Pp. Woskov et al., Accurate and sensitive metals emissions monitoring with an atmospheric microwave-plasma having a real-time span calibration, WASTE MAN, 20(5-6), 2000, pp. 395-402
Both high sensitivity and good measurement accuracy are required in a metal
s continuous emissions monitor in order for it to be acceptable for complia
nce monitoring. An atmospheric plasma sustained by microwaves with an attac
hed source of a calibrated trace metals aerosol has been shown to be capabl
e of achieving both of these requirements. The microwave plasma is continuo
us and operates in undiluted stack exhaust for atomic emission spectroscopy
of trace metals. The plasma is sustained in a shorted waveguide that is at
tached to the stack by a short sample line (< 50 cm). It is powered at 1.5
kW, at a frequency of 2.45 GHz. An undiluted stack slipstream is isokinetic
ally drawn into the plasma by a suction pump at a nominal flow of 14 l/min.
The pneumatic nebulizer attached to the sample line can momentarily, on co
mmand, inject a known concentration of metals solution providing a real-tim
e span calibration. The system was tested on the exhaust stack of the rotar
y kiln incinerator simulator facility at the Environmental Protection Agenc
y (EPA) National Risk Management Laboratory in Research Triangle Park. Thre
e hazardous metals were monitored, lead, chromium, and beryllium. These mea
surements were referenced to EPA Method-29. A total of 20 spiked stack exha
ust tests were carried out. Ten one-hour tests at high concentration (40-60
mu g/actual m(3)) and ten one and half-hour tests at low concentration (10
-15 mu g/actual m(3)). The microwave plasma monitor achieved relative accur
acies of approximately 20% for lead and beryllium and 40% for chromium with
a threshold detection capability of < 3 mu g/actual m3 for a time response
of similar to 1 min. The relative accuracy deviation from the EPA Method-2
9 was found to be mostly systematic. This suggests the possibility of using
a site-specific calibration to bring the microwave plasma into compliance
with EPA's goal of 20% relative accuracy to the reference method. Laborator
y work is continuing to add mercury, arsenic, and cadmium to the monitored
metals. Mercury and arsenic present a particular challenge to achieving hig
h detection sensitivity in undiluted stack exhaust because the plasma is le
ss efficient in exciting Hg and because UV absorption interferes with As de
tection. This is a problem that is also in common with air ICP and in situ
spark plasma methods for continuous emissions monitoring of metals. (C) 200
0 Elsevier Science Ltd. All rights reserved.