Inside the world of chemometrics, a fundamental role is played by the experimental design. Although introduced almost a century ago (1935), it is still not widely employed by chemists and its usefulness continues to be underestimated. When asking why, the answers are often the following: it is too difficult to apply and too much experimental effort is required. Actually, a deeper knowledge on the topic could offer a different point of view. The aim of the present tutorial is to introduce the experimental design to beginners, by providing the theoretical basic principles as well as a practical guide to use the most common designs, from the experimental plan to the final optimization. Response surface methodology will be discussed, and the main terms related to model computation and statistical evaluations usually performed by software will be explained, in order to give suitable tools to properly use them.

Relevant 2004–2018 papers on arsenic speciation analysis of environmental samples have been reviewed, with the focus on the analytical procedures according to the matrix investigated and the species researched.

A new procedure for the determination of iron in sea-water at sub-nanomolar concentration levels was developed. The method applied the low-blank magnesium hydroxide co-precipitation procedure in combination with quadrupole ICP-MS and used the dynamic reaction cell technique to resolve the polyatomic interferences arising from the residual matrix and the solvent. Although the interference-free determination at m/z = 56 could be obtained by using methane as the reaction gas, the best signal-to-background ratio was achieved by using ammonia and performing the measurements at m/z = 54. Accurate quantification of low levels of iron at m/z = 54 also required the mathematical correction for the isobaric interference due to the occurrence of chromium. The main analytical figures of the optimized method were evaluated, with special attention to the determination of the procedural, reagent and field blanks. A detection limit as low as 0.09 nM was achieved, using a pre-concentration ratio of 10. The accuracy of the analytical procedure was evaluated by the analysis of the sea-water reference materials CASS-4, NASS-5, SAFe D2 and SAFe S1, having certified concentration values ranging from 12.8 nM to about 0.1 nM. The analytical precision for replicated analyses of the CRMs ranged from 1.3% to 14% (n = 7). Finally, the developed procedure was applied to a number of open-ocean sea-water samples from the Ross Sea (Southern Ocean, Antarctica)

The behaviour of rare earth elements (REEs) ions within a dynamic reaction cell (DRC) pressurized with oxygen was investigated, in order to explore their determination at m/z +16 by inductively coupled plasma mass spectrometry (ICP-MS). Under the optimized conditions, the conversion of REEs to molecular monoxide ions was nearly quantitative (>96%), with the exception of Tm (78%), Eu (11%) and Yb (6%). Moreover, the formation of dioxide ions was generally lower than 1%. According to these results, a new method for the determination of REEs in digests from geological samples was developed. After microwave-assisted acid digestion, the solutions were directly analyzed by ICP-MS, determining REEs at m/z +16. The interferences due to dioxide and un-reacted ions were carefully evaluated by taking into account the actual concentration of REEs in sediment samples. In addition, the possible interferences due to other constituents of sediments (e.g. barium) were considered. The accurate determination of Y, La, Ce, Pr, Nd, Sm, Tb, Dy, Ho, Er and Tm was achieved, as demonstrated by the analysis of both certified reference materials and Antarctic marine sediment samples, using ICP sector field mass spectrometry (ICP-SFMS) for comparison. On the other hand, Eu, Gd, Yb and Lu could not be measured by the O-atom addition approach and their separate determination in the standard mode using mathematical correction is hence necessary to obtain the complete REE pattern. The quantification limits (10 times the standard deviation of 10 procedural blanks) were at least two orders of magnitude lower than the REEs concentrations in sediment samples. Precision of the procedure ranged from 1.6% to 6.4%.