Investigation of the influence of the number of spectral channels in colorimetric analysis

Colorimetric analysis is widely used for fast and cost effective determination of the concentration of chemical compounds in both liquids and gases [1] – [3] . Some well-known examples include test strips for pH measurement, and urine test, but such technique is also used in safety and health relevant applications, e.g. fire detection [2] and point-of-care diagnosis [3] . The basic working principle relies on a change in the absorption/reflection of the material when interacting with the substance to be detected. The spectral change, which can be proportional to the concentration of a chemical substance or parameter or just indicate the presence of it, results in a visible color change. The chemistry of the color change has been, so far, developed such that it is detectable with the human eye. A visual inspection is in many cases very subjective and dependent on conditions in which the inspection is made, e.g. lighting. Moreover, there are only so many color variations that the human eye can unambiguously distinguish. Both factors lead to errors and a limited accuracy. Such limitations can be overcome by a spectral measurement of the color changing substance, i.e. a digitalization. On top of removing the limitations, having access to the spectrum, allows to measure the influence of multiple reagents at the same time and to expand the spectral region in which color-changing materials can be used, i.e. the non-visible. However, using a state of the art spectrometer is not really an option, as the main applications of colorimetric analysis require compact and cheap devices. Suitable alternatives are small sensors with the ability to obtain spectral information. The simplest example of which is a RGB sensor. A sensor of this type can be seen as the equivalent of a spectrometer with only 3 channels and it closely replicates the human perception of colors. Other similar solutions with a larger number of channels, and with better spectral resolution, are also already available or are currently being developed [1] , [4] .