Question

Why is it that temperature fluctuations during atomization (i.e. in the plasma or flame) are a big problem for atomic emission spectroscopy, but less of a problem for atomic absorbance spectroscopy? Include an equation in your answer.

Answer 1

Influence of Flame Temperature in Atomic Emission and Atomic Absorption Flame Spectrometry

Currently, it is generally agreed that atomic absorption flame spectrometry is complementary rather than competitive to the older method of flame emission spectrometry. It is readily possible to compare these two flame spectrometric methods because of the similarity in instrumentation necessary for measurements. Comparison of these methods is often made, but it appears that some confusion exists as to the role of flame temperature in these flame spectrometric methods. This confusion is particularly great when reading some review articles on atomic absorption flame spectrometry (2, 3). Therefore, the influence of flame temperature on the emission signal in atomic emission flame spectrometry and on the absorption signal in atomic absorption flame spectrometry will be considered in this manuscript. The arguments mainly used by other authors to indicate the influence of flame temperature on results in atomic emission and atomic absorption flame spectrometry are summarized in the following statements in this paragraph. The absorption signal (generally absorbance A = log Io/I where lo is the intensity of radiation transmitted through the flame gases with only blank solution being aspirated, and I is the intensity of radiation transmitted through the flame gases with sample solution being aspirated) is proportional to the number of atoms in the ground atomic state per cubic centimeter of flame gases, whereas the emission signal (emission intensity) in atomic emission flame spectrometry is proportional to the fraction of this number in a particular excited state. According to the Boltzmann law (assuming the flame is approximately in thermal equilibrium), this fraction is proportional to exp -Eu/kT, where Eu is the excitation energy of state u (the upper state of the transition), T is the flame temperature in ' K., and k is the Boltzmann constant. Therefore, it is evident that this fraction will generally be much less than unity-e.g., which corresponds to one excited atom per 106 total atoms in the ground stateand that this fraction will be very sensitive to changes in flame temperature. Consequently, it would appear that the absorption signal should respond to smaller solution concentrations of the element, in concern and should be virtually independent of flame temperature. Both of these conclusions, which are often made in the literature are incorrect because of the omission of certain basic considerations. A more thorough and accurate discussion of the influence of flame temperature in the flame spectrometric methods will be given in the following paragraphs, The existence of a significantly greater number of atoms in the ground state than in excited states does not necessarily mean that the limits of detection in atomic absorption flame spectrometry will be lower than in atomic emission flame spectrometry. Indeed, a correct calculation of the limit of detection must be based on the signal-to-noise ratio rather than only on the magnitude of the signal (6, 6). Variation in flame temperature does not generally constitute the main contribution to the noise in atomic absorption flame spectrometry. The reader is referred to the articles by Winefordner and Vickers (6, 6) for a discussion of the influence of flame temperature on limits of detection. In this manuscript primarily the influence of temperature on the measured signal will be considered.

THEORY The absorbance, A, and the emission intensity, I_E, are related to the ground state atom concentration, N_{0 ,}according to the following expressions A = log (l₀/l) = C_AN₀/_D   (1)

I_E = C_EN₀exp(-Eu/kT), (2)

where the factors C_Aand C_E include such factors as wavelength, transition probability, At, statistical weights, and the atomic partition function, B. It must be noted that for some transition metals and rare earth elements the change of the partition function between 2000' and 3000' K. may amount to 10%. Because major influences of the temperature are of primary interest here, this variation will be ignored. Consequently, the factors C_Aand C_Eare considered to be independent of the flame temperature for this discussion. In Equation 1,_D is the Doppler half width of the absorption line and 6 accounts for the decrease in absorption signal caused by broadening of the absorption line and finite width of the source line. In the ideal case = 1- i.e., measurements are made at the wavelength peak of the spectral lime and Doppler broadening is the major source of absorption line broadening and so the temperature dependence of /_D reduces to that of the Doppler half-width, which is proportional to T^1/2 .