"\
PB 210 671
Field Study
on Application of Laser
Coincidence Absorption
Measurement Techniques
General Electric Co.
February 1972
Distributed By:
National Technical Information Service
U. S. DEPARTMENT OF COMMERCE
5285 Port Royal Road, Springfield Va. 22151
-------
BIBLIOGRAPHIC DATA 1. Report No. 2-
SHEET APTD-0981
3. Accession Ni* ^
Kb-210 671 1
5- Kcpoir Date
February 1972
4. Title and Subtitle
Field Study on Application of Laser Coincidence Absorption
Measurement Techniques
5.
7. Author(s)
8. Pe forming Organization Rept.
No.
9. Performing Organization Name and Address
General Electric
Electronics Laboratory
Syracuse, New York
10. Pioject/Task/Work Unit No.
11. ContractNo.
EHSD 71-8
1 2* Sponsoring Organization Name and Address
ENVIRONMENTAL PROTECTION AGENCY
Durham, North Carolina 27701
13« Type of Report & Period
Covered
Final
H.
15. Supplementary Notes
16. Abstracts The purpose1 was to conduct a field study on the merits and
limitations of laser coincidence absorption measurement technique applied
to long-path monitoring of a gaseous pollutant in an urban atmosphere.
Two gaseous pollutants, ethylene and ammonia, were selected and spectral
interferences identified. Laser wavelengths were selected and appropri-
ate weighting functions calculated and read into the signal processor.
Using a spectrally tunable CO2 laser, measurements and system evaluation
were conducted at Cazenovia, New York. This rural test site was used to
calibrate the system and conduct measurements under variable conditions
of weather, time of day, temperature, etc. A site was selected in Syra-
cuse, New York which provided a good average sampling of the pollutant
gases. The laser system was moved to this location. Measurements made
at the rural test site were repeated under similar conditions of time
and weather. Selected pollutant concentrations and spectral interference
effects were recorded. Concurrent point measurements were made by gas
chromatograph for ethylene concentrations. Overall system effectiveness
17. Key Words and Document Analysis. 17o. Descriptors
Air Pollution
Lasers
Field tests
Ethylene
Amnion ia
Rural areas
Urban areas
and test results wero analyzed and per-
formance evaluated.
17b. ldentifiers/Open-Ended Terms
Reproduced by
NATIONAL TECHNICAL
INFORMATION SERVICE
US Department of Comme/ce
Springfield, VA 22T51
17c. COSATI Field/Group
18. Availability Statement
FORM NTIS-38 ( 10-70)
13 B
Unlimited
19. Security Class (Thi
Report)
WCLASSIFH&P
20. Security Class (This
Page
UNCLASSIFIED
21. No. of Pages
22. Price
-------
Final Report:
FIELD STUDY ON APPLICATION OF
LASER COINCIDENCE ABSORPTION
MEASUREMENT TECHNIQUES
February 1972
Contract EHSD 71-8
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Durham, North Carolina 27701
-------
TABLE OF CONTENTS
Section Title Page
I SCOPE 1/2
E INTRODUCTION AND SUMMARY 1/2
IH DESCRIPTION OF THE LASER SYSTEM 3
A. INTRODUCTION 3
IV PRELIMINARY STUDY AND SYSTEM OPTIMIZATION ... 9
A. OBJECTIVES 9
B. SUMMARY OF RESULTS IN PRELIMINARY STUDY . . 9
C. AMMONIA SURVEY 9
D. ETHYLENE SURVEY 13
E. SELECTION OF SPECTRAL LINES 14
F. SELECTION OF LINEAR WEIGHTS 27
V TEST PLAN 29
A. OBJECTIVE 29
B. EQUIPMENT 29
1. Cazenovia Test Site 29
2. Urban Site 30
C. MEASUREMENT PROCEDURES 30
1. Cazenovia 30
2. Urban Site 33
D. DATA RECORDING 33
1. Data Format 33
2. Ambient Conditions 35
E. DATA ANALYSIS 36
VI RURAL TEST SITE MEASUREMENTS PROGRAM 37
A. SUMMARY OF RURAL TEST RESULTS 37
B. RURAL TEST SITE 38
C. RURAL TEST RESULTS 38
li
-------
TABLE OF CONTENTS
(concluded)
Section Title Page
VII TRANSFER TO URBAN SITE 55
VIII URBAN TEST SITE MEASUREMENTS PROGRAM .... 59
A. SUMMARY OF URBAN TEST RESULTS 59
B. URBAN TEST RESULTS 59
C. SYSTEM IMPROVEMENT WITH A CLEANUP
APERTURE 65
DC SUMMARY OF SYSTEM PERFORMANCE EVALUATION . . 67
1. Severe Scintillation Induced by Local
Heat Sources 67
2. Long Term Unbalance or Drift Error Produced
by Changing the Attitude or Pointing Direction
of the Laser Transceiver and then Correcting
for this by Translating the Germanium Focusing
Lens at the Entrance to the Beam Expander ... 67
3. Unbalance Caused by Aiming Error from
Transmitter to Retroreflector 67
4. Direct Coupled Amplifier Drift and Paper
Slippage in the Chart Recorders 68
5. Errors Caused by System Noise 68
6. Unusually Large Scintillation Caused by
Defocusing the Retrotelescope 68
7. Detection of 14 ppb Ethylene, 19 ppb Ammonia . . 69
8. Projected Performance Capability 69
X REFERENCES 70
lii
-------
LIST OF ILLUSTRATIONS
Figure No. Title Page
1. ILAMS Block Diagram 4
2. nVn Laser Optical Layout 5
3. Four-Wavelength Signal Processing Block Diagram 6
4. Breadboard System at Rural Site 7/8
5a. Power Output versus Wavenumber for "V" Laser Using a
Mode-limiting Circular Mode Stop 16
5b. Power Output versus Wavenumber for "V" Laser Using a
Mode-limiting Circular Mode Stop 17
6. Transmission Spectrum of Ethylene 18
7. Transmission Spectrum of Ammonia 19
8. Transmission Spectrum of Ozone 20
9. Transmission Spectrum of an Automobile Exhaust Sample
Automotive Exhaust - Cold Idle - 10 Meter Cell 21
10. Atmospheric Transmission over a 0.3 km Path in the
Chesapeake Bay area 22
11. Carbon Dioxide Absorption and Water Vapor Absorption in
Solar Spectra 23
12. Measured 10.59/i Transmittance of Water Vapor in Air
at 23 C versus the Partial Pressure of Water Vapor for a
980-m Path 24
13. The Logarithms of the 9. 55/j. and 10. 59/i Transmittance of
Water Vapor at 25°C versus Pressure for a 980-m Path 25
14. Rural Test Facility Profile 39
15. Photo of Cazenovia Valley 39
16. ILAMS Transceiver and Recording Equipment at Rural Test Site 40
17. 27-foot Gas Cell at Rural Test Site 40
18. Contours of Equal Power Density at the Beam Expander Output . 42
19. Chart Record 51/52
