the following part:

Now to verify whether this model works well, we set the input as mixed light and select a certain value of chemical potential to watch the transmission of energy in this overall model. Note: if you cannot open the video, please use the chrome browser on the computer.