No description available.

The reactivity of chemical warfare agents (CWAs) and CWA simulants on organic and oxide surfaces is not currently well understood, but is of substantial importance to the development of effective sensors, filters and sorbent materials. Polyurethane coatings are used by the armed forces as chemical agent resistive paints to limit the uptake of CWAs on surfaces, while the use of metal oxides has been explored for decontamination and protection purposes. To better understand the chemical nature of the interactions of organophosphonate simulants with these surfaces, an ultra-high vacuum environment was used to isolate the target interactions from environmental gaseous interferences. The use of highly-characterized surfaces, coupled with molecular beam and dosing capabilities, allows for the elucidation of adsorption, desorption, and reaction mechanisms of CWA simulants on a variety of materials. Model urethane-containing organic coatings were designed and applied toward the creation of well-ordered thin films containing carbamate linkages. In addition, novel trisilanolphenyl-polyhedral oligomeric silsesquioxane (POSS) molecules were used to create Langmuir-Blodgett films containing reactive silanol groups that have potential use as sensors and coatings. The uptake and reactivity of organophosphonates and chlorophosphates on these surfaces is the focus of this study. Surfaces were characterized before and after exposure to the phosphates using a number of surface sensitive techniques including: contact angle goniometry, reflection-absorption infrared spectroscopy (RAIRS), X-ray photoelectron spectroscopy (XPS) and temperature-programmed desorption (TPD) measurements. In conjunction with surface probes, uptake coefficients were monitored according to the King and Wells direct reflection technique. The integration of these analytical techniques provides insight and direction towards the design of more effective chemical agent resistant coatings and aids in the development of more functional strategies for chemical warfare agent decontamination and sensing. / Ph. D.

Single wavelength ellipsometry measurements at Brewster's angle provide a powerful technique for characterizing ultrathin polymeric films. By conducting the experiments in different ambient media, multiple incident media (MIM) ellipsometry, simultaneous determinations of a film's thickness and refractive index are possible. Poly(tert-butyl acrylate) (PtBA) films serve as a model system for the simultaneous determination of thickness and refractive index (1.45 at 632 nm). Thickness measurements on films of variable thickness agree with X-ray reflectivity results. The method is also applicable to spincoated films where refractive indices of PtBA, polystyrene and poly(methyl methacrylate) are found to agree with literature values within experimental error. Likewise, MIM ellipsometry is utilized to simultaneously obtain the refractive indices and thicknesses of thin films of trimethylsilylcellulose (TMSC), regenerated cellulose, and cellulose nanocrystals where Langmuir-Blodgett (LB) films of TMSC serve as a model system. Ellipsometry measurements not only provide thickness and optical constants of thin films, but can also detect thermally induced structural changes like surface glass transition temperatures (Tg) and layer deformation in LB-films. Understanding the thermal properties of the polymer thin films is crucial for designing nanoscale coatings, where thermal properties are expected to differ from their corresponding bulk properties because of greater fractional free volume in thin films and residual stresses that remain from film preparation. Polyhedral oligomeric silsesquioxane (POSS) derivatives may be useful as a nanofiller in nanocomposite formulations to enhance thermal properties. As a model system, thin films of trisilanolphenyl-POSS (TPP) and two different molar mass PtBA were prepared as blends by Y-type Langmuir-Blodgett film deposition. For comparison, bulk blends were prepared by solution casting and the samples were characterized via differential scanning calorimetry (DSC). Our observations show that surface Tg is depressed relative to bulk Tg and that magnitude of depression is molar mass dependent for pure PtBA films. By adding TPP as a nanofiller both bulk and surface Tg increase. Whereas, bulk Tg shows comparable increases for both molar masses, the increase in surface Tg for higher molar mass PtBA is greater than for lower molar mass PtBA. These studies show that POSS can serve as a model nanofiller for controlling Tg in nanoscale coatings. / Ph. D.