20. Relationship Between Scintillation and Weather at Rural Site ... 53
21. Relationship Between Scintillation and Noise and Drift 54
22. Urban Test Site Photos 57
iv
-------
LIST OF ILLUSTRATIONS (concluded)
Figure No. Title Page
23. Relationship Between Scintillation and Weather at Urban
Site (Skytop) 63
24. Relationship Between Scintillation and Noise and Drift
(Urban Site) 64
26. Location of Cleanup Aperture 66
LIST OF TABLES
Table No. Title Page
I. SUMMARY OF ETHYLENE SURVEY IN SYRACUSE, N. Y 10
II. RESULTS OF AMMONIA MONITORING BY THE ONONDAGA
COUNTY HEALTH DEPARTMENT IN 1969 12
IH. AMMONIA SOURCES IN SYRACUSE, NEW YORK —
COUNTY EMISSION SURVEY FOR 1969 13
IV. POWER OUTPUT VERSUS WAVELENGTH AND WAVE
NUMBER FOR THE CO2 LASER SYSTEM 17
V. 10. 59/jl WATER VAPOR AND CO2 EXTINCTION COEFFICIENTS
AND LOSS PER KILOMETER AT 25°C (77°F) 24
VI. RELATIVE SYSTEM RESPONSE ASSUMING TWO WAVE-
LENGTHS 25
VII. MEASURED ABSORPTION COEFFICIENTS 27
Vin. SETS OF OPTIMUM LINEAR WEIGHTS 28
IX. STANDARD DATA RECORDING SHEET 34
X. LAST 24 HOURS OF DATA TAKEN AT RURAL TEST SITE IN
CAZENOVIA, N.Y 49/50
XI. URBAN TEST SITE DATA 62
v/vi
-------
I. SCOPE
The purpose of this program is to conduct a field study on the merits and
limitations of laser coincidence absorption measurement technique applied to
long-path monitoring of a gaseous pollutant in an urban atmosphere.
II. INTRODUCTION AND SUMMARY
The program was conducted in five phases:
Preliminary Study and System Optimization - During this phase, two
gaseous pollutants, ethylene and ammonia, were selected and
spectral interferences identified. Laser wavelengths were selected
and appropriate weighting functions calculated and read into the
signal processor. The phase objective was to optimize system
sensitivity to the selected gases while rejecting spectral inter-
ference effects.
Rural Test Site Measurements - Using QE's spectrally tunable CO2
laser, measurements and system evaluation were conducted at
Cazenovia, New York. This rural test site was used to calibrate
the system and conduct measurements under variable conditions of
weather> time of day, temperature, etc.
Transfer to Urban Test Site - With EPA approval, a site was selected
in Syracuse, New York which provided a good average sampling of
the pollutant gases. The laser system was moved to this location.
Urban Test Site Measurements - Measurements made 'at the rural test
site were repeated under similar conditions of tim^ and weather.
Selected pollutant concentrations and spectral interference effects
were recorded. Concurrent point measurements were made by gas
chromatograph for ethylene concentrations.
System Performance Evaluation - Overall system effectiveness and test
results were analyzed and performance evaluated. The conclusions
are summarized in this report.
1/2
-------
III. DESCRIPTION OF THE LASER SYSTEM
A. INTRODUCTION
In order to understand the meaning of this measurements program, it is
necessary to understand the system. The laser system used in this program
is a breadboard system constructed by the General Electric Company under an
internally funded program. The system operates in the middle region of the
infrared spectrum and identifies atmospheric constituents by absorption spec-
troscopy. It measures average pollutant concentrations over long ranges at
relatively low, safe power levels. ILAMS (Infrared Laser Atmospueric Moni-
toring System) consists of a spectrally scanning laser, transmit-receive optics,
a signal processor and a retroreflector. The system transmits energy at four
wavelengths m a rapid sequence to a remote retroreflector located (depending
upon the operational situation) from one to ten or more miles away. This laser
function is called spectral scanning.
Laser power densities are kept below safety limits by expanding the beam
0. 5 watt (5 x 10"3 watts/cm^) at the laser and 0. 015 watts (2 x 10^ watts/cm^)
at the retroreflector. Energy returned from the retroreflector is collected in
a collinear transmit-receive optical system and referenced to lase • output
energy to compensate for power differences at each wavelength. p'nergy atten-
uation at the selected lasing wavelengths produces a pattern of absorption
versus wavelength with which the system identifies pollutants and registers
their average concentration over the path (double the laser to retroreflector
distance) traversed by the beam.
An ILAMS block diagram is shown in Figure I. The output power from the
laser is directed to a 50-percent beamsplitter. The energy transmitted through
the beamsplitter is focused down to a 0.004-inch aperture that serves as an
attenuator for the detector and as a simulator for the retroreflector. Behind
this aperture is the reference energy detector. The energy at this focal point
of the output beam has the same energy distribution in cross-section as does
the far-field pattern of the laser, but on a smaller scale. Any ch3nges in the
energy distribution in the far-field due to laser noise will have the same effect
on the reference detector as it does on the energy collected by the retroreflector.
Because of diffraction, the energy that passes through the aperture spreads out
in a 5-degree cone. A thermistor bolometer detector behind (he aperture picks
up the central part of this cone and the attenuation is varied by controlling the
aperture to bolometer spacing.
The reflected power from the beamsplitter goes through a germanium lens
which focuses the energy near tb? focal point of an off-axis parabolic mirror.
The energy then reflects off a 45: degree flat mirror to the parabola, and the
expanded, nearly collimated beali is transmitted to the retroreflector. The
3
-------
c:
V.
tK-
APERTURE
SPECTRALLY SCANNING
LASER
•M?
