Discriminating among different tea leaves using an operating temperature-modulated tin oxide gas sensor

Abstract

We report distinguishing different types of tea leaves from each other based on their aroma using a thermal shock-induced generic tin oxide gas sensor. The sensor used in this work consists of a microheater and a tin oxide pellet, both connected to outside circuitry with noble metal contacts. The heater is powered with a series of narrow high magnitude voltage impulses of predetermined thermal impacts adjusted to produce step-like temperature rises of different magnitudes on the gas sensitive pellet. The sensor is exposed to aromas collected from various types of tea leaves at different concentrations. Within 4.5 s, nine 500 ms-wide voltage pulses, each as high as 9.3 V in magnitude, are applied to the microheater. Each pulse causes a step-like temperature jump on the pellet temperature. The transient responses recorded for different tea leaves look different even after amplitude normalization. The sensor profiles are recorded, digitized, and compared with the database of previous experiences. A heuristically defined high dimensional feature vector is automatically generated for each analyte. Classifications are graphically achieved in a 3-D feature space after applying principle component analysis for dimension reduction.