Sampling assumptions about contamination

This volumetric method makes no assumptions about particle shape, and indeed is not greatly affected by particle shape except in extreme cases like flaky materials, such as some clays. To count the number of particles in a known volume of suspension, such as for particulate contamination studies or for a blood cell count, the sample volume is accurately metered by means of a calibrated "manometer". Figure 1 illustrates the original, simple, mercury siphon and metering system. 

In the fractionation scheme shown in Figure 1, fraction A could be contaminated by pentane, benzene, or ethanol fractions B and C are most likely to be contaminated by benzene and fractions D and E are most likely to be contaminated by methanol. The NMR spectrum of fraction A showed ethanol contamination amounting to 7% of the sample, but apparently no pentane was present, as none was picked up in the hexane elutriation while preparing subfraction Al. No methanol was detected in the NMR spectra of fractions D and E. All fractions showed monoaromatic hydrogen, which could result from monoaromatic structures in the original coal, from combined phenol, or from contamination by benzene used in the fractionation scheme. Thus apart from the small amount of ethanol in fraction A, the main constituent of the added material was phenol and/or benzene. Fortunately, because of the structural similarities of these two, the final answer obtained for the coal structure will not be very sensitive to assumptions made about the relative quantities of these two present. 

In the mass titration method, the PZC is determined as the natnral pH of a concentrated dispersion. A detailed description of the experimental procedure can be found in [667], Mass titration become popular in the late 1980s [668,669], but the same method was already known in the 1960s as the pH drift method [183], Usually, a series of natural pH values of dispersions with increasing solid loads is reported, but only the natural pH of the most concentrated dispersion is actually used. The only role of the data points obtained at lower solid loads is to confirm that a plateau was reached in pH as a function of solid load that is, a further increase in the solid load is unlikely to bring about a change in pH. The mass titration method is based on the assumption that the solid does not contain acid, base, or other surface-active impurities. This is seldom the case, thus mass titration often produces erroneous PZCs. In this respect mass titration is similar to the potentiometric titration without correction illustrated in Figure 2.7, only the solid-to-liquid ratio is different. The experimental conditions in mass titration (solid-to-liquid ratio, time of equilibration, nature and concentration of electrolyte, and initial pH) can vary, but little attention has been paid to the possible effects of experimental conditions on the apparent PZC. The effect of an acid or base associated with solid particles on the course of mass titration was studied in [670], To this end, a series of artificially contaminated samples was prepared by the addition of an acid or base to a commercial powder. The apparent PZC of silicon nitride obtained in [671] by mass titration varied from 4.2 (extrapolated to zero time of equilibration) to 8.2 for time of equilibration longer than 20 days. The method termed mass titration was used in [672], but it was different from the method discussed above. 

The analysis of model biochemical compounds [glucose derivatives, poly(amino acids), etc.] with the SSX 100/206 spectrometer at a pass energy of 50 eV has shown that two systematic errors may compensate each other. The presence of organic contamination at the sample surface reduced the apparent 0/C ratio, because the contamination overlayer contained less oxygen than most of the analyzed standards. On the other hand, the assumption of a constant transmission function exagerated (by about 10%), the computed 0/C concentration ratio. The two errors compensated each other and the apparent 0/C concentrations ratios were in excellent agreement with the values expected from stoichiometry. 

Also, it would be exceedingly rare that we are interested in the contaminants contained in the specimens themselves. These have been removed from the field and in most cases will be disposed of by the laboratory, not returned to their original locations. An assumption that underlies the method as described is that the sample of specimens is representative of conditions in the field, and statistical methods can be employed in attempt to ensure that a representative sample has been obtained. But in reality this methodology can only ever give definitive information about the individual specimens themselves. For practical reasons, soil specimens are relatively small in lateral extent and mass (they are unlikely to be greater than about 10 cm in diameter and 2 kg mass). Any conclusions about the greater mass of soil left behind have to be inferred using statistical methods. 

---