REFERENCE
DETECTOR
^DOUBLET LENS
£V SIGNAL DETECTOR
ELECTRONICS
LINEAR WEIGHTS
SIGNAL PROCESSING
concentration
READOUT
Figure 1. ILAMS Block Diagram
return energy from the retroreflector retraces the path through the beam-
expanding parabolic mirror and the germanium lens to the beamsp.'itter. The
energy reflected from the beamsplitter is lost, but that which is transmitted
is collected by a germanium lens doublet and focused on the signal detector.
Preamplifiers are mounted directly behind the signal and reference detectors;
the preamp outputs go to the signal processor. The detectors used in the sys-
tem are thermistor bolometers operating at ambient temperature (uncooled)
and giving a characteristic flat response across the middle infrared spectral
region.
The attenuation of laser energy at many wavelengths produces absorption
patterns enabling the system's signal processor to effectively separate pollutant
effects from spectral interferences. To aid in the selection of these wavelengths,
off-line computer programs have been developed using inputs such ?s the avail-
able number of lasing lines, their wavelengths and power, the number of pollu-
tants, interferences and spectral data. The computer sorts through the data
produced, to rank wavelength combinations according to their effectiveness in
pollutant discrimination. Linear weights for use in the system's signal proces-
sor are calculated with efficient computer programs developed for this purpose.
The weights are manually set into the system's signal processor. Returned
signals at each of the lasing wavelengths are weighted and summed by the
processor to cancel interference effects and produce pollutant concentration
readouts.
The COo laser in the breadboard system is alignable to four wavelengths. ^
These wavefengths may be selected from the more than 70 available with C Og .
The laser is presently designed to detect ethylene and ammonia in a spectral
environment that is expected to contain neutral attenuation, ozone, water vapor,
and gasoline engine exhaust fumes and a random spectral attenuator such as
atmospheric scintillation might produce. The wavelengths selected were 9. 505,
10.532, 10. 675 and 10. 719 microns.
4
-------
The optical configuration of the laser is shown in Figure 2. The term
"V-laser" comes from the shape of the plasma tube. A relatively long laser
cavity is used for sufficient gain to overcome the losses inherent in the spectral
tuner and to obtain lasing action on a large number of spectral lines. A beam
travels through the plasma tube with aid of a mirror at the point of the V.
Leaving the tube through a germanium Brewster window, the beam is directed
by mirrors through an iris (for mode control) and onto a 105 Lnes/mm diffrac-
tion grating having a dispersion of 105 mr/micron which disperses the beam
spectrally and spatially. The four wavelengths of interest are than relayed
(intercepted) first by two mirrors, and after passing through holes in the chop-
per wheel, by the four end mirrors of the laser cavity. These holes are so
located that, as the wheel turns, only one wavelength at a time is permitted to
pass through to the end mirrors. The four end or wavelength selection mirrors
are adjusted so that the beams are directed back on themselves and through the
laser cavity. In this way, selected laser wavelengths are transmitted sequen-
tially at a rapid rate. A typical chopping wheel speed, 4800 r(Ai, sends all
four wavelengths out in l/80th of a second.
A block diagram for the four-channel (four wavelength) signal processor is
shown in Figure 3. The design is general and can be expanded to accommodate
more wavelengths by duplication of the circuits following the log amplifier and
difference amplifier.
A low-noise preamplifier having a low frequency voltage gain of 2,000 in-
creases the detector signal to a usable level. High-frequency compensation in
the preamplifier corrects for the 1.3 millisecond response time of the bolometer.
The frequency response of the two together is ilat out to 3, 000 riz with a 3-dB
break at 7, 500 Hz.
The AGC amplifier, "A", is required to provide a constant average output
with variations of return signal strength. This AGC is desirable m order to
stay within the optimum dynamic range of tht> log amplifier. AGC ampli ler "B"
is required to compensate for the average eitects of AGC amplifier "A", as well
as to correct for gain changes in the reference energy channel.
5
-------
Figure 3. Four-Wavelength Signal Processing Block Diagram
The function of the d-c restorer is to set the d-c zero level of the signal
after it passes through the a-c coupled preamplifier. Sampling ano controlling
the d-c level of the AGC output provides offset compensation and a convenient
sensing point for AGC control. A peak detector and meter provido a visual
indication of the return signal level. A low-signal threshold presents a visual
indication if the signal drops below a preset level.
After the log of the ratio of the return energy to the reference energy is ob-
tained at the output of the difference amplifier, the transmission information is
extracted by synchronous detectors gated in synchronism with the laser wave-
length scanning. Each synchronous detector is balanced to zero by an added
voltage to compensate for any fixed spectral attenuation. An absorption in any
channel is detected rs a calibrated d-c voltage which is indicated on the signal
meters.
The balanced synchronous detector outputs are weighted and summed. The
resultant d-c voltage is indicated on the output meters, each of which represents
an absorber or class of absorbers. The gating control circuits translate informa-
tion from the magnetic pickups on the chopper wheel into timing signals which
control the operation of the synchronous detectors and the d-c restorers.
6
-------
Breadboard transmit-receive optics expand the laser beam u a diameter
of 5 inches which reduces transmitted power densities to values below 0.01 watt
per square centimeter while also contributing to beam collimation The "cat's
eye" style retroreflector is a 12-inch-diameter f/4 parabolic miri » with a 4b-
inch focal length and a one-inch-diameter,40-inch-radius-ol eurvai ire.concave,
spherical mirror at its focal point. The resolution of the retrorei lector is suf
ficient to return 90% of the incident laser energy to the receiver at one-mile
range. Such operational characteristics give the system an open r n>ge of
10 miles or more under conditions of "good visibility"
Figure 4 shows the present unit at the rural test site. While satisfactory
for current feasibility studies, these are breadboard components, with .subsiau
tial reduction in size possible. The V laser is mounted on an alu linuni channel,
6 feet in length. Repackaging the present analog functions will af >j d a con-
siderable reduction in size of the signal processing electronics, ^readiioaro
electronic components, excluding meters and substituting fixed r< ^lskir; and
trimpots for weighting potentiometers could be packaged ir a volu ne of about
2x4 inches. According to our studies, digital signal processing (cost
effective when the number of transmitted wavelengths exceeds six would not
have an adverse effect on processor size. A 10 x 40 inch laser breadboard
having comparable performance characteristics has also been dn k-|.ed by
under a separate program.