Thin film bilayers and blends composed of polymers and nanoparticles are increasingly important for technological applications that range from space survivable coatings to novel drug delivery systems. Dewetting or spontaneous hole formation in amorphous polymer films and phase separation in multicomponent polymer films can hinder the stability of these systems at elevated temperatures. Hence, fundamental understanding of dewetting and phase separation in polymer/nanoparticle bilayer and blend films is crucial for controlling transport and thermomechanical properties and surface morphologies of these systems. This dissertation provides studies on morphological evolution driven by phase separation and/or dewetting in model polymer/nanoparticle thin film bilayers and blends at elevated temperatures. Morphological evolution in dewetting bilayers of poly(t-butyl acrylate) (PtBA) or polystyrene (PS) and a polyhedral oligomeric silsesquioxane (POSS), trisilanolphenyl-POSS (TPP) is explored at elevated temperatures. The results demonstrate unique dewetting morphologies in both PtBA/TPP and PS/TPP bilayers that are significantly different from those typically observed in dewetting polymer/polymer bilayers. Upon annealing the PtBA/TPP bilayers at 95°C, a two-step dewetting process is observed. PtBA immediately diffuses into the upper TPP layer leading to hole formation and subsequently the holes merge to form interconnected rim structures in the upper TPP layer. Dewetting of both the TPP and PtBA layers at longer annealing times leads to the evolution of scattered holes containing TPP-rich, fractal aggregates. The fractal dimensions of the TPP-rich, fractal aggregates are ~2.2 suggesting fractal pattern formation via cluster-cluster aggregation. Dewetting in PS/TPP bilayers also proceeds via a two-step process; however, the observed dewetting morphologies are dramatically different from those observed in PtBA/TPP bilayers. Cracks immediately form in the upper TPP layer during annealing of PS/TPP bilayers at 200°C. With increasing annealing times, the cracks in the TPP layer act as nucleation sites for dewetting and aggregation of the TPP layer and subsequent dewetting of the underlying PS layer. Complete dewetting of both the TPP and PS layers results in the formation of TPP encapsulated PS droplets. Phase separation in PtBA/TPP thin film blends is investigated as functions of annealing temperature and time. The PtBA/TPP thin film blend system exhibits an upper critical solution temperature (LCST) phase diagram with a critical composition and temperature of 60 wt% PtBA and ~70°C, respectively. Spinodal decomposition (SD) is observed for 60 wt% PtBA blend films and off-critical SD is seen for 58 and 62 wt% PtBA blend films. The temporal evolution of SD in 60 wt% PtBA blend films is also explored. Power law scaling for the characteristic wavevector with time (q ~ t^n with n = -1/4 to -1/3) during the early stages of phase separation yields to domain pinning at the later stages for films annealed at 75, 85, and 95°C. In contrast, domain growth is instantly pinned for films annealed at 105°C. Our work provides an important first step towards understanding how nanoparticles affect polymer thin film stability and this knowledge may be utilized to fabricate surfaces with tunable morphologies via controlled dewetting and/or phase separation. / Ph. D.

For over a century, Langmuir films have served as excellent two-dimensional model systems for studying the conformation and ordering of amphiphilic molecules at the air/water (A/W) interface. With the equipment of Wilhelmy plate technique, Brewster angle microscopy (BAM), and surface light scattering (SLS), the interfacial phase and rheological behavior of Langmuir films can be investigated. In this dissertation, these techniques are employed to examine Langmuir films of polyhedral oligomeric silsesquioxane (POSS), polymer blends, and magnetic nanoparticles (MNPs). In a first time, SLS is employed to study POSS molecules. The interfacial rheological properties of trisilanolisobutyl-POSS (TiBuP) indicate that TiBuP forms a viscoelastic Langmuir film that is almost perfectly elastic in the monolayer state with a maximum dynamic dilational elasticity of around 50 mNâ m-1 prior to film collapse. This result suggests that TiBuP can serve as model nanofiller with polymers. As an interesting next step, blends of TiBuP and polydimethylsiloxane (PDMS) with different compositions are examined via surface pressure (surface pressureâ surface area occupied per molecule (A) isotherms and SLS. The results show that TiBuP, with its attendant water, serves as a plasticizer and lowers the dilational modulus of the films at low surface pressure. As surface pressure increases, composition dependent behavior occurs. Around the collapse pressure of PDMS, the TiBuP component is able to form networks at the A/W interface as PDMS collapse into the upper layer. Blends of non-amphiphilic octaisobutyl-POSS (OiBuP) and PDMS are also studied as an interesting comparison to TiBuP/PDMS blends. In these blends, OiBuP serves