Smartphone-Based Device for Monitoring Chemical Contaminants in Water

Samuel Ozeh, Purdue University


Optimized monitoring of water-borne chemical pollutants is important to curbing inadvertent consumption of contaminated water. Conventional analytical instruments such as atomic absorption fluorescent spectrometer and inductively coupled plasma mass spectrometer are not amenable for such frequent monitoring needs as they are expensive to purchase and maintain, require highly trained personnel and delicate sample handling and transportation. Thus it is pertinent to develop sensitive techniques for on-site detection of target water borne chemical contaminants. Colorimetry has been widely explored as a viable alternative for the conventional analytical instruments. However, there is still room for optimization and innovating a test system that is cost efficient, ubiquitous, easy to operate, and yet possessing fast assay time. The system in consideration is also expected to quantitatively detect trace target ions in water, yielding results that are within the range of values as obtainable from conventional analytical instruments. Smartphone based sensor systems are emerging as an elegant approach for on-site testing of water due to their widespread availability, simplicity and potential for remote contaminant detection. This research work presents a smartphone based colorimetric detection of low-concentration chemical contaminants in water. The RGB components of the analytes’ images are converted to its Hue Saturation Value (HSV) equivalents. Components of the Hue, Saturation and Value (HSV) color space within a global region of interest (ROI) of the images were utilized in developing an algorithm for the colorimetric detection. The study which is aimed at developing a point-of-use technique for high accuracy colorimetric analysis utilized the CMOS sensor of the smartphone to obtain the image in a controlled environment. The initial calibration and smartphone application’s computational algorithm was done using MATLAB and XCode respectively. The smartphone application is installed and thereafter utilized to systematically obtain the concentration of target ions in arbitrary solutions. The effect of varying light intensity was studied in effort to simulate fouling of the prototype test cell as well as changes in ambient illumination. Since fouling of the test cell will tend to slightly impede the light intensity as perceived from images, the light intensity study is a good simulation of the fouling dynamics as well as variation in ambient illumination. Also, four different brands of smartphones were evaluated for the calibration of the smartphone-based contaminant monitoring system. Region of Interest studies were also conducted so as to ascertain the optimized ROI for image analysis. Results obtained show that the colorimetric readings in the image chamber of the test prototype, obtained at a light intensity between 30 and 85 Watts/m 2 is fairly constant. However, there is appreciable undulation in the colorimetric reading below 30 Watts/m2. This therefore suggests that the proposed prototype should be operated at light intensity above 30 Watts/m2. Also, the nuance in light intensity above 30 Watts/m2 within the test chamber of the prototype will have little or no effect in the colorimetric readings. The brand studies for smartphones reveals that the colorimetric readings obtained from different brands of phone in the same ambient condition differ. However, there is consistency in the variation pattern for individual smartphones as the concentration of the target ion increases. The ROI studied revealed that the image property within the global ROI is relatively more consistent than ones obtained within local ROI. Typical calibrations were conducted with hexavalent Chromium and Nitrite analytes ranging between 0–250ppb and 0–3000ppb respectively. The calibration equation and coefficient of correlation as obtained for hexavalent Chromium and Nitrite are HS2 = -0.00918 C + 6.46; R2 = 9.997 and HS2 = -0.00584C + 4.85; R2 = 9.823 respectively. The limits of detection (LOD) of this smartphone based colorimetric procedure for Nitrite and Chromium (VI) were approximately 285ppb and 16ppb respectively. The LODs as obtained are lower than the Maximum Contaminant Level (MCL) as stipulated by World Health Organization (WHO) and United Stated Environmental Protection Agency (USEPA). Appreciable agreements between conventional and proposed techniques were obtained with a maximum percentage difference of 8.21% in the concentration of an arbitrary sample.




Nnanna, Purdue University.

Subject Area

Civil engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server