EPA-650/2-74-059
May 1974
Environmental Protection Technology Series
wmmmmmvm
I
55
o
a)
O
-------
EPA-650/2-74-059
N02 MEASURING SYSTEM
by
M. Birnbaum, and A. W. Tucker
The Aerospace Corporation
Electronics Research Laboratory
Post Office Box 92957
Los Angeles, California 90009
Contract No. 68-02-1225
ROAP No. 56AAI
Program Element No. 1AA010
EPA Project Officer: W.A. McClenny
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVLEOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
May 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
NO2 MEASURING SYSTEM
FINAL REPORT
20 May 1974
Prepared for the
Chemistry and Physics Division
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina 27711
by
THE AEROSPACE CORPORATION
Electronics Research Laboratory
Post Office Box 92957
Los Angeles, California 90009
Contract Number 68-02-1225
Authors
M. Birnbaum
A. W. Tucker
-------
FINAL REPORT EPA CONTRACT NO. 68-06-1255
NO9 MEASURING SYSTEM
L*
TABLE OF CONTENTS
I. Purpose and Scope of Program 2
II. Outline of Tasks 2
IE. Characteristics of the LASER NO2 Monitor Prototype 3
1. Description 3
2. Optical Subsystem 3
3. Electronics Subsystem 6
IV. Calibration and System Performance 9
V. Ambient Air NO~ Levels at Location of the Aerospace
Corporation 11
VI. Conclusions 11
VII. References 13
VIII. Abbreviations 14
IX. Tables 15
X. Figures 18
- 1 -
-------
I. PURPOSE AND SCOPE OF PROGRAM
The purpose of this program is to develop a prototype "NO-
Measuring System" based upon a method originated at The Aerospace
Corporation. A laser is used to excite the NO? molecules, whose
concentration in the atmosphere is monitored by their resultant
fluorescence. The method is of great sensitivity, i. e. , 1 ppbv can be
readily detected. The signal is obtained from the NO- molecules and
therefore the method does not involve chemical processing or other
procedures which could compromise the accuracy of the measurements.
The technique has been found to be free of interference due to the presence
of other gases. The fluorescence methods can provide real-time, automatic
monitoring of ambient atmosphere NO_ levels. In short, the prototype
"NO9 Measuring System" can provide the EPA with a unique means of
O
monitoring ambient NO_ levels.
II. OUTLINE OF TASKS
The plan outline of work under EPA Contract No. 68-02-1255,
"NO_ Measuring System" consisted of the following: (1) procurement of
LJ
major components, (2) design and (3) assembly of the electronics which
•was tested with the Aerospace NO~ measurement system, (4) packaging
of the NO- measurement system, (5) calibration and (6) system performance
tests as specified in the EPA contract.
In the main, the prototype is similar to the systems operated at
the Aerospace Corporation. Two important features incorporated into the
system for the EPA are: a He-Cd cw laser operating at 442 nm used to
excite the NO_ fluorescence and (2) the fluorescence chamber redesigned
for improved performance.
The cw He-Cd laser is a compact instrument of approximately 10 mW
output at 442 nm. Its virtues, compared to the cw argon ion laser used in
1 2
prior work at Aerospace ' are its compactness and relatively lower cost.
The output power of the He-Cd laser (10 mW) is a factor of 100 less than
that of the cw 488 nm argon ion laser output. The utilization of the He-Cd
laser did, of course, require careful design to obtain the required perform-
ance accuracy.
- 2 -
-------
III. CHARACTERISTICS OF THE. LNMP (LASER NO,, MONITOR PROTOTYPE)
1. Description
The appearance of the LNMP may be inferred from the photographs
reproduced in Figures 1 to 4. The arrangement of the components is clearly
visible in Figures 2 and 3 which show respectively the front view (with cover
panels removed) and the back view (-with rear access doors open). For ease
in operation, adjustment and servicing, the Laser NO- Monitor is configured
of two subsystems (1) an optical assembly and (Z) the electronics and data
recording system. The optical subsystem assembly is easily identified by
the He-Cd laser (Figure 2) and the chamber and photomultiplier tube housing
(Figure 3). A cooled photomultiplier tube in conjunction with a photon counting
system is used to detect the weak fluorescence. The relevant spectroscopic
properties of NO7 in absorption and fluorescence have been described in
172
earlier reports.