as a filler and reinforces the blends prior to the collapse of PDMS by forming "bridge" structure on top of PDMS monolayer. However, OiBuP is non-amphiphilic and fails to anchor PDMS chains to the A/W interface. Hence, OiBuP/PDMS blends exhibit negligible dilational viscoelasticity after the collapse of PDMS. Furthermore, the phase behavior of PDMS blended with a trisilanol-POSS derivative containing different substituents, trisilanolcyclopentyl-POSS (TCpP), is also investigated via the Wilhelmy plate technique and BAM. These TCpP/PDMS blends exhibit dramatically different phase behavior and morphological features from previously studied POSS/PDMS blends, showing that the organic substituents on trisilanol-POSS have considerable impact on the phase behavior of POSS/PDMS blends. The interfacial rheological behavior of tricarboxylic acid terminated PDMS (PDMS-Stabilizer) and PDMS stabilized MNPs are investigated and compared with "regular" PDMS containing non-polar end groups. The tricarboxylic acid end group of the PDMS-Stabilizer leads to a different collapse mechanism. The PDMS stabilized MNPs exhibit viscoelastic behavior that is similar to PDMS showing all the tricarboxylic acid end groups are bound to the magnetite cores. Studying the interfacial behavior of different Langmuir films at the A/W interface provides us insight into the impact of molecule-molecule and molecule-subphase interactions on film morphology and rheology. These results are able to serve as important guides for designing surface films with preferred morphological and mechanical properties. / Ph. D.

Langmuir-monolayer formation at the air/water interface (A/W) can be achieved by spreading amphiphilic molecules on a liquid subphase and compressing them into an ordered arrangement. The use of the Langmuir-Blodgett technique (LB) to prepare ultra thin films on solid surfaces from monolayers at A/W has considerable utility for studying surface interactions. In particular, the phase behavior of polyhedral oligomeric silsesquioxanes (POSS) was examined using a combination of LB and Brewster angle microscopy (BAM).Polymer fillers have been shown to reduce the cost and often improve the properties of high performance polymer composites. The utility of POSS as a potential nanofiller in blends with polymers such as poly(dimethylsiloxane) (PDMS) and poly(vinylacetate) (PVAc) was explored using surface pressure-area per monomer isotherms (P-A) and BAM. Substantial morphological differences are seen between polymer blends with heptasubstituted trisilanol-POSS and fully condensed octasubstituted-POSS due to differences in subphase affinity.Several poly(1-alkene-alt-maleic anhydride) (PXcMA) polymers were studied at both the gas/liquid interface as Langmuir films and at the gas/solid interface as Langmuir-Blodgett thin films on silicon substrates. A 0.01 M HCl solution (pH~2) was used during film deposition to ensure the carboxylic acids were fully protonated. The PXcMA polymers included X=1-hexene, 1-octene, 1-decene, and 1-octadecene (represented as PHcMA, POcMA, PDcMA, and PODcMA respectively). The P-A isotherms of these polymers were consistent with those obtained previously.1Tensiometry was used to determine the critical micelle concentrations (c.m.c.) of variable molar mass poly(dimethylsiloxane-b-(3-cyanopropyl)methylsiloxane-b-dimethylsiloxane) (PDMS-PCPMS-PDMS) triblock copolymers and a poly(dimethylsiloxane-b-2-ethyl-2-oxazoline) diblock copolymer. Dynamic light scattering (DLS) corroborated interfacial tension results. The polymers exhibited well-defined temperature-independent c.m.c.'s. These measurements ensured that the synthesis of cobalt nanoparticles for biocompatible magnetic fluids occurred above the c.m.c. / Master of Science

Supervised learning of piecewise linear models is a well studied problem in machine learning community. The key idea in piecewise linear modeling is to properly partition the input space and learn a linear model for every partition. Decision trees and regression trees are classic examples of piecewise linear models for classification and regression problems. The existing approaches for learning decision/regression trees can be broadly classified in to two classes, namely, fixed structure approaches and greedy approaches. In the fixed structure approaches, tree structure is fixed before hand by fixing the number of non leaf nodes, height of the tree and paths from root node to every leaf node of the tree. Mixture of experts and hierarchical mixture of experts are examples of fixed structure approaches for learning piecewise linear models. Parameters of the models are found using, e.g., maximum likelihood estimation, for which expectation maximization(EM) algorithm can be used. Fixed structure piecewise linear models can also be learnt using risk minimization under an appropriate loss function. Learning an optimal decision tree using fixed structure approach is a hard problem. Constructing an optimal binary decision tree is known to be NP Complete. On the other hand, greedy approaches do not assume any parametric form or