Learning Center
Plans & pricing Sign in
Sign Out

Chemical _ Biomolecular Engineering


									                                                        Martha Grover, Associate Professor
                                                   School of Chemical & Biomolecular Engineering
                                                           Georgia Institute of Technology
                                                                 Atlanta, GA USA

Research program
We apply the systems engineering approach to materials synthesis and
macromolecular structure. Mathematical models are not accurate enough for a
purely model-based design. Instead, we combine mechanistic modeling together
with statistics and experimentation. In the group we develop general methodology
and algorithms, and then apply the techniques in a number of specific applications.

Areas of expertise: systems engineering, materials processing
              Systems engineering:                 Process modeling                  Experimental design                                         Sensing and estimation
Applications in materials
Chemical vapor deposition
Supercritical processing of metal
Hyperbranched and crosslinking
Polymer nanoparticles
Pharmaceutical crystallization

 Research Results on Chemical Vapor                 Experimental Design for Process Optimization         In Situ Optical Sensing and Robust Estimation
 Deposition of Metal Oxide Thin Films                                                                    Motivation
                                                                          f(x)                           Interpreting optical sensor data is a major
 Key publications:                                  • Models can be
                                                                                            UB           hurdle to real-time control of thin films.
                                                    helpful in finding
 R. Xiong and M. A. Grover, “In situ estimation
                                                    best process, but      H                                  y     j=8                     m=5
 of thin film growth rate, complex refractive
                                                    they are not                                                                            y6

 index, and roughness during chemical vapor                                           LB
                                                                                                                                       y5        y7

                                                    completely accurate                                            y1             y4
 deposition using a modified moving horizon                                                                   y0                                      y8
                                                    • Models should be                           x                      y2
 estimator,” Journal of Applied Physics, 103(12)                        Conclusions
                                                    used to design the                                                                                     t

 124901 (2008).                                                         • Models can be used to
                                                    next experiment.                                     Conclusions
                                                                        determine which
 P. J. Wissmann and M. A. Grover, “A new                                                                 • Moving horizon
                                                                        process settings could
 approach to batch process optimization using                                                            estimation combines
                                                                        be optimal (based on
 experimental design,” AIChE Journal, 55(2),                                                             models with knowledge of
                                                                        statistical data analysis)
 342-353 (2009).                                                                                         uncertainties.
                                                                        • The next experiment
                                                                        should be performed in           • MHE is more robust than
 Research support:                                                                                       the current approach based
                                                                        this region of potential
 National Science Foundation CAREER award “A                            optima.                          on fitting an optical model.
 Systems Approach to Materials Processing

Research partners needed                                                                             Research facilities available
• Modeling of metal nanoparticle nucleation on                                                       In Professor Grover’s lab
surfaces                                                                                             • Chemical vapor deposition system (custom
• Expertise in robust process design                                                                 built)
• Experimental data with in situ optical                                                             • Atomic force microscope (Molecular Imaging)
measurements of surfaces                                                                             Shared user facilities at Georgia Tech (for fee)
• Other applications in materials processing that                                                    • Electron microscopy center
could benefit from the “systems approach”                                                            • Microelectronics Research Center
                                                                                                     • Marcus Nanotechnology Research Center

To top