2. Optical Subsystem
The heart of the LNMP lies in the optical subsystem shown in the
photograph of Figure 5 and the block diagram of Figure 6. The Liconix Model
401 He-Cd laser emitting at 442 nm with a nominal output of 10 mW is used
for excitation.
A major technical problem in the design of a sensitive LNMP is that
of background signals. It is found that most materials emit a broad-band
red fluorescence when excited by visible light. Thus, scattered laser light
striking components of the chamber results in a broad-band fluorescence .
This signal is observed when the laser beam is introduced into the evacuated
chamber (pressures less 10 mtorr) and when the chamber is filled with
pure gases (such as He, N?, etc) which (as far as is known) do not contain
fluorescing species. This signal, we refer to as background. A development
of an instrument with a low background signal is the key to the attainment of
high sensitivity. Many of the details of the optical arrangement described
below are used to maximize the NO_ fluorescence while minimizing the
LJ
background.
- 3 -
-------
In the LNMP developed for the EPA, key features which have
resulted in high sensitivity, low background and ease of operation are:
(1) development of an almost optimal fluorescence band-pass filter in
conjunction with excitation at 442 nm and (2) the liquid filter cells (Nos.
1 and 2 of Figure 6) which reduce background signals.
The layout of the laser and chamber, shown in Figure 6, is dictated
in part by requirements of compactness and low background signals. Prac-
tially all components which appear between the laser output end and the
chamber end have been selected with a view to minimize the background.
These are: (1) A Corning Glass Filter No. CS-5-61 to absorb the red end-
light emitted by the laser while transmitting the 442 nm laser light; (2)
442nm mirrors which have high reflectivity at 442nm and low reflectivity
at longer wavelengths, particularly in the orange-red spectral band; (3)
aperture and lens to reduce scattered laser light and improve laser beam
collimation. This reduces the background signal and permits utilization
of a small aperture (1. 8 mm dia. ) just behind the entrance Brewster angle
window. It should be noted that the LNMP operates without a noticeable
increase in background at low ambient light levels. This desirable feature
has been achieved with the combination of lens and baffles (particularly
the two baffles just behind the entrance Brewster angle-window). The
features described above which function primarily to reduce background
cannot be considered an invariable feature of the LNMP. We have opera-
ted satisfactorily a breadboard model in our laboratory without the lens
and baffles of Figure 6.
The design of the chamber (Figures 5 and 6) for the LNMP is a
practical implementation of a working design and should not necessarily
be considered optimal. Most of the components within the chamber have
been configured to minimize the background signal. The entrance and
exit windows are fabricated of fused silica (low visible fluorescence)
and are carefully polished to minimize scattering of laser light. In the
- 4 -
-------
arrangement of Figure 6, a unique feature of the LNMP is the specially
designed liquid filters.
Filter section 2 (the center section) is filled -with an aqueous solution
of Na2 Cr2Oy . 2H2O (304 g/1). This solution strongly absorbs the scat-
tered 442 nm laser light without any detectable fluorescence and, in ad-
dition, strongly absorbs the scattered Raman light due to O_, N_ and
water vapor. When using 442 nm excitation, the water vapor line
(furthest towards the red) occurs at 526 nm. The transmission charac-
teristics of a Na~ Cr_ O_ filter are shown in Figure 7 for a 304 g/1 solu-
tion of 1 cm path length which is identical to the concentration and path
length of filter section 2 (Figure 6). From Figure 7, it is evident that
the 304 g/1 Na_ Cr_ O_ solution of 1 cm path begins transmitting at 550 nm
and sharply rises to 90% transmission at 600 nm. The absorption coeffi-
cient of this filter at 442 nm is, k = 1034 cm . The absorption coefficient
is defined by: I = I e (Beer's Law). A Corning glass filter No. CS
3-66 (low wavelength cutoff at 555 nm) was placed in front of the PMT
to provide additional attenuation at the Raman line of water vapor (526 nm).
