Description:

Zhekai Deng (Ottino/Lueptow group)

Continuum Modeling of Segregation in Polydisperse Granular Materials

Segregation of polydisperse granular materials (i.e., granular materials with a continuous distribution of a particular particle property) remains a challenging problem in bulk solids handling across many industries because it can detrimentally affect product quality. However, most segregation studies have focused on either size bidisperse materials (two different particle sizes), which are not true representation of real mixtures encountered in the physical system, or on polydisperse materials in idealized periodic chute flow using Discrete Element Method (DEM) simulations. In this talk, I will share how we developed a continuum model that captures the effects of advection, diffusion, and segregation to model polydisperse segregation in the more realistic and challenging situations of developing (spatially dependent) segregation in a bounded heap and transient (time dependent) segregation in a rotating tumbler. The model was then extended to predict polydisperse segregation in hopper discharge flow in collaboration with the Dow Chemical Company. Quantitative agreement of continuum model results with DEM simulations and experiments over a wide range of flow conditions and particle distributions supports our modeling approach. The model has the potential to accurately predict polydisperse segregation at industrial scales and, thereby, enhance solids processing unit design to minimize segregation.

Hanyu Shi (Amaral group)

Development of benchmarks to evaluate the performance of topic modeling algorithms

In the information age, a large volume of information is produced in the form of unstructured text. There is a desire to develop computer algorithms to extract meaningful information from these “big data”. Topic Models have become a standard tool in the automatic organization and classification of large collections of unstructured texts. Many topic model algorithms have been proposed in last decades, however, there is still a considerable lack of understanding of how accurately topic models infer the underlying topical structure.
In this work, we propose synthetic benchmark corpora which are an artificially constructed collection of documents with a given topical structure. The main advantage of these synthetic corpora is that they are completely tunable in size (such as number of documents), as well as the degree of structure. Since the “true” topical structure (including the number of topics) of synthetic benchmark corpora is completely known, this allows us to rigorously evaluate the results of different topic model algorithms. We assess how robust topic model algorithms are with respect to noise or how prone they are to overfitting in the absence of structure. Our main result is that there are principle limits on the detectability of topical structure similar to “Detectability Thresholds” in the community detection in complex networks. We emphasize the generality of these findings by i) applying commonly used “real-world” tasks to the synthetic benchmark corpora (such as document classification).

Graham Spicer (Backman and Shea groups)

Development of inverse spectroscopic optical coherence tomography for cancer screening at a synthetic pre-metastatic niche

Metastasis is responsible for 90% of cancer-related deaths. The arrival and colonization of metastatic tumor cells in an organ is facilitated by formation of a “pre-metastatic niche”, a local tissue microenvironment primed by immune cells and soluble factors. An emerging approach for early detection of systemic metastasis utilizes the concept of the pre-metastatic niche to engineer a synthetic niche that can be easily monitored. Our lab has developed a porous polymer scaffold that mimics the behavior of the pre-metastatic niche in vivo upon implantation into the subcutaneous space to facilitate detection of metastatic progression. We have used an extension of spectral domain optical coherence tomography, termed inverse spectroscopic OCT (ISOCT) for the early detection of metastasis via measurement of structural changes in the scaffold implant. I will present work using ISOCT to successfully differentiate healthy and tumor-bearing mice in an orthotopic model of metastatic breast cancer by measuring the scaffold mass fractal dimension. Limited penetration of OCT through the skin led to development of a dual-bandwidth benchtop OCT and an ISOCT needle probe designed for in situ imaging of the scaffold. Investigations into the sensitivity of ISOCT to cancer-associated remodeling of constituent scaffold compartments and histological study of scaffold structure will be discussed. The last part of the talk will outline development of a clinical paradigm for scaffold measurement and metastasis detection.

Description:

Michelle O'Malley
Associate Professor in Chemical Engineering
University of California, Santa Barbara

Exploiting Anaerobes for Biomass Breakdown and Sustainable Chemistry

Renewable chemicals derived from plant biomass are attractive alternatives to those made from petroleum. To produce chemicals from biomass, enzymes are used to break down cellulose into simple sugars, which are later fermented into value-added products. However, since cellulose is tightly bound within a network of crystalline cellulosic fibers and lignin, existing biomass degrading enzymes are not very efficient. To develop new technologies that break down plant material into sugar, much can be learned by studying how microbes digest lignocellulose in biomass-rich environments, such as the digestive tract of large herbivores. Our goal is to develop new experimental tools to engineer anaerobic microbial gut consortia for lignocellulose breakdown and chemical production. To accomplish this goal, we isolated a panel of anaerobic fungi and associated microbes from different herbivores and screened for their ability to degrade several types of lignin-rich agricultural waste. By focusing on model anaerobic fungi from the Piromyces, Neocallimastix, Anaeromyces, and Caecomyces genera, we discovered thousands of new genes from these systems, revealing hundreds of novel biomass-degrading enzymes that are already competitive with industrial standards. We characterized key regulatory patterns for these enzymes, which depend on the environment of the fungus. Combined with proteomic approaches, a number of enzymes with non-catalytic fungal dockerin domains were also characterized, providing the first comprehensive insight into the composition and architecture of fungal cellulosomes. Using this information, we are developing new genetic engineering strategies to manipulate gut fungi at the molecular level, along with ‘bottom-up’ strategies to synthesize microbial consortia for compartmentalized breakdown and bioproduction.