Figure 4. Breadboard System at Rural Site
7/8
Tins page js icpm-'"u-t'il at
liack of lid.' ri'juu I ,i i i, Jt i
reproduction rictlm to p.o>
better detail.
-------
IV. PRELIMINARY STUDY AND SYSTEM OPTIMIZATION
A. OBJECTIVES
The goals of this preliminary portion of the program are to select a pol-
lutant and determine the significant spectral interferences from pollution
monitoring data and laboratory analysis of atmospheric samples. Using this
information, the system is adjusted to optimize its sensitivity to the selected
pollutant gas while rejecting the effects of spectral interfererces.
B. SUMMARY OF RESULTS IN PRELIMINARY STUDY
Ammonia measurements made in central Syracuse, New York in 1969 gave
measured concentrations ranging from 0.043 parts per million to 0.148 parts
per million. These data are twenty-four hour averages taken every twelve days
throughout 1969, and analyzed by NAPCA in Cincinatti.
Ethylene measurements, made by point sampling with a Saran sample bag
at downtown locations in Syracuse, gave detectable concentrations of ethylene
(5ppb) about fifty percent of the time. The maximum concentration measured
in the open air was 125 ppb at ground level (see Table L).
Four spectral lines were selected for the detection of bath ethylene and
ammonia and the laser was aligned to those wavelengths. They are: 10. 719
microns (P-32), 10.675 microns (P-28), 10. 532 microns (P-14} in the 00°1-
10°0 transition and 9. 505 microns (P-14) in the 00°1-02°0 transition.
Direct measurements were made of the absorption coefficients of ethylene
and ammonia, using the laser as the source. These data were used to deter-
mine weighting functions that were set into the signal processor.
C. AMMONIA SURVEY
The data on the concentrations of ammonia in the city ci Syracuse, New
York were taken from the results of ammonia monitoring by the Onondaga
County Health Department in 1969. These data are from twenty-iour hour
bubbler samples taken from a window in the Water Department building in
central Syracuse approximately once every twelve days. The data were ana-
lyzed by NAPCA in Cincinatti and the results are presented in Table II.
Table in lists the ammonia sources in Syracuse, New York based on the
Onondaga County emission survey for 1969.
9
-------
TABLE 1
SUMMARY OF ETHYLENE SURVEY IN SYRACUSE, N. Y.
DATE
LOCATION
ETHYLENE,
CONCENTRATION
12/1/70
637 N. Salina Street
downtown at street
level*, heavy traffic
50 ppb
12/1/70
Corner Water and
Salina Streets
downtown-street level-
heavy traffic
125 ppb
12/1/70
Under Route 81 skyway
near Medical Center-
heavy traffic
54 ppb
12/23/70
Top floor (5) of
Warren Street Parking
Garage, 10:30a.m.
27 ppb
12/24/70
Cazenovia Samples.
Two samples collected
at test site in country.
0
12/29/70
Inside of Sibleys'
parking garage
460 ppb
12/29/70
On top of Sibleys'
parking garage in
cold cross-wind.
0
1/05/71
Inside of Sibleys'
parking garage, no cars
5 ppb
(minimum detectable level)
1/05/71
On top of Sibleys1
parking garage
0
1/05/71
Erie Blvd. East
in Holiday Bowl
Parking lot
5 ppb
1/05/71
Corner Teal Avenue
and Erie Blvd.
13 ppb
1/11/71
Two samples in East
Syracuse Freight Yards
0
(Continued on
next page.)
10
-------
TABLE I (concluded)
SUMMARY OF ETHYLENE SURVEY IN SYRACUSE. N. Y.
DATE
LOCATION
ETHYLENE,
CONCENTRATION
1/13/71
Syracuse Airport
Runway Sample
during use
5 ppb
1/20/71
Three more samples
collected around perimeter
of airport
0
2/16/71
Four Samples taken at
street level on Hiawatha
Blvd. near end of lake
0
9/3/71
Two Samples. One at
Skytop near laser site,
another in Oakwood
cemetary on hill top
0
9/9/71
Oakwood Cemetary
hill top and Colvin
Ave. near gym
n
10/5/71
Top of Sibleys' parking
garage
0
I
10/5/71
Street level at Jefferson
i
i
|
12/22/71
Men's dormitory at
Syracuse University
1
0 i
11
-------
TABLE H
RESULTS OF AMMONIA MONITORING B\r THE ONONDAGA COUNTY
HEALTH DEPARTMENT IN 1969
DATE AMMONIA
PPM
01/7 0.052
01/19 0.057
01/31 0.063
02/16 0.066
03/08 0.069
03/20 0.073
04/1 0.075
04/13 0.079
04/25 0.058 These data are from twenty-four hours bubbler
05/7 0. 043 samples taken from a window m the Water
05/19 0.043 Department building once every twelve days.
05/31 0.076
06/12 0.127
06/24 0.068
07/6 0.086
07/18 0.047
07/30 0.073
08/11 0.043
08/23 0.055
09/4 0.148
09/16 0.063
09/28 0.073
10/10 0.046
10/22 0.062
11/3 0.052
The above data were received from Bill Compton of Syracuse
University Research Corporation on September 25, 1970.
12
-------
TABLE III
AMMONIA SOURCES IN SYRACUSE, NEW YORK -- COUNTY EMISSION
SURVEY FOR 1969
SOURCE
TONS/YEAR
PERCENTAGE OF TOTAL
Industrial Process
Industrial Fuel
Private Fuel
Government Fuel
Commercial Fuel
Refuse Disposal
Transportation
193
16.1
719
*59.9
47
3.9
17
1.4
67
5.6
5
0.4
153
12. 7
D. ETHYLENE SURVEY
The results of the ethylene survey were presented m Table I. Ambient
ethylene was point-monitored by a grab-sample technique. Ten liter samples
were pumped into saran bags (from the Anspec Company, Ann Arbor, Michi-
gan) with the aid of a one-liter Hamilton syringe. The sample bags were
transported back to the laboratory and analyzed as soon as possible, usually
within 30 minutes and always within one hour. Analysis of standard samples
by this technique showed no loss of ethylene due to diffusion or wall effects.
