|Home | Overview | Traineeships | Program Activities | People | Research | Publications | Communications | Contact|
Improving Control of Atomic Force Microscopes for Biomedical Applications
The original aim of my research is to design a control system for an atomic force microscope (AFM) that will allow a new highly accurate and reliable method potentially suitable for early detection of cervical cancer viable for use in research and general medical use. In this method, developed in Dr. Igor Sokolov's lab, cervical cells are collected and the cells are imaged in the AFM, producing an adhesion map. Fractal dimensional analysis is conducted on the adhesion map to identify malignancy of the cell.
Figure 1: Histogram of fractal dimensionality for 300 cervical cells [M. Dokukin et. al. 2011]
In Figure 1, we see a clear distinction between healthy cells and cancerous cells. At this time, adhesion fractal dimensionality is the only cellular characteristic that shows such a clear distinction. Despite remarkable accuracy surpassing all existing methods, there is serious deficiency of this method. The time of collecting one adhesion map is too long to be practical. When the AFM scanning velocity crosses a threshold, we start to see artifacts in the adhesion map. Artifacts make that particular map useless for analysis and it can be hard to tell when artifacts start appearing.
Figure 2: Adhesion maps of cervical cells (upper) tip velocity is 0.1 um/s (lower) tip velocity is 0.5 um/s [Maxim Dokukin]
In Figure 2, we can see artifacts in the lower image in the form of streaking or blurring. Images are obtained using Peak Force mode. The cause of the artifacts in this case could be tip-crashing, accidental excitation of piezo or sensor dynamics, or some other reason and investigation is ongoing.
The goal of my proposed research is to investigate control methods to significantly decrease the time of recording the adhesion map. The immediate goals of my research are to study AFM probe tip to cell surface interactions, study non-linear control methods, and to research the effectiveness of this technique on a wide variety of biological and synthetic tissues.
Figure 3: AFM z-direction block diagram, relevant functions inside the subsystem blocks
Figure 4: AFM probe deflection excited at 1ms intervals on flat Si wafer (left) experiment (right) simulation
So far, research has yielded a Simulink (Matlab) simulation of
probe dynamics and interactions and control in the z-direction. The
block diagram and simulation output can be seen above in Figures 3
and 4. With the exclusion of the small interference frequency, the
simulation almost exactly matches experiment.
|© 2019 Tufts University | Photo credits: Tufts University and Getty Images | Privacy|