Vat photopolymerization's (VP) use in large-scale industrial manufacturing is limited due to poor scalability, and limited catalogue of engineering polymers. The challenges in scalability stem from an inherent process paradox: the feature resolution, part size, and manufacturing throughput cannot be maximized simultaneously in standard VP platforms. In addition, VP's inability to process viscous and high-molecular weight engineering polymers limits the VP materials catalogue. To address these limitations, the research presented in this work was conducted in two stages: (1) Development and modeling of new VP platforms to address the scalability and viscosity challenges, and (2) Investigating the influence of using the new processes on the cured polymer network structure and mechanical properties. First, a scanning mask projection vat photopolymerization (S-MPVP) system was developed to address the scalability limitations in VP systems. The process paradox was resolved by scanning the mask projection device across the resin surface while simultaneously projecting the layer as a movie. Using actual projected pixel irradiance distribution, a process model was developed to capture the interaction between projected pixels and the resin, and predict the resulting cure profile with an error of 2.9%. The S-MPVP model was then extended for processing heterogeneous UV scattering resins (i.e. UV curable polymer colloids). Using computer vision, the scattering of incident UV radiation on the resin surface was successfully captured and used to predict scattering-compensated printing parameters (bitmap pattern, exposure time , scanning speed). The developed reverse-curing model was used to successfully fabricate complex features using photocurable SBR latex with XY errors < 1.3%. To address the low manufacturing throughput of VP systems, a recoat-less, volumetric curing VP system that fabricates parts by continuously irradiating the resin surface with a movie composed of different gray-scaled bitmap images ( Free-surface movie mask projection (FreeMMaP)) was developed. The effect of cumulative exposure on the cure profile (X,Y,Z dimensions) was investigated and used to develop an iterative gray-scaling algorithm that generated a combination of gray-scaled bitmap images and exposure times that result in accurate volumetric curing (errors in XY plane and Z axis < 5% and 3% respectively). Results of this work demonstrate that the elimination of the recoating process increased manufacturing speed by 8.05 times and enabled high-resolution fabrication with highly viscous resins or soft gels. Then, highly viscous resins were made processible in VP systems by using elevated processing temperatures to lower resin viscosity. New characterization techniques were developed to determine the threshold printing temperature and time that prevented the onset of thermally-induced polymerization. The effect of printing temperature on curing, cured polymer structure, cured polymer mechanical properties, and printable aspect ratio was also investigated using diacrylate and dimethacrylate resins. Results of this investigation revealed increasing printing temperature resulted in improvements in crosslink density, tensile strength, and printability. However, presence of hydroxl groups on the resin backbone caused deterioration of crosslink density, mechanical properties, and curing properties at elevated printing temperatures. Finally, the lack of a systematic, constraint based approach to resin design was bridged by using the results of earlier process-structure-property explorations to create an intuitive framework for resin screening and design. Key screening parameters (such as UV absorptivity, plateau storage modulus) and design parameters (such as photoinitiator concentration, polymer concentration, UV blocker concentration) were identified and the methods to optimize them to meet the desired printability metrics were demonstrated using case studies. Most work in vat photopolymerization either deal with materials development or process development and modeling. This dissertation is placed at the intersection of process development and materials development, thus giving it an unique perspective for exploring the inter-dependency of machine and material. The process models, machines and techniques used in this work to make a material printable will serve as a guide for chemists and engineers working on the next generation of vat photopolymerization machines and materials. / Doctor of Philosophy / Vat Photopolymerization (VP) is a polymer-based additive manufacturing platform that uses UV light to cure a photo-sensitive polymer into the desired shape. While parts fabricated via VP exhibit excellent surface finish and high-feature resolution, their use for commercial manufacturing is limited because of its poor scalability for large-scale manufacturing and limited selection of engineering materials. This work focuses on the development of new VP platforms, process models and the investigation of the process-structure-property relationships to mitigate these limitations and enable fabrication of performance polymers. The first section of the dissertation presents the development of two new VP platforms to address the limitations in scalability. The Scanning Mask Projection Vat Photopolymerization (S-MPVP)) was developed to fabricate large area parts with high-resolution features and the Free-surface movie mask projection (FreeMMaP) VP platform was developed to enable high-speed, recoat-less, volumetric fabrication of 3D objects. Computer-vision based models were developed to investigate the influence of these new processes on the resultant cure shape and dimensional accuracy. Process models that can: (1) predict the cure profile for given input printing parameters (error < 3%), (2) predict the printing parameters (exposure time, bitmap gray-scaling) required for accurate part fabrication in homogeneous and UV scattering resins, and (3) generate gray-scaled bitmap images that can induce volumetric curing inside the resin (dimensional accuracy of 97% Z axis, 95% XY axis), were designed and demonstrated successfully. In the second portion of this work, the use of high-temperature VP to enable processing of high-viscosity resins and expansion of materials catalogue is presented. New methods to characterize the resin's thermal stability are developed. Techniques to determine the printing temperature and time that will prevent the occurrence of thermally-induced polymerization are demonstrated. Parts were fabricated at different printing temperatures and the influence of printing temperature on the resultant mechanical properties and polymer network structure was studied. Results of this work indicate that elevated printing temperature can be used to alter the final mechanical properties of the printed part and improve the printability of the high-resolution, slender features. Finally, the results of the process-structure-property investigations conducted in this work were used to guide the development of a resin design framework that highlights the parameters, metrics, and methods required to (1) identify printable resin formulations, and (2) tune printable formulations for optimal photocuring. Elements of this framework were then combined into an intuitive flowchart to serve as a design tool for chemists and engineers.

