P. Ljung et al., DETECTION OF TITANIUM IN ELECTROTHERMAL ATOMIZERS BY LASER-INDUCED FLUORESCENCE .2. INVESTIGATION OF VARIOUS TYPES OF ATOMIZERS, Spectrochimica acta, Part B: Atomic spectroscopy, 52(6), 1997, pp. 703-716
Different types of electrothermal atomizers (graphite and tungsten fur
naces) have been studied for the determination of Ti by laser-induced
fluorescence (LIF). It was found that all graphite furnaces suffer fro
m both bulk contamination by Ti, corresponding to tens of pg per firin
g, and significant memory effects in such a way that the high sensitiv
ity of the LIF technique could not be fully utilized. The detectabilit
y of Ti by the LIF technique was limited by fluctuations of the atomiz
er blank signals and therefore depended mainly on the Ti history of th
e furnaces. The detection limit of Ti was found to be similar for all
graphite furnaces: around 1 pg for furnaces not exposed to any substan
tial amounts of Ti, and a few pg for those which had been exposed to a
few hundreds of pg of Ti (or more). It was found that the transversel
y heated graphite atomizer (THGA) from Perkin-Elmer gives the most rep
roducible atomization of Ti. The tungsten furnace showed no memory eff
ects and therefore no empty firings (bum-outs) between determinations
are needed. The tungsten furnace, however, afforded significantly weak
er signals (approximately 250-fold) than the graphite furnaces; this w
as attributed to incomplete atomization processes. The Ti signal from
the tungsten furnace could be increased by the addition of H-2 in the
buffer gas. For 20% of H-2 added, however, the analyte signal was stil
l weaker by a factor of 8 than that seen in a graphite furnace shielde
d with Ar (the remaining difference is primarily attributed to fluores
cence quenching from H-2) In addition, the precision of the measuremen
ts was poorer than that obtained from graphite furnaces. (C) 1997 Else
vier Science B.V.