The Rapid Identification of Solid Materials Using the ACP Method ^†

Abstract

In this work, we will present a new protocol to identify solid matter quickly, and at low cost, by physico-chemical analysis methods. The fast and less expensive methods for identification that we investigate, using principal component analysis (ACP), are hydrogen potential (pH), electrical conductivity, density and solubility.

1. Introduction

Principal component analysis (ACP) is one of the multivariate data analysis methods, and is very useful when there are large amounts of quantitative data to process and interpret. ACP is a tool, whereby it is possible to make correlations between different variables by their projections in reduced spaces made up of axes designated as principal components [1]. The methods for identifying solid compounds include Infrared Spectroscopy, Scanning Electron Microscopy and X-ray Diffraction. However, these methods are very expensive and difficult to handle. In this article we study and validate different rapid techniques using ACP, that facilitate identification of solid compounds at low cost [2,3,4]. The identification methods which will be studied are solubility, hydrogen potential (pH), electrical conductivity and density.

2. Materials and Methods

The study was carried out on four phosphate fertilizer materials purchased from Sigma-Aldrich to test and validate the method for each material We prepared four series of measurements under repeatability conditions. These four series corresponded, respectively, to hydrogen potential (pH), electrical conductivity, density and solubility [5]. In order to determine the solubility of the material, we prepared a saturated solution of the sample. Then, the prepared solution was filtered under vacuum. The solid part of the extraction was recovered and put in an oven to dry until solid. Finally, the solubility was the ratio of the soluble mass to the volume [6]. To measure the pH and the electrical conductivity of the samples, we measured the solubility pH and the conductivity of the filtra (saturated solution), respectively, by making use of a pH-meter (PREMIUM) and a conductivity meter [7,8]. The density was measured by filling a cylinder (test tube) with a determined volume, having a determined mass, and calculating the density as mass over volume [9].

3. Results and Discussion

At the end of validating the identification, the solubility, the pH, the electrical conductivity and the volumetric mass, we calculated the coefficient of variation (CV (%)) and the lower and upper repeatability limits for each series of measurements, with a probability of 95% [10]. The results show that for the four phosphate fertilizers and for the four methods, all the values did not exceed the lower and upper values of repeatability. Furthermore, the relative error in all the cases was lower than 5%. Then, the repeatability of the four methods was checked and validated. After checking the validation of the four methods, we applied the ACP method using the results of measurements for the different materials, with the aim of immersing two relevant pieces of information. For each material, ten measurements were included in the analysis and were introduced without any specification of the type considered. The samples were numbered from 1 to 40 and the matrix, thus formed. contained 160 pieces of data to be processed.

Figure 1 presents the eigenvalues obtained in the data processing. The first two main axes mainly carry the relevant information representing about 98% of the cumulative variability [11]. In this case, the ACP application allowed us to project the 160 pieces of data on a two-dimensional space.

Figure 2 shows that the four variables are well represented in the space of the two components F1 and F2. From this figure the representation of the samples shows four distinct groups; each linked to the type of material considered. The possibility of discriminating between the types of materials must be exploited in the attribution of any type of sample analyzed.

4. Conclusions

In this work, the following physico-chemical parameters were measured for four types of phosphate fertilizer materials, with the aim of being able to set up a rapid identification method for these materials: solubility, pH, conductivity and density. Validity tests applied to the results of these measurements validated the four methods. Data from repeatability tests were processed using the Principal Component Analysis method. We noticed that the tests of each type of material formed a group. Therefore, thanks to the analysis of the principal components (PCA) we were able to discriminate any unknown analyzed sample and were, therefore, able to recognize it and easily attribute it to the type of matching material, quickly and unambiguously.