Page 279 - Vibrational Spectroscopic Imaging for Biomedical Applications

Raman Micr oscopy for Biomedical Applications 255

rbf kernel. Recognition rates in Table 8.3 ranged from 96 percent to
100 percent correct classification on the strain level, which gives an
average recognition rate of 98.7 percent. Only one out of 1150 spec-
tra was misclassified on the species level, resulting in a recognition
rate of 99.6 percent for Staphylococcus epidermidis and an overall rec-
61
ognition rate of 99.9 percent. In another study, UV resonance
Raman spectroscopy was applied to identify eight different strains
62
of lactic acid bacteria from yoghurt. Classification was accom-
plished using different chemometric methods. In a first attempt, the
unsupervised methods hierarchical cluster analysis and principal
component analysis were applied to investigate natural grouping in
the data. In a second step the spectra were analyzed using several
supervised methods: k-nearest neighbor classifier, nearest mean
classifier, linear discriminant analysis, and SVMs.

8.4.2 Imaging Single Bacteria
After bacterial identification, the infectious ones require treatment
with efficient antibiotics. A suitable target structure of many antibi-
otics, e.g., of the group of the β-lactams, penicillines and glycopep-
tides is the bacterial cell wall. An intact cell wall is crucial for the
correct function of many vital processes such as, e.g., signal trans-
duction, mass transport, adhesion on surfaces, cell recognition, and
enzyme reactions. Many of these processes that take place at and
through the cell membrane are not completely understood so far.
Furthermore, some extracellular substances were found to form a
biofilm around the bacteria which may alter the effectiveness of the
antibiotics. Therefore, a detailed knowledge about the chemical
composition of the bacterial surface and its spatial arrangement on
a molecular level, as well as an in-depth understanding of the
dynamics of the cell membrane are important for the development
of more effective drugs.
TERS is a recently developed technique which combines SERS
63
with atomic force microscopy (AFM). In SERS a metal surface
with roughness in the nanometer scale is brought into close contact
to a sample to experience electromagnetic and chemical enhance-
14
8
ment factors up to 10 to 10 upon illumination with appropriate
wavelengths. For TERS, the SERS active metal is reduced to the
size of an AFM tip with apex sizes of less than 50 nm in diameter
(Fig 8.11a). This tip is moved across the sample to record the surface
features. At the same time, that part of the sample which is in the
close vicinity of the tip apex experiences the enhanced electromag-
netic field if the wavelength of the laser excitation is tuned to the
plasmon absorption of the nanoparticle. This evanescent electro-
magnetic field decays very rapidly with increasing distance from
the metal surface, so that at only 50 nm away from a metal tip no