In the two filter sections (1, Figure 6), a solution of 600 g/1 CoSO.
and a 600 g/1 NiSO. solution was used. This was found to result in fewer
background counts than that obtained with a 312 g/1 solution of Na? Cr~ O_
The cells containing the solutions were fabricated of fused silica.
Fused silica is preferable to pyrex glass because its broad-band red
fluorescence is much weaker than that of the pyrex glass. The inner
diameter of the filter cells was approximately 2 cm.
Six baffles (shown in Figure 6) were used to reduce scattered laser
light and, thereby, resulted in lowering the background signal. The two
baffles just behind the front Brewster window provided the greatest re-
duction in the background.
- 5 -
-------
The final element in the chamber is the beam absorber, designed
to absorb the laser beam with a minimal increase in background. This
is accomplished with a liquid filter cell with a fused silica Brewster angle
entrance window filled with the 304 g/1 sodium dichromate solution. It
is obvious that improved sensitivity could be obtained by utilizing a mirror
to reflect the laser light back along the same path and increasing the path
length over which the fluorescence is observed.
A simple means of increasing the sensitivity was employed and
consisted of a cylindrical front surface aluminized mirror on a fused
silica substrate placed around the rear of the center section filter to
reflect fluorescence light into the PMT. This resulted in a signal increase
by a factor of about 1. 5 and a small increase in the background
3. Electronics Subsystem
The fluorescence is monitored by a cooled PMT (EMI 9659 QAM)
followed by a wide band amplifier and a high speed electronic counter
to count the electrical pulses produced by light impinging on the photocathode.
An RF shielded PMT refrigerator chamber minimizes interference from
external sources of RF radiation. The PMT is operated in a grounded
cathode configuration with the cathode attached to the RF shield. The PMT
is cooled to -25 C to reduce the number of dark current pulses. Under
separate cover, a full set of the manufacturer's instruction booklets and
detailed circuit diagrams of each component has been delivered to EPA
personnel together with the LNMP at Research Triangle Park, North
Carolina. A detailed circuit diagram of an integrator which is incorporated
into the LNMP designed by the Aerospace Corporation is part of the detailed
description of the unit. In addition, a very detailed set of instructions for
operation and adjustment of the instrument has been included with the opera-
ting manuals. Thus in this report only a brief description of the LNMP
electronics will be included.
- 6 -
-------
The output pulses of the cooled PMT are fed to a wide band, high
gain amplifier/discriminator (SSR Model # 1120). The output of this
amplifier consists of pulses of standard 25 nsec width and -1. 5V ampli-
tude and are fed to a 100 MHz counter (Newport Labs, Model # 700).
A clock and printer, together with the required storage circuit are
provided to secure several types of operation. In the repetitive timed
mode of operation, the counter records the number of counts in a pre-
set interval (10 sec, 100 sec, and etc. ), determined by the internal
time base of the counter. At the end of this period, this information is
transferred to a storage circuit. The counter is automatically reset
and begins recording counts for the next consecutive period. The coun-
ter dead time during the resetting is about 0. 1 sec. The number of
counts recorded is transferred to the dials of the print unit together
with the date and time and is printed out on a 2-1/4 inch wide strip of
paper.
The alternative mode of operation is provided by means of an
electronic integrator. In the integrator mode, after a preset laser
energy is measured (for example, 10 mW for 100 sec results in 1 J total
output), the number of counts accumulated by the counter is transferred
to the storage circuit then to the printer wheels of the printer unit and
is printed out on paper tape together with the date and time. The advan-
tage of the integrator mode of operation is that the signal counts recorded
during the interval depend only on the total preset laser energy delivered
to the sample and not on the instantaneous power levels during the interval.