Quantization was accomplished by use of a Varian-Aerograph Model 1520 gas
chromatograph equipped with a flame ionization detector and fitted with a 1/8"
x 61 stainless steel column packed with well-aged Chromosorb 112, 60/80
mesh, and operated at 57°C with a nitrogen Carrier gas flow of 50 ml/min.
Samples were transferred from the saran bags to the chromatograph by means
of 10 mi - Hamilton gas-tight syringes. Calibration curves were constructed
using the same type of equipment. The calibration gases were primary
standard ethylenes in nitrogen purchased from Matheson.
In earlier experiments, removable sampling loops m £. gas sampling
valve were packed with an absorbent to trap the ethylene when an air sample
was drawn through the loop. Various adsorbents were investigated for this
purpose. Calibrated volume vacuuir chambers were constructed so that known
air volumes might be drawn througi the sampling loops tn the field without
the aid of pumps. The planned prrpedure was that, once a proper adsorbent
was chosen, the instrument could je calibrated and measurements of ethylene
levels could begin. Extensive teehng of the packed sampling loops for am-
bient temperature collection of etjylene indicated that the efficiency of
adsorption was somewhat lower tian exoected. This technique was abandoned
in favor of the saran bags.
13
-------
The selection of both Ethylene and ArrJnonia as pollutants to be monitored
by the laser system was based upon two faltors:
1) Both pollutants were found to exisA in the Syracuse atmosphere
in detectable concentrations
2) The four-wavelength GE laser coulld be aligned to drtect both
pollutants simultaneously without yompromising the system's
sensitivity to either one.
E. SELECTION OF SPECTRAL LINES
This laser instrument measures directly the transmission (T., T2, T^,
T^) of the sample region between the transmitter and the retrorefrector at
four wavelengths. Assuming that the line width of the resonance absorption
is broad compared to the transmitted laser line width, and assuming also
that there is no saturation in the absorbing media, then, for a uniform con-
centration of the absorber over the path, the transmission at each discrete
wavelength is of the form, Tm = exp (-AmC^L); where Am is the absorption
coefficient of absorber A at wavelength m, C^ is the concentration of ab-
sorber A over the total optical path, and L is the total optical path through
the sample region. If the concentration is non-uniform over the path, as is
the more usual case, then C^L can be replaced by the integrated concentra-
tion over the path or, more simply, let C^ represent the average concentra-
tion over the path.
Typically, C has units of grams/liter or atmospheres of partial pressure,
and L is in centimeters. Am is in units to make AmC^L dimensionless.
If a second absorber B with absorption coefficients B is introduced into
the region, the net transmission will be the product of the transmission due to
each absorber.
-[A C.L + B C„L]
_ = e 1 m A m B 1
m
If the natural log of the transmission at each wavelength is taken elec-
tronically, then
InT =-(A C.L + B CnL)
m m A m B
and the resulting signals have two convenient properties:
1) System response to any absorber is a linear function of the quantity
(concentration x path) present
2) System response to several absorbers is the sum of the responses
to the individual absorbers.
14
-------
Therefore, if the system can be designed to give a zero response to spectrally
interfering absorbers, the system will respond only to the pollutant to be
measured, and the response will be proportional to the quantity present.
Using decision theory and multivariate statistical analysis, it can be
shown that the optimum signal processing involves the use of single or
multiple linear weights. Let Sj = -lnT^. Application of a single linear
weight, W, means taking a linear sum of the signals Sj, ... Sn to give
a new signal, S = + W2S2 + ... WnSn. The quantity of absorbers
present can be accurately determined by examining the magnitude of such
linear sums.
Techniques may also be applied for choosing linear weights to accurately
measure the quantity of a given pollutant in the presence of known interfering
spectral absorbers, random spectral absorbers, scintillation, and other
"noises"
On the basis of both analytical and experimental work, jeveral basic
conclusions about wavelength selection can be drawn. These are:
1) The relative success of a group of wavelengths depends directly
on the measurement problem. The pollutants to be measured,
pollutants and absorbers to be ignored, expected quantities of
absorbers and the system noise levels, all affect the choice of
wavelengths.
2) For a given problem, there will be an optimum set of wavelengths.
Increasing the number of pollutants to be estimated or ignored will
tend to increase the optimum number of wavelengths to be used;
i.e., the more complex the environment, the more wavelengths
are necessary.
3) The finer the spectral structure of an absorber, the fewer number
of wavelengths are needed to best measure the quantities of it
present. The laser system is not limited to detecting pollutants
with fine structure such as ammonia and ethylene. In fact, it does
remarkably well in detecting or rejecting absorbers with rather
smooth spectral characteristics.
4} On the basis of past experience, four wavelengths have done an
excellent job in handling spectral recognition problems m environ-
ments representative of the real world.
Selection of the best four wavelengths depends upon the relative intensity
of the available spectral lines from the kisei, the absorption spectra of the
target pollutants, the absorption spectra of the other gases and particulate
material in the air, and their expected concentrations. The measured spectral
lines from the CO2 laser system m single-mode operation ar3 given in Tables
IVa. and IVb. The accompanying Figures (5a. and 5b.) show the same informa-
tion. Figures 6, 7 and 8 show measured spectra of ethylene, ammonia and
ozone, respectively (*). These data were taken with a single-beam spectrom-
eter having a resolution of 0. 1 m"1. Figure 9 shows a low-resolution spec-
trum of a typical automobile exl/iust fume sample measured with a ten-meter
^ P. L. Hanst, private commun Cation
15
-------
TABLE IVa.
