Researchers have created artificial neural networks that can distinguish between different kinds of tea leaves - most people can't do that. But they do it by analyzing the mineral content.
Their method makes it possible to distinguish between the five main tea varieties (white, green, black, Oolong and red) using chemometrics, a branch of chemistry that uses mathematics to extract useful information from data obtained in the laboratory.
The concentrations of the chemical elements in the leaves were determined using 'inductively-coupled plasma atomic emission spectroscopy', which showed the most abundant elements to be calcium, magnesium, potassium, aluminium, phosphorus and sulfur. Other essential elements were also identified in the tea, such as copper, manganese, iron and zinc, according to their paper.
Once the mineral content of the leaves was established, probabilistic neural networks were used to find out which type of tea a sample belonged to. These networks are "mathematical algorithms that mimic the behaviour of the neurons in the human nervous system in order to process the information", they say.
"This method makes it possible to clearly differentiate between the five types of tea – something that is often not easy to do by eye alone – by using analysis of the leaves' mineral content and then mathematically processing these data," José Marcos Jurado, co-author of the study told Servicio de Información y Noticias Científicas (SINC).
Can you identify the 5 common types of tea? An artificial neural network can. Credit: J. Marcos Jurado et al.
This generates a model that receives an input signal (chemical data) and produces an output one, making it possible to predict the type of tea in the sample with a probability of 97%.
The differences for people are subtle but chemically noticeable. While four of the most common tea types is from the plant Camellia sinensis, the process is different for each. White tea is a non-fermented tea made of new buds and leaves that are protected from sunlight to limit chlorophyll production. Green tea is also unfermented but is made by using older green leaves. Oolong and black tea varieties are made by fermenting the leaves, though in the first case these are completely fermented, while black tea undergoes an intermediate controlled fermentation process of between 10% and 70%. Red, or Pu-erh, tea is a fermented product obtained from another variety of the plant, Camellia sinensis var assamica, which is cultivated in the Chinese region of Yunnan.
Citation: James S. McKenzie, José Marcos Jurado y Fernando de Pablos, “Characterisation of tea leaves according to their total mineral content by means of probabilistic neural networks”, Food Chemistry 123 (3): 859–864, 2010, DOI: 10.1016/j.foodchem.2010.05.007.
Artificial Neural Network Can Distinguish Between Tea Leaves