Instrument optimal bandwidth

The instrument variables Rs, RB, and Rs + 2RB are used in instrument optimization for example, an improved matching of the laser bandwidth to the HO absorption could increase Rs, a reduction in illumination of walls near the detection zone by ambient light or scattered or diffracted laser light could decrease RB, and an increase in photon collection efficiency could increase (Rs + 2RB). The remaining quantities fav, MAT, SNR, and MDC may be traded off during data processing, but the choice of their values is restricted by the instrument variables. 

The bandwidth values in the table are those calculated for the total system the instrument plus the column. The values for number of plates are for the number of plates realized in the total system. It can be seen that the optimized system does not greatly impact column efficiency, the total loss in plates being only about ten percent at total exclusion for a 24,000 plate column. This is consistent with an instrumental bandwidth equal to a third of the bandwidth of the column. The conventional system, with a bandwidth equal to or greater than that of the column, exhibited a severe loss in realized efficiency, particularly at or near exclusion. 

The second reason for the appeal of Fourier transform methods is cost. While the cost for a large, visible/UV Fourier transform instrument is high, two factors should be kept in mind. First, less expensive instruments can be constructed which will provide good spectroscopic data as shown by Horlick and Yuen. Second, the true cost to be considered is the cost per spectrum. Since, in general, Fourier transform instruments can collect broad bandwidth spectra relatively quickly, many more spectra can usually be collected by this method than by conventional dispersive methods. Since spectra can be collected much more rapidly than they can be analyzed, it is not unrealistic to think that a single instrument could serve many research groups. This method of joint or shared operation keeps the instrument in a state of use which optimizes its efficiency, and spreads the cost over many users. 

Altogether, the results deliver a clear answer to the question, what is the optimal instrumental bandwidth for CS AAS. The bandwidth should never be chosen smaller than two times the FWHM of absorbing lines. Higher values for AAinstr will not significantly reduce the shot-noise limited Amin- Nevertheless they should be avoided to allow the best possible background correction (BC). With respect to the results presented in Section 2.1.5, an optimized value for the instrumental resolving power / = A / AA is between 50 000 and 150000. 

---