Total-reflection X-ray fluorescence analysis (TXRF) is nowadays a well-established technique for the multielemental analysis of different types of samples, from water solutions to ores. The rapid diffusion of the technique is mainly due to the advantages of requiring very low amount of sample (in the µL or µg range), and a reduced sample preparation. For the analysis, samples are usually directly deposited as solutions or liquid suspensions on suitable supports (reflectors). In the case of suspensions, several approaches for sample preparation (e.g., considering different amounts of sample mass, dispersant agent volume, internal standard concentration) have been developed for different types of samples and matrixes. The accuracy of the method is usually assessed by analysing suitable certified reference materials, and the recovery is the parameter most often adopted to check the quality of the analysis. However, through this approach, the identification of the best sample preparation procedure is not always guaranteed. In fact, such classical univariate approach cannot provide a full understanding of the effects of the parameters’ variation on the result of the analysis. Indeed, it has been demonstrated that the combination of factors can influence the quality of the analysis more than the single parameters. For this reason, a multivariate experimental design should be applied for the identification of the effect of each factor and their combinations on the analytical performances. In this work, a full factorial experimental design was applied to develop the most suitable protocol for the preparation of different types of soil organic matter and amendments prepared as suspensions for TXRF analysis. Two factors were considered for sample preparation: sample mass (10– 100 mg) and dispersant volume (1 – 5 mL of 1% Triton X-100 solution). The analyses were performed with two different Bruker S2 Picofox TXRF spectrometers equipped with a Mo and a W sources, respectively. In the first step, the study was done on the certified reference material NIST 1635a and using the recovery as monitoring parameter. The best results were obtained by dispersing 10 mg of material in in 3 or 5 mL of dispersant solution. Secondly, the validation of the two preparations was performed using a second certified reference material (NIST 1573a) and evaluating the recovery, relative standard deviation, limits of detection and quantification. Based on these parameters, 10 mg in 3 mL was considered the best sample preparation procedure and was applied to different soil organic matter and amendment samples. The set of samples is composed of three humic acids, three humates, three peats, two biochars, one digestate and one compost. The obtained results are statistically similar to those obtained using ICP-OES, demonstrating the good analytical performances of the new developed method.
Sample preparation customization for TXRF analysis using a multivariate experimental design: the case of soil organic matter and amendments
Zaccone C.;
2023-01-01
Abstract
Total-reflection X-ray fluorescence analysis (TXRF) is nowadays a well-established technique for the multielemental analysis of different types of samples, from water solutions to ores. The rapid diffusion of the technique is mainly due to the advantages of requiring very low amount of sample (in the µL or µg range), and a reduced sample preparation. For the analysis, samples are usually directly deposited as solutions or liquid suspensions on suitable supports (reflectors). In the case of suspensions, several approaches for sample preparation (e.g., considering different amounts of sample mass, dispersant agent volume, internal standard concentration) have been developed for different types of samples and matrixes. The accuracy of the method is usually assessed by analysing suitable certified reference materials, and the recovery is the parameter most often adopted to check the quality of the analysis. However, through this approach, the identification of the best sample preparation procedure is not always guaranteed. In fact, such classical univariate approach cannot provide a full understanding of the effects of the parameters’ variation on the result of the analysis. Indeed, it has been demonstrated that the combination of factors can influence the quality of the analysis more than the single parameters. For this reason, a multivariate experimental design should be applied for the identification of the effect of each factor and their combinations on the analytical performances. In this work, a full factorial experimental design was applied to develop the most suitable protocol for the preparation of different types of soil organic matter and amendments prepared as suspensions for TXRF analysis. Two factors were considered for sample preparation: sample mass (10– 100 mg) and dispersant volume (1 – 5 mL of 1% Triton X-100 solution). The analyses were performed with two different Bruker S2 Picofox TXRF spectrometers equipped with a Mo and a W sources, respectively. In the first step, the study was done on the certified reference material NIST 1635a and using the recovery as monitoring parameter. The best results were obtained by dispersing 10 mg of material in in 3 or 5 mL of dispersant solution. Secondly, the validation of the two preparations was performed using a second certified reference material (NIST 1573a) and evaluating the recovery, relative standard deviation, limits of detection and quantification. Based on these parameters, 10 mg in 3 mL was considered the best sample preparation procedure and was applied to different soil organic matter and amendment samples. The set of samples is composed of three humic acids, three humates, three peats, two biochars, one digestate and one compost. The obtained results are statistically similar to those obtained using ICP-OES, demonstrating the good analytical performances of the new developed method.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
