About

Richard D. Braatz is the Edwin R. Gilliland Professor of Chemical Engineering at MIT. His research focuses on chemical engineering, with particular emphasis on areas related to his department's expertise. As a faculty member, he contributes to the academic community through teaching and research, although specific details about his research interests and contributions are not provided in the page text.

Research topics

• Computer Science
• Machine Learning
• Artificial Intelligence
• Engineering
• Political Science
• Reliability engineering
• Waste management
• Operating system
• Data science
• Business
• Chemistry
• Physics
• Environmental science
• Manufacturing engineering
• Engineering ethics

Selected publications

• Least-Squares Multi-Step Koopman Operator Learning for Model Predictive Control

  arXiv (Cornell University) · 2026-01-17

  preprintOpen access

  MPC is widely used in real-time applications, but practical implementations are typically restricted to convex QP formulations to ensure fast and certified execution. Koopman-based MPC enables QP-based control of nonlinear systems by lifting the dynamics to a higher-dimensional linear representation. However, existing approaches rely on single-step EDMD. Consequently, prediction errors may accumulate over long horizons when the EDMD operator is applied recursively. Moreover, the multi-step prediction loss is nonconvex with respect to the single-step EDMD operator, making long-horizon model identification particularly challenging. This paper proposes a multi-step EDMD framework that directly learns the condensed multi-step state-control mapping required for Koopman-MPC, thereby bypassing explicit identification of the lifted system matrices and subsequent model condensation. The resulting identification problem admits a convex least-squares formulation. We further show that the problem decomposes across prediction horizons and state coordinates, enabling parallel computation and row-wise $\ell_1$-regularization for automatic dictionary pruning. A non-asymptotic finite-sample analysis demonstrates that, unlike one-step EDMD, the proposed method avoids error compounding and yields error bounds that depend only on the target multi-step mapping. Numerical examples validate improved long-horizon prediction accuracy and closed-loop performance.

  PublisherDOI

• GlycoPy: A CasADi-based Python Framework for Hierarchical Modeling, Optimization, and Control of Bioprocesses

  arXiv (Cornell University) · 2026-01-04

  preprintOpen accessSenior author

  Efficient implementation of nonlinear model predictive control (NMPC) for bioprocesses remains challenging because large nonlinear models are difficult to organize, simulate, and embed within optimization and control workflows. This difficulty is particularly pronounced for large-scale and multiscale systems that require hierarchical model construction and customized simulation strategies. To address this issue, we present GlycoPy, a CasADi-based Python framework for hierarchical modeling, optimization, and control of bioprocesses. GlycoPy combines an equation-oriented, object-oriented modeling architecture with CasADi's symbolic and differentiable computational capabilities, enabling hierarchical model composition, numerical and symbolic simulation, parameter estimation, dynamic optimization, and NMPC within a unified workflow. A key feature of the framework is its support for customized differentiable simulation algorithms that can be embedded directly in gradient-based optimization and control. GlycoPy is demonstrated on a multiscale monoclonal antibody glycosylation process in Chinese hamster ovary cell culture, where it is used for hierarchical model construction, quasi-steady-state simulation, and adaptive NMPC. The results show that GlycoPy provides a practical and reusable framework for applying advanced optimization and control methods to computationally demanding bioprocesses.

  PublisherDOI

• Adaptive nonlinear model predictive control of monoclonal antibody glycosylation in CHO cell culture

  Control Engineering Practice · 2026-01-07

  articleSenior authorCorresponding
  PublisherDOI

• A Lego Block Approach to Flow in Complex Microfluidic Networks

  arXiv (Cornell University) · 2026-03-23

  articleOpen accessSenior author

  We present a new way to construct analytical solutions for flow in complex microfluidic channel networks, as well as planar disordered media. Using a combination of Schwarz-Christoffel maps and segmentation techniques inspired by integrated circuit analysis, we build a library of base building blocks which can be reassembled to model complex geometries, in the style of ``Lego Blocks''. Our approach requires minimal numerical computation, and can then generate analytical solutions for any combination of inlet and outlet flow rates. Moreover, our method can tackle multiply connected domains which are usually difficult to model using typical conformal transform approaches. The solutions are developed for microfluidic Hele-Shaw cell devices, but also apply to ideal flow and Darcy flow in complex geometries, or any other flow problem adequately modeled by Laplace's equation. We end by showing how the procedure can be used to model complex disordered media, fractal-like flow geometries, as well as problems of steady advection-diffusion in microfluidic mixers.

