Next Generation Graphene Photonics Enabled by Ultrafast Light-Matter Interactions and Machine Learning

Graphene was first experimentally studied in 2004, featuring an atomically-thin structure. Since then, many unique photonic and electrical properties of graphene and other 2D materials were reported. However, additional efforts are necessary to convert these findings in physics to successful industrial applications. This thesis presents works exploiting the picosecond-scale ultrafast light-matter interactions in graphene to meet the growing demands in IR sensing, 3D detection, and THz light source. We will start from graphene’s interactions with ultrafast lasers. The hot carrier generation, relaxation, and transport will be discussed in graphene and graphene heterostructures. We present a graphene phototransistor with decent near- and mid-infrared (IR) responsivity. Moreover, the detector’s responsivity is tunable with a gate voltage. The responsivity has different gate dependence under different illumination wavelengths. Based on the spectrally-resolved response, we adopt least square regression algorithms to extract the light source’s spectral information at near-infrared. We further perform first-principle photocurrent simulations and spectral reconstructions on defect-free ideal devices with optimized band structure. The results indicate the detector's potential as an ultra-compact on-chip spectrometer for multispectral imaging after further developments. Then we discuss how the graphene detector’s high transparency enables a novel 3D detection and imaging technology. Our graphene phototransistors absorb < 10% of light and give a 3 A/W photoresponse at 532 nm wavelength. The high transparency and sensitivity enable transparent photodetector arrays built on glass substrates, with over 85% of incident light power transmits through such an imager chip. We stack multiple transparent arrays at different focal depths in a camera system. The setup enables simultaneous light intensity (image) acquisition at different depths. We use artificial neural networks to process the image stack data into 3D position and configuration of the objects. For a proof-of-concept demonstration, we used the setup to achieve 3D ranging and tracking of a point source. The technical approach benefits from compactness, high speed, and decent power efficiency for real-time 3D tracking applications. Lastly, we explore the potential of graphene heterostructures as terahertz (THz) emitters and ultrafast photodetectors. The picosecond-scale light-matter interaction of graphene allows us to engineer its optical and electrical structures for THz field emission. We insert a graphene layer in the channel of a silicon photoconductive switch. The device works as a THz electromagnetic wave emitter under femtosecond laser pulse illumination. We use an on-chip pump-probe system to study the temporal and spatial behavior of the THz generation. Our device’s emission amplitude is 80 times larger than a graphene-free control group under identical device geometry and test conditions. Moreover, we also observe strong photocurrent generation below 0.5 ps verified by the photocurrent autocorrelation test. The responsivity is 800 times larger than that in the graphene-free control group. The substantial enhancements are attributed to the high mobility in graphene and the strong absorption in silicon. Gate dependence observations indicate vertical hot-carrier transfer from the silicon layer to the graphene layer, followed by efficient lateral charge separation inside graphene. The results open the gate for more research and development of graphene-based strong THz sources and sensitive ultrafast photodetectors. We conclude the works with strategies to convert graphene’s unique properties to practical and competitive applications. The strategies are extended to general nanodevice and nano-system development methodologies. Specifically, we propose the synergic design of nanodevices and machine learning algorithms as a feasible approach towards many new applications.PHDElectrical and Computer EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/169683/1/dehui_1.pd