The colorimetric sensor array (CSA) is an array of colored chemical indicators of diverse reactivities embedded in a nanoporous sol-gel matrix, developed in the laboratory of Ken Suslick at the University of Illinois and reported a series of academic publications starting with an article in Nature in 2000. Each indicators in the array very sensitively changes color, creating a high dimensional and specific fingerprint allowing identification of the species or mixture presented. Because the reactive indicators are highly diverse chemically, a very wide range of chemical species can be selective detected. Further, the sensor has proven very well suited to identifying the fingerprint of highly complex mixtures.
Sensor Background and Capabilities.
Regardless of the analytes, the detection event is observed as numerous printed dots changing color. These events can be recorded with a typical imaging device. While kinetic information may be obtained by repeatedly scanning the CSA, typically ‘before’ and ‘after’ images are sufficient to determine identity. Given two (before and after exposure) images of the CSA, a color difference map may be generated by superimposing and subtracting the two images. This color difference map can be quantified as a high dimensional vector of RGB values. These vectors can be used to discriminate different analytes while retaining their chemical resemblances.
Toxic Industrial Chemicals. As a demonstration of our classification power, we have previously reported the detection of 20 different toxic industrial chemicals (TICs). Clear differentiation among 20 different TICs was been easily achieved within minutes of exposure as shown below.
On another example, the discrimination of different kinds of coffees was achieved with the use of our colorimetric sensor array. Coffee provides a readily available example of discrimination among closely similar complex mixtures. Using HCA, 10 commercial coffees and controls were differentiated without errors. The method underlying this successful identification of complex mixtures rely on treating the mixture as a single analyte. Compared with traditional techniques using complicated separation techniques such as gas chromatography, this approach is less expensive and easier to operate. While mixtures in the field are difficult to predict, diversifying our pesticide library with different background gases will be part of our path to success.
Another example of the sensor’s ability to distinguish complex mixtures is demonstrated in the ability to rapidly identify pathogenic microorganisms based on the sensor “fingerprint” of volatile small molecule metabolites produced by cells during growth. An ongoing study at Stanford University of now over 3000 cultures including more than 20 species and 50 strains has shown identification accuracy of 95%
Placed in a culture the resulting pattern of color changes comprises a high-dimensional fingerprint of the cell type for bacteria the species and strain. The figure below demonstrates the patterns produced by 9 distinct bacteria grown in culture.