Qualitative Reproducibility

A high degree of qualitative reproducibility, that is the continuous presence of the same features in analysis patterns obtained from the same sample, is a prime requirement for any fingerprinting technique (ref. 107). In our experience, Curiepoint Py-MS techniques show excellent qualitative reproducibility for all kinds of complex biomaterials analysed so far. Similar conclusions were drawn by Hickman and Jane from a comparative study of three different Py-MS techniques in the analysis of various polymeric materials (ref. 121). Considering the large number of different, possibly even competing, pyrolysis reactions that occur in highly complex biomaterials, the stability of the pyrolysis patterns of such materials, as also observed in Py-GC (ref. 122), is an intriguing phenomenon. In fact, many deliberate changes in pyrolysis conditions, e.g. a 15-fold increase (from 0.1 -1.5 s) in ty (refs. 100, 101), fail to change the qualitative composition of the pyrolysate, provided that the occurrence of secondary reactions is minimized and is high enough to ensure complete pyrolysis within ty. Marked qualitative changes in pyrograms, e.g. due to formation of recombination products, may occur, however, when samples become excessively large, or when pyrolysis is carried out inside a capillary tube rather than on the surface of a filament. Drastic changes in pyrograms can also be expected to occur if the selected Tgq is so low that the half-lives of degradation reactions become longer than ty, allowing isothermal pyrolysis processes to take place at, or close to, Teq of the filament. Windig et al. (ref. 109), who applied factor analysis (see Section 6.6.2) to the qualitative evaluation of the influence of changes in a number of pyrolytic and mass spectrometric conditions, demonstrated that an increase in Tgq (358 - 770°C) for polysaccharide materials results in a stronger degradation of the compounds to small, less characteristic fragment molecules, whereas the effect for a typical protein is completely different, viz. the formation of more strongly dehydrogenated pyrolysis products. Since qualitative changes in the pyrolysates are brought about only by such gross, deliberate changes in pyrolysis conditions, such changes hardly ever pose a significant problem under controlled analytical conditions.

A problem with regard to qualitative reproducibility may be encountered when a high instrument background is present due to insufficient trapping efficiency of the liquid nitrogen-cooled screen or unusually heavy contamination of the expansion chamber or ion source. However, Curie-point Py-MS systems such as that shown in Figure 20 are specifically designed to eliminate such problems as rigorously as possible. Naturally, even a low instrumental background may strongly influence the pyrolysis patterns if the amount of sample pyrolysed is too low to provide an adequate signal-to-background ratio. For the type of instrument shown in Figures 18 and 20, the minimum sample size for obtaining satisfactory signal-to-background ratios is approximately 1 microgram for substances yielding complex pyrolysates, e.g. biopolymers (see Figure 21), as compared with 10 ng or less for compounds which yield a single dominant component, e.g. polystyrene (see Figure 1). Some types of sample tend to give low signal yields, especially samples containing relatively large amounts of inorganic constituents, e.g. whole soil samples, compounds forming large

Figure 21. Pyrolysis mass spectra of (a) background produced by heating a clean wire (b) and (c) 1 and 5 yg, respectively, of a galacto-gluco-glucurono-araban. Note that the spectrum of the 1 yg sample still shows an overall signal-to-background ratio of better than 10:1. Further note that the intensity of the 5 yg spectrum is only about 4 times that of the 1 yg spectrum, indicating possible inaccuracy of actual sample size.

Figure 21. Pyrolysis mass spectra of (a) background produced by heating a clean wire (b) and (c) 1 and 5 yg, respectively, of a galacto-gluco-glucurono-araban. Note that the spectrum of the 1 yg sample still shows an overall signal-to-background ratio of better than 10:1. Further note that the intensity of the 5 yg spectrum is only about 4 times that of the 1 yg spectrum, indicating possible inaccuracy of actual sample size.

amounts of char, e.g. condensed polynuclear structures such as are found in high rank coals, or compounds evaporating intact and condensing on the relatively cold reaction tube wall, such as cholesterol and other relatively volatile compounds. With such samples, low signal-to-background ratios may be obtained in spite of a seemingly adequate sample size.

0 0

Post a comment