Michelle A. O’Malley is an Associate Professor in the Department of Chemical Engineering at the University of California, Santa Barbara. She earned a B.S. in Chemical Engineering and Biomedical Engineering from Carnegie Mellon University in 2004 and a PhD in Chemical Engineering from the University of Delaware in 2009, where she worked with Prof. Anne Robinson to engineer overproduction of membrane proteins in yeast. O’Malley was a USDA-NIFA postdoctoral fellow in the Department of Biology at MIT, where she developed new strategies for cellulosic biofuel production. At UCSB, her research group engineers protein synthesis within anaerobes and consortia for sustainable chemical production, bioremediation, and natural product discovery. O’Malley’s research has been featured on NPR’s Science Friday, the BBC Newshour, the LA Times, and several other media outlets. She was named one of the 35 Top Innovators Under 35 in the world by MIT Technology Review in 2015, and is the recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE), a DOE Early Career Award, an NSF CAREER award, the Camille Dreyfus Teacher-Scholar Award, the ACS BIOT Division Young Investigator Award, an ACS PMSE Division Young Investigator Award, an ACS WCC “Rising Star” Award, and a Hellman Faculty Fellowship.

Description:

Weekly guest and student speakers in the Department of Chemical and Biological Engineering.

PhD Candidate Presentations

Aysenur Iscen (Schatz Group)

Title: Design of Stimuli-Responsive Bioinspired Materials using Computational Methods

Abstract: The development of soft matter that mimics living organisms in its capacity to autonomously translate or morph reversibly is a grand challenge for science. Such soft materials could be transformative in emerging areas such as soft robotics, interactive actuators, and sensor-effector devices. Design of switchable soft actuators that can be controlled by external stimuli, such as temperature, pH and light, requires an understanding of specific molecular interactions. Using atomistic and coarse-grained models with classical molecular dynamics simulations, we aim to provide insight into the conformational changes of biomolecular assemblies in response to external stimuli. In this talk, I will focus on computational methods to design functional materials, including a pH-responsive ultrasmall nano-grafted nanobin for drug delivery applications, supramolecular-covalent hybrid polymers for light-activated robotic functions and multivalent cation-induced actuation of DNA-mediated colloidal superlattices.

Kazi Y. Helal (Mrksich Group)

Title: Traceless Immobilization of Analytes for High-Throughput Experiments with SAMDI Mass Spectrometry 

Abstract: Label-free assays, and particularly those based on the combination of mass spectroscopy with surface chemistries, enable high-throughput experiments of a broad range of reactions. However, these methods can still require the incorporation of functional groups that allow immobilization of reactants and products to surfaces prior to analysis. Our lab developed a traceless method for attaching molecules to a self-assembled monolayer for matrix-assisted laser desorption and ionization (SAMDI) mass spectrometry. This method uses monolayers that are functionalized with a 3- trifluoromethyl-3-phenyl-diazirine group that liberates nitrogen when irradiated and gives a carbene that inserts into a wide range of bonds to covalently immobilize molecules. Analysis of the monolayer with SAMDI then reveals peaks for each of the adducts formed from molecules in the sample. This method is applied to characterize a P450 drug metabolizing enzyme and to monitor a Suzuki−Miyaura coupling chemical reaction and is important because modification of the substrates with a functional group would alter their activities. This method will be important for high-throughput experiments in many areas, including reaction discovery and optimization.

Description:

Weekly guest and student speakers in the Department of Chemical and Biological Engineering.

Gavin Towler
Honeywell

Title: Connected Plant

Abstract: Advances in process instrumentation, internet connectivity and cloud computing are setting the stage for a revolution in process automation technology. Chemical plant operators are increasingly able to fit sophisticated models to plant data in close to real time, enabling better decisions on process optimization, plant reliability and worker safety and productivity. While the Industrial Internet of Things (IIoT) / Industrie 4.0 provides the potential to transform the way chemical plants are operated, the success of these technologies depends on deployment of chemical engineering insights and the transfer of chemical engineering methodologies from steady-state design mode to dynamic operational mode. Examples will be given to illustrate how engineers at Honeywell are approaching these challenges.

Bio: Gavin Towler is the Vice President and Chief Technology Officer of Honeywell Performance Materials & Technologies (PMT). Honeywell PMT is a global leader in providing customers with high-performance specialty materials, process technologies and materials for petroleum refining, petrochemical production, and natural gas processing as well as products, services and solutions for industrial process automation. Gavin has 27 years of broad experience of process and product design in the chemicals and fuels industries and has 72 US patents. He is co-author of “Chemical Engineering Design”, a textbook on process design, and is an Adjunct Professor at Northwestern University and the National University of Singapore, where he helps teach the senior design classes. Gavin has a B.A. and M.Eng. in chemical engineering from Cambridge University and a Ph.D. from U.C. Berkeley. He is a Chartered Engineer and Fellow of the Institute of Chemical Engineers, a Fellow of the AIChE and a member of the United States National Academy of Engineering.

Description:

Weekly guest and student speakers in the Department of Chemical and Biological Engineering.

Ashley Smart
MIT

Title: We are All Science Storytellers

Summary: Scientists and the journalists who write about them share a common goal: raising awareness of the science that shapes our lives. Yet they often take different tacks toward achieving that goal. At a time when society is increasingly looking to science for answers, how are scientists and journalists working together to ensure that science stories are meaningful, accurate, and impactful? Can we do better?

Bio: Ashley Smart is associate director of the Knight Science Journalism Program at MIT and a senior editor at Undark magazine. Prior to joining MIT, he spent eight years as an editor and reporter at Physics Today, and was a 2015-16 Knight Science Journalism Fellow. He has a PhD in chemical and biological engineering from Northwestern University.