  PublisherOA PDF

• Understanding Size Distributions during Lipid Nanoparticle Manufacturing through Mechanistic Modeling

  ACS Omega · 2026-04-27

  articleOpen accessSenior authorCorresponding

  Recent breakthroughs in messenger RNA (mRNA) therapeutics have highlighted the importance of lipid nanoparticles (LNPs) as delivery vehicles that protect fragile mRNA and facilitate cellular uptake. While manufacturing technologies that rely on self-assembly through mixing, including microfluidic systems and turbulent jet mixers, have enabled scalable production, a first-principles model explaining the fundamental mechanisms responsible for LNP size and its distribution has remained elusive. This has limited predictive process control and optimization. To address this gap, we present here a model based on crystallization of a lipid during turbulent mixing of water-ethanol, employing population balance equations to predict particle size distributions. The model successfully predicts the influence of key operating parameters including total flow rate, flow rate ratio, and lipid concentration, demonstrating favorable comparison with experimental results and enabling the development of predictive manufacturing processes for consistent critical quality attributes. The mechanistic understanding provides the foundation necessary to optimize manufacturing of LNPs used as delivery agents of mRNA vaccines, which extends beyond the current application to the broader field of nucleic acid therapeutics.

  PublisherDOI

• DNA Nanoflowers Assemble via Template-Induced Crystallization during mRNA Synthesis

  Crystal Growth & Design · 2026-02-12

  article

  mRNA-based vaccines and therapeutics are synthesized using the in vitro transcription (IVT) reaction. However, the byproduct pyrophosphate (PPi) forms insoluble magnesium pyrophosphate (Mg2PPi) during IVT, which coprecipitates with DNA, the transcription blueprint, interrupting the reaction and reducing RNA yield. The severity of this crystallization problem has prompted the routine inclusion of inorganic pyrophosphatase (PPase) in industrial processes to hydrolyze pyrophosphate into phosphate ions and prevent crystallization. To date, the mechanism by which Mg2PPi crystals interact with DNA has remained unclear. In this work, we demonstrate that DNA promotes Mg2PPi crystallization through a template-directed mechanism, reducing the nucleation induction time by a factor of 2. The resulting crystals form compact spherulites with a rosette habit, distinct from the morphologies observed in DNA-free systems. We propose that DNA preferentially templates the nucleation of Mg2PPi·3.5H2O crystals by interacting with the {100} crystal face. These findings reveal how DNA influences Mg2PPi crystallization and offer mechanistic insights relevant to optimizing RNA manufacturing systems.

  PublisherDOI

• Codesigning Tubular Flow and Noncontact Sonication for Antisolvent Crystallization: Rapid Fouling-Free Production of Uniform Rifapentine Crystals with Reduced Aspect Ratios

  Crystal Growth & Design · 2026-02-17

  article

  Rifapentine (RPT), a critical antituberculosis active pharmaceutical ingredient (API), faces persistent supply shortages compounded by crystallization challenges. Conventional antisolvent crystallization often produces RPT crystals with high aspect ratios (>7) and sizes outside the preferred 5–20 μm range. RPT’s strong adhesion to solid surfaces further increases fouling risk, and makes conventional milling impractical due to equipment damage. To avoid fouling, prior processes have relied on low supersaturation (≤1.5), resulting in long residence times of 4–6 hours. This study introduces a modular, continuous, fouling-free crystallization process that simultaneously addresses these limitations by codesigning precise antisolvent injection, controlled supersaturation, and dynamic noncontact ultrasonication within a tubular slug-flow platform. We establish design criteria enabling reproducible antisolvent (water) injection into flowing RPT solution slugs (in tetrahydrofuran) via continuous droplet formation and merging, minimizing needs for feedback control or complex operation. Guided by simplified semiempirical models and rapid droplet/slug tests, noncontact sonication is strategically applied at three stages─nucleation, growth, and breakage─to suppress fouling and enhance uniformity, without needing milling. Within 0.5 hour of residence time, these steps collectively reduce crystal dimensions while maintaining small variations (e.g., length from 81 μm to 6 μm, coefficient of variation or CVL ∼ 0.20; aspect ratio from 33 to 3.5, CVAR ∼ 0.29) and 100% form purity. This integrated approach offers a scalable pathway toward reliable RPT crystallization and supply resilience.