POWER OUTPUT VS. WAVELENGTH ND WAVE NUMBER) FOR THE
C02 LASER SYSTEM
p
R
Wave
Wave
Power
in
Watts
P
P
Wave
Wave
Power
in
Watts
J Values
Number
Length
J Values
Number
Length
38
926.9
10. 789
0.24
18
973.2
10.275
0. 75
36
928.9
10.765
0.34
20
974. 5
10. 262
0.74
34
930.9
10. 742
0.44
22
976.0
10.248
0. 73
32
932.9
10.719
0. 50
24
977.0
10.236
0. 71
30
934.9
10.696
0. 56
26
978.1
10.224
0. 65
28
936.8
10.675
0. 60
28
979.3
10.ill
0.63
26
938.6
10.654
0. 63
30
980.6
10.198
0. 58
24
940. 5
10.633
0. 65
32
981.9
10.184
0. 52
22
942.3
10.612
0.66
34
983.0
10.173
0.44
20
944. 1
10. 592
0.66
36
983.8
10.165
0. 40
18
945.9
10.572
0. 66
38
986. 5
10.137
0.15
16
947. 7
10.552
0.65
40
987. 5
10.126
0.00
14
944. 8
10.532
0. 64
42
988.6
10.115
0.00
12
951. 3
10.512
0.60
10
953.0
10.493
0. 57
42
922. 8
10.836
0.04
8
954.6
10.476
0. 54
40
924. 8
10.812
0.15
6
956. 1
10.459
0.40
4
957.6
10.442
0. 00
2
2
4
6
3
10
12
14
16
959.2
961. 5
963.0
964. 5
965.7
967.2
968.6
970. 2
971.9
10.425
10.400
10.385
10.369
10.355
10.339
10.323
10.307
10.290
0.00
0.00
0.02
0. 40
0. 56
0. 65
0.70
0. 73
0.75
WAVENUMBER CM'1
-i 1 1 1 1 1 1 i I r~
11.0 10.9 10 e 10.7 10.6 10.3 10 4 It 3 10.2 10.1
MICRONS
TRANSITION 00°l — 10°0
Figure 5a. Power Output vs, Wavenumber for "V" Laser Using a
Mode-limiting Circular Mode Stop
16
-------
TABLE IVb.
POWER OUTPUT VS. WAVELENGTH (AND WAVE NUMBER) FOR THE
C02 LASER SYSTEM
p
R
Wave
Wave
Power
in
Watts
P
R
Wave
Wave
Power
in
Watts
J Values
Number
Length
J
Values
Number
Le.igth
38
1029.4
9. 714
0. 50
18
1075. 7
9. 296
1.10
36
1031.5
9.694
0.60
20
1077.0
9.285
1.10
34
1033. 5
9. 676
0.70
22
1078. 4
9.273
1.08
32
1035.4
9.658
0. 76
24
1079.8
9. 262
1.04
30
1037.4
9.639
0.80
26
1081.0
9. 251
0.98
28
1039.4
9.621
0.83
28
1082.2
9. 240
0.90
26
1041. 3
9.603
0. 84
30
1083.4
9. 230
0.84
24
1043.2
9. 586
0.83
32
1084.5
9.221
0.76
22
1045.0
9. 569
0.82
34
1086.0
9.201
0.68
20
1046. 8
9. 553
0. 81
36
1087.9
9. 192
0. 52
18
1048.6
9. 537
0. 78
38
1088.9
9. 188
0. 38
16
1050.4
9. 521
0.75
40
1090.0
i). 174
0.00
14
1052.1
9. 505
0.70
42
1090.9
9. 166
0.00
12
1053.8
9.489
0.65
10
1055.6
9.478
0. 57
42
1025.2
9. 754
0.20
B
1057.4
9.457
0. 50
40
1027.4
9.733
0. 38
6
1059.1
9.442
0.34
4
1060.6
9.429
0.02
2
2
4
6
8
10
12
14
16
1062.0
1064.5
1065.8
1067.3
1068.7
1070.1
1071.5
1072.9
1074.3
9.416
9.394
9.382
9. 369
9.357
9. 345
9.333
9. 320
9. 308
0.00
0.00
0.00
0.00
0. 50
0.76
0.90
0.98
1.04
WAVENUMBER CM-1
I I | 1 -| 1 1 1 r
10 9.9 9.8 9.7 9 6 9.5 9 4 9 3 9.2
MICRONS
TRANSITION 00°1 — 02°0
Figure 5b. Power Output vs. Wavenuuoer for "V" Laser Using a
Mode-limiting Circular Mide Stop
17
-------
WAVELENGTH (MICRONS)
11.0 10.8 10.6
10.4
10 2
>
£
UJ
UJ
o
UJ
I-
t
2
CO
z
<
-------
WAVELENGTH (MICRONS)
11.0 10.8
100
Figure 7. Transmission Spectrum of Ammo.ua
19
-------
WAVELENGTH (MICRONS)
10.0 9.8 9.6 9.5 9.4 9.3
EMPTY CELL
>
o
tr
UJ
UJ
o
UJ
I-
3E
(0
z
<
cr
5 TORR. OZONE
IN ONE ATM. OXYGEN
10 CM. CELL
-------
Figure 9. Transmission Spectrum of an Automobile Exhaust Sample
Automotive Exhaust - Cold Idle - 10 Meter Cell
t
21
-------
cell on the Beckman IR-9 spectrophotometer. Note that, in Figure 9, the
wavelength is increasing from left to right, so that the picture is reversed
from that of the other three curves. The selected laser spectral lines are
shown on each of these curves. The structure of the absorption pattern for
exhaust fumes in all five exhaust samples measured indicates that ethylene
is the only significant contributor between 9 and 11 microns. We have con-
cluded that none of the other material in automobile exhaust fumes contributes
any significant spectral interference for the CO2 laser system. The other
spectrally interfering materials considered were water vapor and carbon
dioxide.
The relation between the CO2 laser lines and the H2O absorption line at
10. 542 microns is shown in Figures 10 and 11. In Figure 11, the absorption at
the line centers is lower and the line widths are larger than would be observed
with the laser due to the limiting resolution of the spectrometer; i.e., the
profiles shown are instrumental. Note that the P-14 line at 10. 532 microns
is approximately 0.01 microns (1 cm"^) away from the H2O line center. The
width of the 10. 532 micron laser gain line is approximately 0.001 cm~l and
the half width of the H2O absorption line is 0.1 cm"* (see Reference 3).
It may be noted, at this point, that CO2 in the atmosphere represents a
neutral absorption rather than a spectral absorption to the CO2 laser wave-
lengths. This is true as long as we restrict the laser to the common isotope
of carbon dioxide — The absorption lines are wide compared to the
laser gain lines, they have exactly the same centers, and they are of approxi-
mately the same strength at the wavelengths where this system will lase.