Thus, the results are independent of the power variations during the inter-
vals-
For continuous monitoring, it is desirable that the data be pro-
cessed by a computer. With simple modifications, the LNMP could be
provided with an on-line computer which would process the data and pro-
vide a read-out of the concentration of NO_ in the ambient atmosphere.
c*
- 7 -
-------
Photon counting is among the most sensitive methods of detection
of low light level signals. The heart of the system is the PMT which
provides approximately noiseless amplification of the photon signal.
A photon that produces a photo-electron, is amplified and produces a
pulse of current which signals this event. The current pulse exhibits,
considerable variability in terms of amplitude, width, etc. , but this
does not prevent an unambigous recording of the release of a photo-
electron.
The difficulty with the PMT resides in the existence of other pro-
cesses which result in the release of an electron producing a current
pulse indistinguishable from that generated by photons. Among these
effects are dark current or thermionic emission from the photocathode,
UV radiation produced by cosmic-ray induced Cerenkov pulses, other
345
effects of cosmic rays and unidentified sources of radiation. ' ' A
selected EMI 9659 cooled to -25 C operates with about 45 c/sec (with
the shutter closed).
After adjustment of the PMT high voltage to secure operation on
the ''plateau", it -was hoped that the fluctuations in the counts could be
described by Poisson statistics. Tests with a selected low dark current
tube are summarized in Table I. A minimum detectable concentration
is defined to be that for •which the variation in the number of signal
counts just equals the number of signal counts, namely, S/a = 1, where
S is the signal due to the NO- concentration, and O is one standard devia-
tion of the signal counts. According to the results shown in Table I,
an operating detectability of 0.9 ppbv was obtained for a 100 sec inte-
gration time. Had the fluctuation exhibited a Poisson distribution, a
detectability of 0.41 ppbv would have resulted. Other investigators have
also noted departures from Poisson statistics and have tried to identify
345
the causes. ' ' Note that by increasing the integration time the detec-
tability increases (as the square root of the integration time) and, in
principle, very high sensitivites can be obtained by utilizing long integra-
tion periods. The presence of background counts at low NO~ levels
- 8 -
-------
reduces the sensitivity. For example, if the background counts had been
eliminated, the detectability of the LNMP would be 0. 09 ppbv a ten fold
improvement over the value listed (0.9) for a 100 sec integration time.
IV. CALIBRATION AND SYSTEM PERFORMANCE
In this section, -we will summarize many tests carried out to
delineate the performance of the L/NMP. All tests specified in the contract
have been satisfactorily completed. The detectability of the LNMP is
better by more than a factor of ten than the contract specification of
10 _+ 5 ppbv (parts per billion per unit volume).
In prior •work on a similar system, no interferences with NO_
L*
measurements had been detected as a result of the presence of other gases.
In the present contract, tests were made for possible interference resulting
from water vapor up to the saturation value. A vapor saturated air sample
was prepared and directed through the chamber and did not result in a sig-
nificant increase in the background counts. Therefore, we concluded that
the presence of water vapor up to the saturation point did not show a de-
tectable signal. The fact that water-vapor does not interfere with NO_
12
measurements was already known from earlier tests. '
The tests involving calibration and linearity of the LNMP specified
in the contract have been performed and have shown the LNMP calibrations
to be highly reproducible. Within experimental error the response of the
LNMP was shown to be linear over the range of 1 pphmv to 1 ppmv.
One of the tasks specified in the contract was the determination
of the calibration constant of the unit for 10 consecutive days. The pur-
pose of these tests was to demonstrate the repeatability and reliability
of the LNMP. A summary listing of the measurements performed on
11 consecutive working days is shown in Table II. The uncertainty in
the calibration constant is due to the statistical fluctuation of the back-
ground and signal counts and mostly due to the uncertainty of the known
NO2 sample (_+ 2. 5%). However, 73% of the values fall within 1 standard
deviation of the mean. We consider the minimum detectability (D) of the
- 9 -
-------
system to be \/2 times the standard deviation (a) of the background counts
divided by the calibration constant(C), (D = 1.4
-------
50 pphmv, the background levels begin to show a marked increase. This
is apparently due to adsorption of NO- on the -walls of the chamber and the
connecting tubes. The background can be reduced to a low level by opening
the chamber and cleaning. In normal use where concentrations seldom
go above 10 pphmv, the system was observed to operate for long intervals
without any increase in the background counting rate.
