EPA-R2-72-121
ARKON Sci.n«mc
Labs
DEVELOPMENT OF AN INFRARED FLUORESCENT
GAS ANALYZER
By
E.A. McClatchie
ARFR-14
August 1972
Distribution of this report is provided in the interest of in-
formation exchange. Responsibility for the contents resides
with the author or organization that prepared it.
The work upon which this publication was based was performed pur-
suant to Contract No. CPA-70-152 with the National Air Pollution
Control Administration by:
ARKON SCIENTIFIC LABS
O3O OwIgM Way. Berkeley. Calif. 94710
(415) 849-1377
-------
DEVELOPMENT OF AN INFRARED FLUORESCENT
GAS ANALYZER
By
E.A. McClatchie
ARFR-14
August 1972
Distribution of this report is provided in the interest of in-
formation exchange. Responsibility for the contents resides
with the author or organization that prepared it.
The work upon which this publication was based was performed pur-
suant to Contract No. CPA-70-152 with the National Air Pollution
Control Administration by:
ARKON SCIENTIFIC LABS
930 Dwight Way
Berkeley, California 94710
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION ±
2. THEORY OF OPERATION 1
3. RESULTS TAKEN WITH THE PROTOTYPE ANALYZER 4
4. CONCLUSIONS 7
-------
1. INTRODUCTION
This report covers the work performed by Arkon Scientific
Labs under Contract CPA-70-152 to the National Air Pollution Con-
trol Administration (NAPCA).
The aim of this contract was to develop a prototype model
low level carbon monoxide analyzer using the Arkon Scientific Labs
fluorescent cell and negative chopping techniques to achieve a de-
vice superior to state of the art NDIR analyzers in stability and
cross-sensitivity to other gaseous species.
The prototype analyzer, a completely selfcontained unit
powered from 115 V 60 HZ line, was delivered to NAPCA in January
1971. It showed an ultimate detectivity to CO of approximately
1 ppm, and no measurable cross-sensitivity to any other gaseous
species at the level normally found in the ambient atmosphere.
2. THEORY OF OPERATION
The analyzer uses the principle of infrared fluorescence,
where a vibrationally excited molecular species decays electro-
magnetically and emits a line spectrum of infrared radiation which
is an exact match of the infrared absorption spectrum of that mole-
cule. Two isotopic forms of carbon monoxide are used in the
fluorescent mode. One, the normally abundant species CO16, was
used to provide the measuring beam of the instrument since CO16
fluorescent radiation is strongly absorbed by normal carbon monox-
ide. The second beam, CO18 fluorescent radiation, was used to
provide a reference beam since the variety of radiation is only
weakly absorbed by the low abundance of this isotope of carbon
monoxide in nature.
Referring to Figure 1, the operating principles of the in-
strument are as follows:
-------
Infrared radiation from the black body source is used to
vibrationally excite the carbon monoxide molecules in a sealed
fluorescent cell. The fluorescent cell contains a mixture of two
isotopes of carbon monoxide at pressures of approximately 8.0 torr
each and is filled to 1 atmosphere total pressure with Argon as a
buffer infrared neutral gas. The gases in the cell are of extreme-
ly high purity so that the vibrationally-excited molecules de-
excite by re-emission of electromagnetic radiation; i.e., fluores-
cent radiation. Thus, the cell continuously emits both CO16 and
Diode
-w-
o
*a
•o
f
Black Boc
Source
lyc
K
Phototrai
CO18
Argon
*
k
isi
stor —
co"
—
co"
->to processing electronics
1 metre long
sample cell
f /
« /
Interference F:
r"l_!
v\
i^ to processing
electronics
fs
PbSe detector
Chopper
Fluorescent Cell
Figure 1. Arkon 0-20 ppm Carbon Monoxide Analyzer
-------
and CO18 fluorescent radiation. The emitted radiation is chopped
by negative chopper cells containing CO16 and CO18 alternately.
