Concentration measurement of yeast suspensions using high frequency ultrasound backscattering

This work proposes the use of an ultrasound based technique to measure the concentration of yeasts in liquid suspension. This measurement was achieved by the detection and quantification of ultrasonic echoes backscattered by the cells. More specifically, the technique was applied to the detection and quantification of Saccharomyces cerevisiae. A theoretical approach was proposed to get the average density and sound speed of the yeasts, which were found to be 1116 kg/m(3) and 1679 m/s, respectively. These parameters were needed to model the waves backscattered by each single cell. A pulse-echo arrangement working around 50 MHz, being able to detect echoes from single yeasts was used to characterize experimentally yeast solutions from 10(2) to 10(7)cells/ml. The Non-negative Matrix Factorization denoising technique was applied for data analysis. This technique required a previous learning of the spectral patterns of the echoes reflected from yeasts in solution and the base noise from the liquid medium. Comparison between pulse correlation (without denoising) and theoretical and experimental pattern learning was made to select the best signal processing. A linear relation between ultrasound output and concentration was obtained with correlation coefficient R(2)=0.996 for the experimental learning. Concentrations from 10(4) to 10(7)cells/ml were detected above the base noise. These results show the viability of using the ultrasound backscattering technique to detect yeasts and measure their concentration in liquid cultures, improving the sensitivity obtained using spectrophotometric methods by one order of magnitude.