V. AMBIENT AIR NO LEVELS AT LOCATION OF THE
AEROSPACE CORPORATION
After completion of the calibration and linearity tests, the LNMP
was used to monitor the ambient air in the vicinity of our building. The
general geographical location is just southeast of Los Angeles International
Airport near the intersection of Aviation and El Segundo Boulevn Ms. The
air sample was drawn into the lab through PVC tubing with an orifice 15 ft.
above the roof of our building; the overall length of the piping to our cham-
ber was approximately 40 ft. A flow rate of 4. 5 liters/min. was used and
the air sample was filtered to remove aerosols.
The data reduction -was accomplished with our CDC 7600 computer
and the computer generated curves showing the ambient levels of NO9 are
shown in Figures 11-13. A noteworthy feature is the low levels of NO-
during a rainy period (Figure 11) and the higher levels with the return of
the usual Los Angeles climatic pattern (Figures 12 and 13). The 28th and
29th of March were cloudy days. On the 30th of March, it rained and on
this day (Figure 11) the lowest concentrations of NO., were observed.
CONCLUSIONS
The construction and operation of the Laser NO_ Monitor Prototype,
LJ
developed for>the EPA under contract No. 68-02-1255, are described in
this report. Several new features not present in earlier embodiments of
the laser-induced NO_ fluorescence monitor, were incorporated. A He-
L*
Cd laser was used for excitation. This provided an economical and com-
pact excitation source. A new type of low fluorescence solution filter .
- 11 -
-------
•was developed which resulted in greatly improved sensitivity. The LNMP
was found to operate in a highly reproducible manner (Table II) and
possessed a detectability almost a factor of 10 greater than specified
by the contract. Saturated water vapor levels did not result in any
detectable interference. The response of the instrument with respect
to concentration of NO? was found to be strictly linear over the range
tested, 1.0 to 100 pphmv. Measurements of the ambient NO_ levels
in the atmosphere in the vicinity of our laboratory building showed a
highly satisfactory performance of the instrument.
- 12 -
-------
REFERENCES
1. J. A. Gelbwachs, M. Birnbaum, A. W. Tucker and C. L. Fincher,
"Fluorescence Determination of Atmospheric NO_, " Optoelectronics,
Vol. 4, pp. 155-160 (1972).
2. A. W. Tucker, A. Peterson and M. Birnbaum, "Fluorescence
Determination of Atmospheric NO and NO_, " Appl. Optics,
Vol. 12, pp. 2036-2038 (September 1973).
3. M. Gasden, "Some Statistical Properties of Pulses from Photo-
multipliers, " Appl. Optics, Vol. 4, pp. 1446-1452 (November
1965).
4. J. P. Rodman and H. J. Smith, "Tests of Photomultipliers for
Astronomical Pulse-Counting Applications," Appl. Optics,
Vol. 2, pp. 181-185 (February 1963).
5. R. G. Tull, "A Comparison of Photon Counting and Current
Measuring Techniques in Spectrophotometry of Faint Sources,"
Appl. Optics, Vol. 7, pp. 2023-2029 (October 1968).
6. D. C. Baird, "Experimentation: An Introduction to Measurement
Theory and Experiment Design," (Prentice-Hall, Inc., Englewood
Cliffs, New Jersey, 1962).
- 13 -
-------
ABBREVIATIONS
Laser NO7 Monitor Prototype
parts per billion per unit volume
parts per hundred million per unit volume
parts per million per unit volume
photomultiplier tube
LNMP
ppbv
pphmv
ppmv
PMT
- 14 -
-------
LIST OF TABLES
I. Performance of Prototype.
II. Instrument Sensitivity -
Consecutive Daily Calibrations
(11 days).
- 15 -
-------
TABLE I
PERFORMANCE OF PROTOTYPE
lOmW at 441. 6nm
S/<7 = 1
Poisson Statistics
Measured Values
Min. Det. Concentration (ppbv) , D
Integration Time
10 sec 100 sec
1. 3
2.9
0. 4
0.9
S is the number of counts due to the NO_ concentration.
a is the standard deviation in S.