WAVENUMBER ~ cm
-1
1,000
: 1—
950
t
900
L
TO
5.7 MM PRECIPITABLE WATER
77° F TEMP
N« CL PRISM
8.5
9 O
9 3 100
WAVELENGTH (MICRONS)
10 3
II 0
Figure 10. Atmospheric Transmission over a 0. 3 km Path in the
Chesapeake Bay area
22
-------
WAVENUMBER ~ cm
WAVELENGTH IN MICROMETERS
Figure 11. Carbon Dioxide Absorption and Water Vapor Absorption in
Solar Spectra (4)
At long range and under conditions of high absolute humidity, there is an
absorption due to the wings of the water vapor lines on either side of the 9 to
11 micron atmospheric window. The absorption coefficients icr water vapor
and carbon dioxide have been measured by McCoy et al(5) at the 10. 59 micron
and 9. 51 micron lines of the CO2 laser. Their data for the 10. 59 micron line
are presented in Table V and Figure 12. Because of the extremely strong self
pressure broadening effect, the transmission of water vapor in air does not
vary exponentially with the precipitable centimeters of water in ihe path, as
Beers law would predict. However, this effect does not introduce nonlinearity
in the system response. The measured absorption coefficients for the expected
range of partial pressures at 9. 55 microns are 80% as large aa those at 10. 59
microns. That is, ag< 55/aio. 59 = 0.80. This ratio rern?ins constant as a
function of partial pressure, as can be seen from Figure 13. In other words,
the absorption pattern in the form of the natural log of the transmission at each
wavelength does not change with large concentrations.
In the selection of a group of four laser lines, then, the only spectral
interferents that were considered were ozone and water vapor. Of course,
ammonia must be considered as a spectral interferent for the detection of
ethylene and, similarly, ethylene is an interferent for ammonia.
Before the beginning of the contract effort, a two-wavelength selection
computer run was made to choose the best wavelength pair for ethylene detec-
tion from 70 CO2 laser lines, assuming 14 spectrally interfering absorbers.
They include the absorbers listed in Table VI, plus two absorbers (carbon
monoxide and nitric oxide) that do not absorb in the 9-to-ll. 5^ region and
can be ignored. The two wavelengths chosen were 10. 532(jl and 10.675ji.
The (A vacuum) values in Table VI indicate the relative system response to
equal quantities (CL) of absorbers. The most significant interference is
propylene, whose response is approximately 1/40 that of ethylene. This
rejection is not sufficient for operation in an environment where propylene
might occur in concentrations equal to or greater than that of ethylene. For
example, 10 ppm of propylene would give approximately a 25% error in
23
-------
TABLE V
10. 59m WATER VAPOR AND C02 EXTINCTION COEFFICIENTS AND LOSS
PER KILOMETER AT 25°C (77°F)
RELATIVE
HUMIDITY
(%)
k-l
(km A)
k-l
(ft I
loss.,
(dB km )
10
0.0125
3.81 x 10"6
0.054
20
0.0338
1.0 x 10"5
0. 15
30
0.0653
1.99 x 10"5
0. 28
40
0. 107
3. 26 x 10" 5
0.46
50
0.157
4.78X 10"5
0.68
60
0. 215
6. 56 x 10~5
0.93
70
0.284
8.65 x 10"5
1.2
80
0. 363
1.11 x 10"4
1.6
CO2 at 330 ppm:
0.0611
2. 48 x 10" 5
0.35
Q
Figure 12. Measured 10. 59p. Transmittance of Water Vapor m Air at 23 C
vs. the Partial Pressure of Water Vapor for a 980-m Path.
Total pressure = 700 Torr
24
-------
Several sets of weights were calculated for a variety of assumed condi-
tions. Typical results are presented in Table VIII. Spectrally neutral attenua-
tion is always considered one of the spectral interferences. In spite of the fact
that, by definition, its transmission does not vary spectrally, u must be dis-
criminated against. A one-wavGlength system, for example, could not distin-
guish between neutral attenuation and the gas that absorbs specifically at that
wavelength. Neutral attenuation is produced by particulates in the air, angular
misalignment of the optics, any obscuration in the path, dirt on the optics,
change in responsivity of the bolometer, or change in the gain of the preampli-
fier. From the absorption coefficients, it is seen that, using these four wave-
lengths, ozone is the inverse of a combination of water vapor and
neutral attenuation. For this reason we can detect ethylene, for example,
uniquely in the presence of four spectral interferences (i.e., ammonia, ozone,
water vapor and neutral attenuation) with just four wavelengths. -Table VIII
gives some calculated weights for the detection of ethylene, ammonia and also
ozone. This table is the result of just one of several computer runs to deter-
mine the optimum linear weight for different models of the spectral environ-
ment. The results were dependent upon the relative magnitude of the random
spectral interferent assumed, the model for the spectral distribution of the
interferent, and the magnitude of the other interferents. The calculated sets
of optimum linear weights, for the case where the random spectral interferents
had approximately equal magnitude to that of ethylene, ammonia and ozone,
were for all practical purposes identical to those listed in the first part of
Table VIII. In other words, only 10. 53 and 10. 67 microns are used in de-
tecting ethylene, and 10. 719 and 10. 67 microns are used for detecting ammonia.
These results are what one would expect intuitively. The wavelength corre-
sponding to the absorption peak and the nearest available other wavelength are
optimum. Our experience with optimizing the detection of chemical warfare
agents indicates that things usually do not work out so simply.
TABLE VIII. SETS OF OPTIMUM LINEAR WEIGHTS
Wavelength
9. 505
10.532
10. 675)
10.719
Wavenumber
1052.1
959. 5
936.3
932.9
Independent Detection of
Ethylene, Ammonia and Ozone
with no random spectral
interferences
Ethylene
Ammonia
+2.27 x 10"11
+1.3 x 10"10
+0.682
-0.012
-0.658
-1. 39
-0.024
41.40
Ozone
+0. 437
+0.0082
-0.445
+0.0025
Detection oi ethylene and
ammonia assuming only
a large random spectral
interference
Ethylene
-0.225
+ 1
-1. 59
+0.821
Ammonia
-0.057
+0.139
-1.08
+ 1
28
-------
V. TEST PLAN
A. OBJECTIVE
The purpose of this plan was to define the equipment and procedures to be
used in calibration of the laser system, measurements of the atmosphere and
analysis and reporting of data.
B. EQUIPMENT
Measurements were obtained at two locations; at the General Electric
Cazenovia Test Site for calibration and baseline measurement/3 in a pollutant
free environment, and at an urban site at Syracuse, New York, for measure-
ment in the presence of pollutants of interest and ambient interferences.
