University of Rhode Island Sensors and Surface Technology Partnership


General Information:

History of the SST Partnership
Participating Faculty
Facilities
Participating Companies
Student Opportunities
Seminars
Annual Poster Session



Research Areas:

Microsensors
Surface Engineering
Food Pathogens
IR Imaging for Health Monitoring



Instrumentation Reservation:

Reservation Calendar



Other Links:

SST Home Page
Microbeam Analysis Society
URI Home Page


Detection and Identification of Food Pathogens by IR Imaging

Infrared spectroscopic analysis of bacteria has been focused on the mid-IR region of 900 to 1800 cm–1. Much of the differentiation amount species has been confine to the 900 to 1200 cm–1 region due to composition differences in phospholipids and polysaccharides, which absorb in this region. There has also been some emphasis amide I and amide II bands at ~1650 and ~1550 cm–1, and on the C-H stretching region of 2700 to 3100 cm–1.

Since near-IR spectroscopy has been so effective in the analysis of agriculture products, the near-IR spectra of bacteria were measured. Five bacteria including Salmonella, E. coli O157:H7, S. aureus, Listeria monocytogenes, and non-pathogenic E. coli were analyzed in the preliminary study. Significant differences were observed between all five bacteria as shown in the spectra below.

Vertical lines mark similar bands in the spectra of non-pathogenic E. coli and E. coli O157:H7; two new bands were observed in the spectra of pathogenic E. coli. Bands in the region of 4000 to 5000 cm–1 are due to combinations of the bands (vibrations) in the mid-IR fingerprint region of 700-1800 cm–1 with the bands (vibrations) of the C-H stretching region of 2800 to 3000 cm–1. The sensitivity in the near-IR is about one-tenth to one-hundredth weaker than the mid-IR; however, detectors and radiation sources are very good in the near-IR and interferences from water are much less.

Recently, we measured mid-IR hyperspectral images of the 5 bacterial discussed above. All of the images were measured on a microscopic scale, which was 40×40 micron in size. The 64×64 focal plane array camera produced 4096 spectra with a wavenumber range of 4000-900 cm–1 for each sample. The bacteria samples were smeared onto an infrared transparent ZnSe window. The infrared images formed by plotting the intensity ratio of the amide I band at 1663 cm–1 to the amide II band at 1550 cm–1 of the five bacteria are shown below.

The spectrum of all five samples was similar to that of protein; however, each contained distinctive bands and contours, which were associated with characteristics of the particular protein.

A hyperspectral images of two bacteria, E. coli O157:H7 and S. aureus, are shown in the figure below.

The spectra were all normalized to the height of the Amide II band at 1550 cm–1. Plotting intensity of the Amide I (1663 cm–1) formed the image on the left. In this particular image it is not possible to detect the difference between the two bacteria, since they both contain similar amounts of protein. However, by using a self-modeling method we were able to “extract” the spectrum (right side of the figure) of a second chemical component, which corresponded to a phospholipid. The major band in this spectrum was at 1158 cm–1; the image based on the intensity of this band is shown on the right of the figure. It is clear that the E. coli contained this compound possibly as the surface membrane, and this clearly distinguished the two bacteria.

This research has been led by Professors Chris Brown, Steve Letcher, and Garth Rand. Students who have been involved include Scott Huffman and Kara Lukasiewicz.

Partial support for this research was provided by the U. S. Department of Agriculture. Hyperspectral images were obtained using equipment at BioRad Corporation.