  PublisherDOI

• Koopman-BoxQP: Solving Large-Scale NMPC at kHz Rates

  Open MIND · 2026-02-20

  preprint

  Solving large-scale nonlinear model predictive control (NMPC) problems at kilohertz (kHz) rates on standard processors remains a formidable challenge. This paper proposes a Koopman-BoxQP framework that i) learns a linear Koopman high-dimensional model, ii) eliminates the high-dimensional observables to construct a multi-step prediction model of the states and control inputs, iii) penalizes the multi-step prediction model into the objective, which results in a structured box-constrained quadratic program (BoxQP) whose decision variables include both the system states and control inputs, iv) develops a structure-exploited and warm-starting-supported variant of the feasible Mehrotra's interior-point algorithm for BoxQP. Numerical results demonstrate that Koopman-BoxQP can solve a large-scale NMPC problem with $1040$ variables and $2080$ inequalities at a kHz rate.

  DOI

• LQR for Systems with Probabilistic Parametric Uncertainties: A Gradient Method

  ArXiv.org · 2026-03-27

  articleOpen accessSenior author

  A gradient-based method is proposed for solving the linear quadratic regulator (LQR) problem for linear systems with nonlinear dependence on time-invariant probabilistic parametric uncertainties. The approach explicitly accounts for model uncertainty and ensures robust performance. By leveraging polynomial chaos theory (PCT) in conjunction with policy optimization techniques, the original stochastic system is lifted into a high-dimensional linear time-invariant (LTI) system with structured state-feedback control. A first-order gradient descent algorithm is then developed to directly optimize the structured feedback gain and iteratively minimize the LQR cost. We rigorously establish linear convergence of the gradient descent algorithm and show that the PCT-based approximation error decays algebraically at a rate $O(N^{-p})$ for any positive integer $p$, where $N$ denotes the order of the polynomials. Numerical examples demonstrate that the proposed method achieves significantly higher computational efficiency than conventional bilinear matrix inequality (BMI)-based approaches.

  PublisherOA PDF

• Continuous Precipitation of Biotherapeutics: A Review

  ChemBioEng Reviews · 2026-01-07

  articleOpen accessSenior authorCorresponding

  Abstract The increasing efficiency of upstream production of biotherapeutics exposes significant challenges in downstream purification. Chromatography, the traditional gold standard, offers high purity and yield but faces limitations such as restricted mass transfer, limited capacity, and scalability issues, especially with high‐titer strains. As biomanufacturing moves toward fully continuous production, interest grows in alternatives beyond chromatography. Precipitation has emerged as a versatile, scalable, and titer‐independent technique that achieves yields and purities comparable to chromatography while offering simpler and more adaptable processing. It can be operated continuously, enabling seamless bioprocessing. Despite its potential, continuous precipitation remains less explored for biotherapeutics due to biomolecular complexity. Effective design requires deep understanding of thermodynamics, process modeling, hardware, and real‐time monitoring. This review discusses these critical factors, key design elements, and application examples, highlighting precipitation as a promising purification technology and exploring future research directions.

  PublisherOA PDFDOI

Recent grants

• Collaborative Research: Model Reformulation for Lithium-ion Batteries - Parameter Estimation and Dynamic Optimization

  NSF · $106k · 2008–2011

Frequent coauthors

• Dennis S. Bernstein

  University of Michigan–Ann Arbor

  98 shared

• Stephen Yurkovich

  The University of Texas at Dallas

  97 shared

• Hong Yue

  97 shared

• Rodolphe Sepulchre

  97 shared

• Hesuan Hu

  Xi'an Jiaotong University

  97 shared

• Scott Ploen

  Jet Propulsion Laboratory

  97 shared

• Reginald B. H. Tan

  82 shared

• Joshua D. Isom

  University of Cambridge

  75 shared

Labs

• MIT ChemEPI

Education

• Ph.D., Chemical Engineering

  Massachusetts Institute of Technology

  1986

• M.S., Chemical Engineering

  Massachusetts Institute of Technology

  1982

• B.S., Chemical Engineering

  University of California, Berkeley

  1980

Awards & honors

• John R. Ragazzini Education Award (2023)
• Elected to the National Academy of Engineering (2019)
• AIChE Separation Division Innovation Award (2019)
• Elected AIChE Fellow (2018)
• Automatica Paper Prize (2017)