Capability for Determination of Chemical Composition

Any meaningful discussion of current capabilities in the measurement of composition must focus on the real world of practical materials. 

In 1933 G. E. F. Lundell, then the Chief Chemist of the National Bureau of Standards, wrote a very candid article entitled, The Chemical Analysis of Things as They Are. He said he used that title because so many articles on analysis appearing in the literature dealt with the chemical analysis of things as they are not.  

Much the same problem exists in the materials analysis field today. The scientific literature contains numerous references to materials of 5-9 s and 6-9 s purity. Such statements are misleading and of questionable validity because the estimates are in many cases derived from resistivity measurements supplemented by emission spectrographic analyses (with a sensitivity of only 1-10 ppm for most elements). To establish that a sample of material contained less than 1 ppm total impurities would require analyses for all elements present by techniques with sensitivities and accuracies in the range 1-20 ppb. 

The magnitude of effort required is illustrated by the classic work on zone-refined aluminum by Albert of the CNRS Laboratory in Paris. Samples were analyzed for over 60 elements plus the rare earths by high-sensitivity neutron activation using detailed radiochemical procedures. Elements such as carbon, oxygen, and nitrogen were determined by photonuclear or charged-particle activation. This procedure required the efforts of a four-man team for 12 hr, an additional person for nine days, and another person for two weeks to analyze the rare earths. After the amounts of individual contaminants were totaled (for many there were only experimental upper limits of 1-10 ng), it was possible to establish that a particular sample of aluminum contained less than 2 ppm total impurities (i.e., it was not quite 6-9 s pure). 

Unfortunately, little confidence can be placed in most general statements of purity, such as 6-9 s. It is imperative that analytical information be developed from more discriminating techniques. 

---