Ph.D Thesis

Ph.D StudentLeonard Heidi
SubjectRapid Antimicrobial Susceptibility Testing using Photonic
Silicon Microarrays
DepartmentDepartment of Biotechnology and Food Engineering
Supervisors PROF. Ester H. Segal
PROF. Sarel Halachmi


This research succeeded to design an optical, label-free sensing platform using microstructured silicon arrays for rapid antimicrobial susceptibility testing of clinical samples.  The fabricated silicon sensors, produced by standard lithography and reactive ion etching techniques, act as a two-dimensional diffraction grating capable of trapping bacteria within their periodic topology while simultaneously yielding zero-order reflectance spectra. Values of 2nL, in which n corresponds to refractive index of the medium inside the grating and L corresponds to the depth of the grating, are inferred from frequency analysis of the acquired reflectance spectra. This process was termed phase-shift reflectometric interference spectroscopic measurements (PRISM). By employing PRISM, changes in bacterial growth on the silicon sensors were observed as changes in 2nL; in particular, an increase in 2nL over time corresponded to bacterial growth, while a decrease in 2nL corresponded to bacterial death. By monitoring these optical changes, bacteria were classified as susceptible or resistant in the presence of various antibiotics. PRISM was first demonstrated as a proof-of-concept platform for phenotypic antimicrobial susceptibility testing (AST) using Escherichia coli as a model organism. Minimum inhibitory concentration (MIC) values for five different antibiotics and resistant/susceptible profiles were determined within 2 - 3 hours of bacteria incubation on the silicon sensors.

This platform opened the door to the observation of unique bacterial behaviors guided by surfaces interactions, as bacteria adhesion, growth patterns, and antibiotic resistance on different types of micro-architectures and different surface chemistries were then elucidated by PRISM. Characteristics such a cell motility, charge, and biofilm forming abilities of different bacterial species were explored in their ability to affect bacteria adhesion. Furthermore, chemotactic abilities of synthetic bacteria mutants were also observed using PRISM. With this knowledge of bacteria surface-guided behavior, the PRISM platform for AST was further optimized by incorporation of the silicon sensors into disposable, inexpensive microfluidic devices, consolidation of the data acquisition algorithms, and re-interpretation of the optical signal, leading to simple, static experiments. AST using PRISM of clinical samples, including blood, urine, and bacteria isolates collected from Bnai Zion Medical Center, were consequently tested against two commonly used antibiotics, namely gentamicin and ciprofloxacin. MIC values were determined within 30 - 150 minutes using minimal sample preparation, which is significantly faster than any currently existing phenotypic AST platform. Thus, the PRISM platform is expected to be further validated a clinical setting in order to encourage the correct prescription of antibiotics and reduce the instances of antimicrobial resistance.