any fixed structure for the decision tree classifier. Most of the greedy approaches learn tree structured piecewise linear models in a top down fashion. These are built by binary or multi-way recursive partitioning of the input space. The main issues in top down decision tree induction is to choose an appropriate objective function to rate the split rules. The objective function should be easy to optimize. Top-down decision trees are easy to implement and understand, but there are no optimality guarantees due to their greedy nature. Regression trees are built in the similar way as decision trees. In regression trees, every leaf node is associated with a linear regression function. All piece wise linear modeling techniques deal with two main tasks, namely, partitioning of the input space and learning a linear model for every partition. However, Partitioning of the input space and learning linear models for different partitions are not independent problems. Simultaneous optimal estimation of partitions and learning linear models for every partition, is a combinatorial problem and hence computationally hard. However, piecewise linear models provide better insights in to the classification or regression problem by giving explicit representation of the structure in the data. The information captured by piecewise linear models can be summarized in terms of simple rules, so that, they can be used to analyze the properties of the domain from which the data originates. These properties make piecewise linear models, like decision trees and regression trees, extremely useful in many data mining applications and place them among top data mining algorithms. In this thesis, we address the problem of supervised learning of piecewise linear models for classification and regression. We propose novel algorithms for learning piecewise linear classifiers and regression functions. We also address the problem of noise tolerant learning of classifiers in presence of label noise. We propose a novel algorithm for learning polyhedral classifiers which are the simplest form of piecewise linear classifiers. Polyhedral classifiers are useful when points of positive class fall inside a convex region and all the negative class points are distributed outside the convex region. Then the region of positive class can be well approximated by a simple polyhedral set. The key challenge in optimally learning a fixed structure polyhedral classifier is to identify sub problems, where each sub problem is a linear classification problem. This is a hard problem and identifying polyhedral separability is known to be NP complete. The goal of any polyhedral learning algorithm is to efficiently handle underlying combinatorial problem while achieving good classification accuracy. Existing methods for learning a fixed structure polyhedral classifier are based on solving non convex constrained optimization problems. These approaches do not efficiently handle the combinatorial aspect of the problem and are computationally expensive. We propose a method of model based estimation of posterior class probability to learn polyhedral classifiers. We solve an unconstrained optimization problem using a simple two step algorithm (similar to EM algorithm) to find the model parameters. To the best of our knowledge, this is the first attempt to form an unconstrained optimization problem for learning polyhedral classifiers. We then modify our algorithm to find the number of required hyperplanes also automatically. We experimentally show that our approach is better than the existing polyhedral learning algorithms in terms of training time, performance and the complexity. Most often, class conditional densities are multimodal. In such cases, each class region may be represented as a union of polyhedral regions and hence a single polyhedral classifier is not sufficient. To handle such situation, a generic decision tree is required. Learning optimal fixed structure decision tree is a computationally hard problem. On the other hand, top-down decision trees have no optimality guarantees due to the greedy nature. However, top-down decision tree approaches are widely used as they are versatile and easy to implement. Most of the existing top-down decision tree algorithms (CART,OC1,C4.5, etc.) use impurity measures to assess the goodness of hyper planes at each node of the tree. These measures do not properly capture the geometric structures in the data. We propose a novel decision tree algorithm that ,at each node, selects hyperplanes based on an objective function which takes into consideration geometric structure of the class regions. The resulting optimization problem turns out to be a generalized eigen value problem and hence is efficiently solved. We show through empirical studies that our approach leads to smaller size trees and better performance compared to other top-down decision tree approaches. We also provide some theoretical justification for the proposed method of learning decision trees. Piecewise linear regression is similar to the corresponding classification problem. For