Additive and Fiber Nano-manufacturing of Multi-material, Dielectric Photonic Crystals

Abstract

Next-generation manufacturing systems will be largely autonomous, relying on in-situ sample measurements and intelligent feedback to create the desired components. This dissertation takes critical first steps toward realizing this vision in the domain of nano-photonic structures. One- and two-dimensional photonic structures are designed and fabricated using additive processes where multi-functional optimization of material selection and process conditions is performed to negotiate trade-offs presented by the simultaneous need to maximize optical performance and manufacturability. Physics-guided design that accounts for both optical and rheological properties of constituent materials and in situ measurements of the manufactured component’s optical properties are combined to guide on-the-fly adjustment of manufacturing parameters via intelligent reinforcement learning algorithms; thus, moving toward a novel, autonomous manufacturing paradigm. Photonic crystals (PC) are ubiquitous in nature, arising from wavelength-selective refraction by structural arrangements of matter at the micro and nanoscale. Engineered PCs allow precise reflection, transmission, and absorption properties enabling their increasing use in many applications, from biomedical imaging to fiber optic communications. Today, most PC fabrication is largely limited to “subtractive” manufacturing techniques, which are mature, yet generally capital intensive and relatively slow for prototyping. This dissertation considers a “bottom-up” approach to creating PCs, using micro- and nano-scale additive manufacturing (AM) and fiber drawing (FD). First, electromagnetic simulations are used to guide material selection depending on the manufacturing technique. Dielectric inks with moderate refractive index contrast (Δn=0.2) and low parasitic absorption (k~0) in the visible to infrared regions (0.4-5 µm) are employed to create multilayered and multi-material one-dimensional PC (1DPCs) “pixels” using drop-on-demand electrohydrodynamic jet (e-jet) printing. While optical design calls for maximizing refractive index contrast, low-index photopolymer inks tend to present a low-energy surface that results in poor spreading when coated by other inks, and vice-versa. Favorable combinations of ink surface tension and substrate surface energy are engineered to promote uniform spreading; the trade-off between increasing the layer number and the resulting increase in roughness is explored. Building on the understanding of the statics and dynamics of surface wetting, e-jet printing is again employed to create two-dimensional arrays of constituent meta-atoms (2DPCs). Colloidal inks with unusually high (>75 wt%) loading of high refractive index (n~4) germanium nanoparticles and fluorinated, self-assembled surface coatings are selected to maximize droplet contact angle, and thus aspect ratio, in the meta-atoms. After negotiating processing challenges for these inks, printed meta-atoms are obtained with strong photoluminescence in the visible regime (~650 nm), while 2D arrays demonstrated infrared photonic response (~4 µm). The colloidal ink nano-printing approach simplifies fabrication and opens a path to, for example, enabling pathogen detection through surface binding. Returning to the 1DPC concept, novel, multi-material, thermally drawn photonic fibers are fabricated with a strong infrared (1-5.5 µm) reflectance response for an all-polymer (Δn=0.1) layer composition, opening the door to numerous applications not feasible using previously employed materials. Thermal and rheological limitations of the constituent polymers were negotiated to achieve high reflectivity in the infrared range through drawing N > 100 distinct, nanoscale polymer layers while maintaining continuity and uniformity. The layer structure was further engineered to have a unique photonic “barcode” using overtones, achieving a high signal-to-noise ratio versus the background woven fabric. These fibers may find applications in life cycle textile tracing and sorting and for item authentication. Finally, future applications and extensions of the present work are discussed.