Full text
| |
| Author(s): |
Tiago Pucca Araujo
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Química |
| Defense date: | 2007-05-25 |
| Examining board members: |
Ronei Jesus Poppi;
Ieda Spacino Scarminio;
Adriana Vitorino Rossi;
Jarbas José Rodrigues Rohwedder
|
| Advisor: | Ronei Jesus Poppi |
| Abstract | |
This work treats of the application of chemometrics methods in the development of new analytical methodology, without the aid of separation methods or pre-concentration, in the direct analysis of tetracycline in bovine milk. For this purpose, it was used the Infrared Spectroscopy with transformed of Fourier (FT-IR) and the chemometrics methods PLS, iPLS and SIMCA. Samples of skimmed milk were doped with progressive amounts of Tetracycline, Chlortetracycline and Oxitetracycline, three of the antibiotics of the class of Tetracycline more commonly used in veterinary medicine and, therefore, with larger risk of contamination of the milk destined to human consumption. This procedure was adapted to produce samples containing each one of the mentioned tetracycline and also samples containing the binary combination of the same ones. For samples doped with a tetracycline type, it was obtained a matrix with 40 samples, and a matrix with 36 samples for samples containing the binary combination of the three tetracyclines. The spectral range used for the spectra obtaining was 900 to 1800 cm, corresponding to the Medium Infrared region (MIR - Mid Infrared). The SIMCA method was used in the identification of samples, classifying them as belonging to the mentioned matrix. The success degree in the classification was of 100%. For the identification of the samples, the next step was the development of the procedure of quantification of the tetracyclines in milk using the PLS method. The range of dope of the samples was from 20 to 500 ppb (parts per billion), including a wide range, considered by the Brazilian legislations, American and European Union, because initially there was not knowledge of the limit of detection of the method. The iPLS was used for the identification and selection of the spectral range with better correlation with the amount of presented antibiotic in the sample. The obtained results were promising considering the low concentration that was intended to detect: for concentration above 100 ppb, the errors obtained were, on average, less than 13% and for concentration below 100 ppb were, on average, less than 18%. In that way, the developed method, in spite of not being strictly quantitative, assists to the objectives of fast detection, without pre-treatment of the samples and generation of chemical residues, allowing the identification, in an unknown sample of milk, of the residue of present tetracycline. (AU) | |