The goal of this work was to design and build a mask projection microstereolithography (MPμSL) 3D printing system to characterize, process, and quantify the performance of novel photopolymers. MPμSL is an Additive Manufacturing process that uses DLP technology to digitally pattern UV light and selectively cure entire layers of photopolymer resin and fabricate a three dimensional part. For the MPμSL system designed in this body of work, a process was defined to introduce novel photopolymers and characterize their performance. The characterization process first determines the curing characteristics of the photopolymer, namely the Critical Exposure (Ec) and Depth of Penetration (Dp). Performance of the photopolymer is identified via the fabrication of a benchmark test part, designed to determine the minimum feature size, XY plane accuracy, Z-axis minimum feature size, and Z-axis accuracy of each photopolymer with the system. The first characterized photopolymer was poly (propylene glycol) diacrylate, which was used to benchmark the designed MPμSL system. This included the achievable XY resolution (212 micrometers), minimum layer thickness (20 micrometers), vertical build rate (360 layers/hr), and maximum build volume (6x8x36mm3). This system benchmarking process revealed two areas of underperformance when compared to systems of similar design, which lead to the development of the first two research questions: (i) 'How does minimum feature size vary with exposure energy?' and (ii) 'How does Z-axis accuracy vary with increasing Tinuvin 400 concentration in the prepolymer?' The experiment for research question (i) revealed that achievable feature size decreases by 67% with a 420% increase in exposure energy. Introducing 0.25wt% of the photo-inhibitor Tinuvin 400 demonstrated depth of penetration reduction from 398.5 micrometers to 119.7 micrometers. This corresponds to a decrease in Z-axis error from 119% (no Tinuvin 400) to 9% Z-axis error (0.25% Tinuvin 400). Two novel photopolymers were introduced to the system and characterized. Research question (iii) asks 'What are the curing characteristics of Pluronic L-31 how does it perform in the MPμSL system?' while Research Question 4 similarly queries 'What are the curing characteristics of Phosphonium Ionic Liquid and how does it perform in the MPμSL system?' The Pluronic L-31 with 2wt% photo-initiator had an Ec of 17.2 mJ/cm2 and a Dp of 288.8 micrometers, with a minimum feature size of 57.3 ± 5.7 micrometers, with XY plane error of 6% and a Z-axis error of 83%. Phosphonium Ionic Liquid was mixed in various concentrations into two base polymers, Butyl Diacrylate (0% PIL and 10% PIL) and Poly Ethylene Dimethacrylate (5% PIL, 15% PIL, 25% PIL). Introducing PIL into either base polymer caused the Ec to increase in all samples, while there is no significant trend between increasing concentrations of IL in either PEGDMA or BDA and depth of penetration. Any trends previously identified between penetration depth and Z accuracy do not seem to extend from one resin to another. This means that overall, among all resins, depth of penetration is not an accurate way to predict the Z axis accuracy of a part. Furthermore, increasing concentrations of PIL caused increasing % error in both XY plane and Z-axis accuracy . / Master of Science

Mask Projection micro Stereolithography (MPuSLA) is an additive manufacturing process used to build physical components out of a photopolymer resin. Existing MPuSLA technology cut the CAD model of part into slices by horizontal planes and the slices are stored as bitmaps. A layer corresponding to the shape of each bitmap gets cured. This layer is coated with a fresh layer of resin by lowering the Z-stage inside a vat holding the resin and the next layer is cured on top of it. In our Thick-film MPuSLA(TfMPuSLA) system, incident radiation, patterned by a dynamic mask, passes through a fixed transparent substrate to cure photopolymer resin. The existing MPuSLA fabrication models can work only for controlling the lateral dimension, without any control over the thickness of the cured part. The proposed process plan controls both the lateral dimensions and the thickness of profile of the cured part. In this thesis, a novel process planning for TfMPuSLA method is developed, to fabricate films on fixed flat substrate. The process of curing a part using this system is analytically modeled as the column cure model. It is different from the conventional process - layer cure model. Column means that a CAD model of part is discretized into vertical columns instead of being sliced into horizontal layers, and all columns get cured simultaneously till the desired heights. The process planning system is modularized into geometrical, chemical, optical, mathematical and physical modules and validated by curing test parts on our system. The thesis formulates a feasible process planning method, providing a strong basis for continued investigation of MPuSLA technology in microfabrication, such as micro lens fabrication.

Ceramic manufacturing is an expensive process with long lead times between the initial design and final manufactured part. This limits the use of ceramic as a viable material unless there is a large project budget or high production volume associated with the part. Ceramic stereolithography is an alternative to producing low cost parts through the mixing of a photo curable resin and ceramic particles. This is an additive manufacturing process in which each layer is built upon the previous to produce a green body that can be sintered for a fully dense ceramic part. This thesis introduces a new approach to ceramic stereolithography with a top mask projection light source which is much more economical compared to current vector scanning methods. The research goes through the design and development of a stereolithography printer prototype capable of handling ceramics and the testing of different mixtures to provide the best printing results with varying viscosities. The initial testing of this printer has created a starting point for top mask projection as an economical alternative to current ceramic manufacturing techniques.