D = ale
m
C = calibration constant
- 16 -
-------
TABLE II
INSTRUMENT SENSITIVITY
CONSECUTIVE DAILY CALIBRATIONS (11 DAYS)
SIGNAL COUNTS
MINIMUM DETECTABLE
CONCENTRATION 80 SEC
INTEGRATION TIME
Day in March 1974
13
14
15
16
18
19
20
20
21
22
23
for one pphmv
8706 + 211
8430 + 204
8732 + 215
9274 + 229
8720 + 213
8379 + 216
8289 + 205
8870 + 230
8780 + 218
8835 ± 220
8846 + 216
(pphmv)
0.059
0. 065
0.033
0. 044
0. 066
0. 090
0.074
0.057
0. 061
0.059
0.056
INSTRUMENT CAPABILITY
1 pphmv _+ 0. 06 pphmv
CONTRACT SPECIFICATION
1 pphmv _f 0. 50 pphmv
- 17 -
-------
LIST OF FIGURES
1. Laser NO? Monitor Prototype.
2. Laser NO_ Monitor Prototype - front view with cover panels removed.
LJ
3. Laser NO- Monitor Prototype - rear view with rear access doors
open.
4. LNMP electronics and data recording subsystem.
5. LNMP optical subsystem.
6. Block diagram of optical subsystem.
7. Spectral transmission of the Na? Cr_ O_ . 2H?O aqueous solution
(304 g/1, 1 cm path length).
8. Linearity test for NO,, monitor in the 1 to 10 pphmv range for
filtered and unfilterea samples.
9. Linearity test for NO- monitor in the 10 to 100 pphmv range for
filtered and unfilterea samples.
10. Optical power linearity test for NO2 monitor.
11. Atmospheric NO? concentrations, El Segundo, California
March 28-30, 1974.
12. Atmospheric NO- concentrations, El Segundo, California
April 1-3, 1974.
13. Atmospheric NO? concentrations, El Segundo, California
April 4 and 5, 1974.
- 18 -
-------
I
Figure 1, Page 19
-------
Figure 2, Page 20
-------
Figure 3, Page 21
-------
Figure 4, Page 22
-------
Figure 5, Page 23
-------
CD
ob
cu
Figure 6, Page 24
-------
Figure 7, Page 25
-------
CO
CM
CD
UJ
O
V— O
CU CM
(M
i i
CD
2
a
a
O
in
z
LU
"8
CM
O
(M
00
in
CO (M
M3SV1
I M3d .01
F
Figure 8, Page 26
-------
o
(VI
o
o
>
2
o
CO
z °
o
UJ
O
S
-------
CNJ
CO
CD
8
CU
CD
CU
CM
0 8
CM
O H
Z <
CM
•1
cc
CO UJ
1
1^-
o:
u
co
CM
Figure 10 Page 28
-------
I 0
CL
z 3
g
h-
< 2
*
'
o
o
(M 0
o
2 3
Atmospheric NO, Concentrations
MARCH 28, 1974
MARCH 29, 1974
MARCH 30, 1974
8
10 11 12 13 14 15 16 17
PACIFIC DAYLIGHT SAVINGS TIME (hr)
18 19
Figure 11, Page 29
-------
Atmospheric N02 Concentrations
3
2
1
Q.
Q.
z
u
O
O
6
5
3
2
1
APRIL 1, 1974
APRIL 2, 1974
APRIL 3, 1974
J I
I
I I
I I
I I
J I
10 11 12 13 14 15 16 17 18
PACIFIC DAYLIGHT SAVINGS TIME (hr)
19 20
Figure 12, Page 30
-------
Atmospheric NO, Concentrations
12.0
10.5
| 9.0
a.
a
Z 7.5
g
a 6.0
0 4.5
1.5
0
I
15.0
13.5
12.0
2 10.5
a
a.
~ 9-0
7.5
6.0
8
-------
------- |