The negative chopper cells contain impure carbon monoxide so that
they absorb the appropriate variety of fluorescent radiation. The
complete chopper consists, as shown in Figure 1, of two blades
rotating on a common shaft. When rotated, the fluorescent cell is
alternately exposed to exciting radiation from the black body source
and then allowed to shine its fluorescent radiation through a nega-
tive chopper cell and sample tube to the detector.
Thus the infrared detector receives alternate pulses of CO16
and CO18 fluorescent radiation which have passed through the gas
sample to be measured. Referring to Figure 2, if the CO16 and CO18
B, CO
18
A, CO
16
a.
B, CO
18
A, CO
16
B, CO
18
A, CO
16
B, CO
18
A, CO
16
2. b.
Fig. 2. a' Detector signals with no CO in sample tube
b.. Detector signals with CO in sample tube
-------
signals are defined as the "A" and "B" signals, then since the
natural isotopic abundance of CO18 is only 0.2% of CO16, only the
"A" signal will be attenuated by the presence of CO gas in the
analyzer sample tube. A light emitting diode and phototransistor
on the chopper wheel (see Figure 1) provide logic signals to the
electronics for identification of the A and B pulses. The A and
B pulses are then routed into separate demodulators where they are
synchronously demodulated and smoothed. In addition, an automatic
gain control circuit maintains the "B" signals at fixed amplitude
regardless of variations of source power or detector response. The
processing electronics then generates an output proportional to the
quantity
B - A
B
and this expression should only be sensitive to differential absorp-
tion of the A and B signals which to first order can only occur due
to the presence of carbon monoxide in the sample tube.
3. RESULTS TAKEN WITH THE PROTOTYPE ANALYZER
Part of a chart recorder trace of the analyzer output is
shown below in Figure 3. In this case the analyzer was set up for
50 ppm CO full scale and the chart recorder trace shows the varia-
tion of CO level over a two mile run through the downtown section
of Berkeley, California, near our plant. The analyzer is obviously
capable of accurately measuring CO concentration in the 0-50 ppm
range.
-------
10
20
30
40
mins
Figure 3. Carbon monoxide distribution in down-
town Berkeley, California, streets
-------
Under constant temperature laboratory conditions, a
series of carbon monoxide readings taken every 10 seconds over a
period of 5 minutes showed a variation from a highest recorded
value of 50.5 to a lowest value of 46.0 with nominal 49 ppm CO
span gas flowing through the instrument. A root mean square
analysis of the 30 recorded readings showed that the 'R.M.S. noise
of the analyzer was 0.7 ppm CO.
Possible cross-sensitivities of the prototype analyzer
to other gases were investigated specifically for the gases C02,
CH4 and H2O. For 100% CO2 the analyzer showed a downscale reading
of - 25 to 30 ppm CO, or a rejection ratio of - 30,000:1. For
100% CH4 and saturated H2O vapor at room temperature (approximately
1%) the instrument showed no deflection within its noise limit.
The performance of the prototype analyzer can therefore
be summarized as follows:
Minimum detectable value of CO: 0.7 ppm
Rejection ratio to other gases - CO2: 30,000:1
H20: > 10,000:1
CH4: > 106:1
Only limited temperature coefficient data was obtained
with the prototype analyzer. The measured zero drift was of the
order of 5 - 10 ppm CO per °C change in ambient temperature. It
was identified that the major contributor to this drift was the
change in temperature of the infrared detector. The temperature
change in the detector element caused a differential change in
16 18
its response to CO and CO fluorescent radiation and hence a
severe zero drift in the analyzer.
-------
4. CONCLUSIONS
A prototype carbon monoxide analyzer using an infrared
fluorescent source has been built and undergone a preliminary test-
ing program. It is clear that this type of analyzer has the capa-
bility to advance the state of the art in NDIR carbon monoxide mea-
surement devices in the areas of stability and freedom from inter-
ferences by other gases. More effort will have to be devoted to
an investigation of the effects of ambient temperature and pressure
variation on the span and zero stability of the fluorescent CO
analyzer before its true performance advantages can be assessed.
A potential factor of 10 improvement in CO noise level would be
attainable by a thorough upgrading of the optical components of
the analyzer.
------- |