Whenever possible, solid substance should be analyzed in the solid state. This dispenses with the often laborious dissolution process, particularly for solid substances which are chemically resistant such as eg. precious metals, ceramics, and hard metals. The dissolution process often requires a great deal of time; it may lead to contamination and it impairs the LOD according to the degree of dilution. Particular in the case of solid substances with a uniform type of bonding, and usually when a reference from the sample can be used, chemical dissolution is not necessary.

  When working with solution it is often possible, however, to calibrate and recalibrate with synthetic solutions, which is particularly advantageous if no calibration samples similar to materials to be analysed are available, or if their cost is exorbitant.

  Direct solids introduction：

  Direct sample introduction takes place without any change in physical state. Whereas direct delivery of gas and liquid samples to flames and plasmas is based on tried and tested techniques, direct delivery of solid samples and particularly of non-conductors is not “standardized ”. Direct input only gives usable results if the analytes in different sample are present in the same chemical and mineral bond, and a reference from the sample can be used. Direct input will also be successful if only one assessment is to be made on the level of individual analytes eg. an easily vaporizable analyte in a matrix which is vaporize.

  In all other cases direct introduction has not yet got past the experimental stage and will possibly never get beyond it! In the author’s opinion, a two-stage technique must be used for most “real ” nonconductive solid samples. The sample must be available as a powder and must be isoformed in the first stage ie. It must be converted into a uniform type of bonding independent of the form of bonding in the initial sample and, so far as possible, in the form of a gas eg. as fluorides. The other, less elegant way is isoforming in a previous stage by converting it eg. into borates, as has long been known and practiced. The fused product must be finely ground and this powder can be introduced directly to the plasma.

  There is a great deal of interest in direct analysis of solid substances, Conversion into a solution would involve additional work which, with certain materials is expensive. The time required for this does not allow control of process. In addition there is dilution in the sample and the risk of contamination. Because of the increased use of ICP for liquids analyses most sample delivery systems are centred on ICP.