example, in regression trees, each leaf node is associated with a linear regression model. Thus the problem is once again that of (simultaneous) estimation of optimal partitions and learning a linear model for each partition. Regression trees, hinge hyperplane method, mixture of experts are some of the approaches to learn continuous piecewise linear regression models. Many of these algorithms are computationally intensive. We present a method of learning piecewise linear regression model which is computationally simple and is capable of learning discontinuous functions as well. The method is based on the idea of K plane regression that can identify a set of linear models given the training data. K plane regression is a simple algorithm motivated by the philosophy of k means clustering. However this simple algorithm has several problems. It does not give a model function so that we can predict the target value for any given input. Also, it is very sensitive to noise. We propose a modified K plane regression algorithm which can learn continuous as well as discontinuous functions. The proposed algorithm still retains the spirit of k means algorithm and after every iteration it improves the objective function. The proposed method learns a proper Piece wise linear model that can be used for prediction. The algorithm is also more robust to additive noise than K plane regression. While learning classifiers, one normally assumes that the class labels in the training data set are noise free. However, in many applications like Spam filtering, text classification etc., the training data can be mislabeled due to subjective errors. In such cases, the standard learning algorithms (SVM, Adaboost, decision trees etc.) start over fitting on the noisy points and lead to poor test accuracy. Thus analyzing the vulnerabilities of classifiers to label noise has recently attracted growing interest from the machine learning community. The existing noise tolerant learning approaches first try to identify the noisy points and then learn classifier on remaining points. In this thesis, we address the issue of developing learning algorithms which are inherently noise tolerant. An algorithm is inherently noise tolerant if, the classifier it learns with noisy samples would have the same performance on test data as that learnt from noise free samples. Algorithms having such robustness (under suitable assumption on the noise) are attractive for learning with noisy samples. Here, we consider non uniform label noise which is a generic noise model. In non uniform label noise, the probability of the class label for an example being incorrect, is a function of the feature vector of the example.(We assume that this probability is less than 0.5 for all feature vectors.) This can account for most cases of noisy data sets. There is no provably optimal algorithm for learning noise tolerant classifiers in presence of non uniform label noise. We propose a novel characterization of noise tolerance of an algorithm. We analyze noise tolerance properties of risk minimization frame work as risk minimization is a common strategy for classifier learning. We show that risk minimization under 01 loss has the best noise tolerance properties. None of the other convex loss functions have such noise tolerance properties. Empirical risk minimization under 01 loss is a hard problem as 01 loss function is not differentiable. We propose a gradient free stochastic optimization technique to minimize risk under 01 loss function for noise tolerant learning of linear classifiers. We show (under some conditions) that the algorithm converges asymptotically to the global minima of the risk under 01 loss function. We illustrate the noise tolerance of our algorithm through simulations experiments. We demonstrate the noise tolerance of the algorithm through simulations.

Proper modelling of the gravitational fields of irregularly shaped asteroids and comets is an essential yet challenging part of any spacecraft visit and flyby to these bodies. Accurate density representations provide crucial information for proximity missions, which rely heavily on it to design safe and efficient trajectories. This work explores using a spacecraft swarm to maximise the measured gravitational signal in a hypothetical mission around the comet 67P/Churyumov-Gerasimenko. Spacecraft trajectories are simultaneously computed and evaluated using a high-order numerical integrator and an evolutionary optimisation method to maximise overall signal return. The propagation is based on an open-source polyhedral gravity model using a detailed mesh of 67P/C-G and considers the comet’s sidereal rotation. We compare performance on various mission scenarios using one and four spacecraft. The results show that the swarm achieved an expected increase in coverage over a single spacecraft when considering a fixed mission duration. However, optimising for a single spacecraft results in a more effective trajectory. The impact of dimensionality is further studied by introducing an iterative local search strategy, resulting in a generally improved robustness for finding efficient solutions. Overall, this work serves as a testbed for designing a set of trajectories in particularly complex gravitational environments, balancing measured signals and risks in a swarm scenario. / En korrekt modellering av de gravitationsfält som uppstår runt irreguljärt formade asteroider och kometer är en avgörande och utmanande del för alla uppdrag till likartade himlakroppar. Exakta densitetsrepresentationer tillhandahåller viktig information för att säkerställa säkra och effektiva rutter för särsilt närgående rymdfarkoster. I denna studie utforskar vi användningen av en svärm av rymdfarkoster för att maximera den uppmätta gravitationssignalen i ett hypotetisk uppdrag runt kometen 67P/Churyumov-Gerasimenko. Rymdfarkosternas banor beräknas och utvärderas i parallella scheman med hjälp av en högre ordningens numerisk integration och en evolutionär optimeringsmetod i syfte att maximera den totala uppmätta signalen. Beräkningarna baseras på en öppen källkod för en polyhedral gravitationsmodell som använder ett detaljerat rutnät av triangulära polygoner för att representera 67P/C-G och beaktar kometens egna rotation. Vi jämför sedan prestanden för olika uppdragscenarier med en respektive fyra rymdfarkoster. Resultaten visar att svärmen uppnådde en förväntad ökning i täckning jämfört med en enskild rymdfarkost under en fast uppdragsvaraktighet. Dock resulterar optimering för en enskild rymdfarkost i en mer effektiv bana. Påverkan av dimensionshöjningen hos oberoende variabler studeras vidare genom att introducera en iterativ lokal sökstrategi, vilket resulterar i en generellt förbättrad robusthet samt effektivare lösningar. Sammantaget fungerar detta arbete som en testbädd för att studera och utforma rymdfarkosters banor i särskilt komplexa gravitationsmiljöer, samt för att balansera uppmätta signaler och risker i ett svärmscenario.

Master of Science / Department of Industrial & Manufacturing Systems Engineering / Todd W. Easton / Integer programs (IPs) are mathematical models that can provide an optimal solution to a variety of different problems. They have been used to reduce costs and optimize organizations. Additionally, IPs are NP-complete resulting in many IPs that cannot be solved. Cutting planes or valid inequalities have been used to decrease the time required to solve IPs. Lifting is a technique that strengthens existing valid inequalities. Lifting inequalities can result in facet defining inequalities, which are the theoretically strongest valid inequalities. Because of these properties, lifting procedures are used in software to reduce the time required to solve an IP. The thesis introduces a new algorithm for exact synchronized simultaneous uplifting over an arbitrary initial inequality for knapsack problems. Synchronized Simultaneous Lifting (SSL) is a pseudopolynomial time algorithm requiring O(nb+n[superscript]3) effort to solve. It exactly uplifts two sets simultaneously into an initial arbitrary valid inequality and creates multiple inequalities of a particular form. This previously undiscovered class of inequalities generated by SSL can be facet defining. A small computational study shows that SSL is quick to execute, requiring on average less than a quarter of a second. Additionally, applying SSL inequalities to a knapsack problem enabled commercial software to solve problems that it could not solve without them.

Master of Science / Department of Industrial and Manufacturing Systems Engineering / Todd Easton / Integer Programming is used to solve numerous optimization problems. This class of mathematical models aims to maximize or minimize a cost function restricted to some constraints and the solution must be integer. One class of widely studied Integer Program (IP) is the Multiple Knapsack Problem (MKP). Unfortunately, both IPs and MKPs are NP-hard, potentially requiring an exponential time to solve these problems. Utilization of cutting planes is one common method to improve the solution time of IPs. A cutting plane is a valid inequality that cuts off a portion of the linear relaxation space. This thesis presents a new class of cutting planes referred to as merged knapsack cover inequalities (MKCI). These valid inequalities combine information from a cover inequality with a knapsack constraint to generate stronger inequalities. Merged knapsack cover inequalities are generated by the Merging Knapsack Cover Algorithm (MKCA), which runs in linear time. These inequalities may be improved by the Exact Improvement Through Dynamic Programming Algorithm (EITDPA) in order to make them stronger inequalities. Theoretical results have demonstrated that this new class of cutting planes may cut off some space of the linear relaxation region. A computational study was performed to determine whether implementation of merged knapsack cover inequalities is computationally effective. Results demonstrated that MKCIs decrease solution time an average of 8% and decrease the number of ticks in CPLEX, a commercial IP solver, approximately 4% when implemented in appropriate instances.