1. Cazenovia Test Site
a. General Electric "V" laser system including lassr,
retrotelescope and signal processor.
b. Six-inch-diameter, twenty-seven-foot-long gas cell.
c. Recorders:
1) Two dual-channel Mark II Brush Recorders,
2) One dual-channel Sanborn model 320 hot-pen recorder
3) One eight-channel Sanborn hot-pen recorder
d. Tektronix 541 Oscilloscope with Type CA Dual Trace Amplifier
e. Ammonia and ethylene test gas samples and Hamilton sample
handling syringes
f. Meteorological instruments
1) Thermometer,
2) Hygrometer, Serdex model 201
3) Anemometer, Lafayette and Taylor
g. Varian Aerograph Gas Chromatograph, model 1520-IB with
flame ionization detector
h. Hamilton air sample; syringes.
29
-------
2. Urban Site
The same equipment as at the, Cazenovia Test Site except as listed
below.
a. Recorder - A Sanborn 8ichannel recorder will replace
the three test site recorders
b. The twenty-seven-foot gas cell will not be installed at
the urban site
c. A six-inch-long gas cell will be used instead
C. MEASUREMENT PROCEDURES
1. Cazenovia
a. Laser Calibration
Calibration against specific gases was accomplished by
injecting a measured sample of the specified target gas
into the test cell and recording system response. The
sample was allowed to diffuse through the cell for ten or
fifteen minutes to observe: 1) the effects (if any) of
dilution of gas within the cell upon system response, and
2) change in the system noise and system drift (if any)
that occurs when an absorption signal is present.
The gas cell is a twenty-seven-foot section of six-inch-
ID circular copper waveguide. The cell windows are
0. 6 mil (0.0006 inch) polyethylene (handy-wrap or baggies).
The reflection losses from these windows are somewhat
spectral due to the interference between the front and
back surface reflections. This spectral interference
varies across the surface of the polyethylene because
of non-uniformlty in thickness. The spectral unbalance
due to these windows was balanced out when the cell was
set up and does not affect the system sensitivity or signal-
to-noise except for a slight loss m total power received.
An exhaust blower is mounted on the outside of the test
cell at one end and an exhaust port at the other. The
exhaust fan drives ambient air into the cell and out the
exhaust port. When not exhausting, the blower was turned
off and covered with a polyethylene bag and the thr°e-inch-
diameter port was sealed with a rubber septum.
To inject the test sample of gas, a sample cell with a
silicone rubber septum was mounted onto the lecture bottle
containing the pure test gas, i.e., ethylene or r.mmoma
under pressure, and filled with the pure gas. The gas was
extracted from the cell with a ten-milliliter hypodermic
syringe through the septum and a measured amount injected
30
-------
through the rubber partition on the exhaust port. When
calibrating for ammonia it was found necessary to saturate
the test cell walls with ammonia m advance of the calibra-
tion test to prevent rapid adsorption of the test gas sample
onto the cell walls. This step was not necessary with
ethylene.
After fifteen minutes of data recording, the exhrust blower
and exhaust port were uncovered and the system flashed
with outside air. This exhaust system removes wore
than ninety percent of the gas in the cell in ten seconds.
After flushing for two minutes, the blower was Lhut off,
the ports closed and the aero condition of the system
recorded.
b. Laser Baseline Measurements
Field data was recorded by strip chart recorders which
continuously monitor the signal outputs from the absorption
meters and the synchronous detector outputs during each
test run. In addition, the average peak-to-peak signal
return from the retroreflector was recorded on gome
tests. All meteorological data as indicated on the data
format was recorded by hand.
Following calibration, the procedure was to first record
any change in balance since the previous test, turn on the
chart recorders and then record the meteorological data.
The frequency with which weather information was re-
corded depended on how rapidly conditions were changing.
The laser was then peaked up by adjusting the vprnisr
aiming mechanism (which directs the angle of the trans-
mitted beam in azimuth and elevation) to determine if the
unbalance was in any way related to an optical misalignment.
Any change in balance caused by peaking was then recorded.
The balance control dial settings were recorded, the sys-
tem was balanced by resetting the dials and the new settings
were recorded. These balance dials compensate for any
spectral absorption signal due to fixed elements in the
optical path such as lenses, coated mirrors and polyethylene
windows. After balance, the four synchronous detector
output meters read zero. The system was then left on with
the recorders running for a minimum of two hours. During
this time, surveillance was maintained and any changes in
operating conditions that might affect the system were noted
on the chart paper along with the time. The chart speed is
one millimeter per second on the narrow chart paper and
1/4 mm/sec on the wide paper unless otherwise noted on the
chart paper. At the end of the run, the recorder sensitivity
31
-------
setting (volts/centimeter), the time and the ciuta were
recorded. Weather conditions were recorded agun only
if conditions had changed during the run.
In addition to the meteorological data already described, an
attempt was made to measure the temperature gradient in the
air near the optical test cell for the laser transceiver. Six
thermometers were placed every two feet on a vertical ten-
foot pole and shielded from heat radiation from the ground,
sky and sun. The thermometers were read during conditions
of both high and low optical noise (scintillation). Less than
1/2 0 F was observed.
c. Supplementary Measurements
Independent measurements of ethylene concentration were
made on ambient air samples taken at the two ends of the
two-mile test path. Meteorological conditions w^re also
measured at each end of the path including temperature,
wind speed and direction, cloud cover, humidity, subjective
measures of atmospheric scintillation and solai conditions;
and time was recorded for all tests. (See IV - Data Recording.)
Independent measurement of ethylene concentration was made
as follows:
1) Air samples were taken with a one-liter Hamilton syringe
and pumped into a Saran sample bag. Sample size was
about five (5) liters. Anspec Company 12-liter bags were
used.
2) Filled sample bags were returned to the Technical Services
Laboratory. Sample bags were kneaded briefly to insure
mixing of the sample constituents and a 1 to 10 milliliter
sample extracted with a 10-ml Hamilton syringe. Sample
extraction was at room ambient temperature.
3) The extracted sample was injected into the gas chromatograph.
The peak height of the ethylene peak on the GC output
recorder was used to indicate the concentration of ethylene
in the sample.
4) The GC was calibrated before and after sample measurements
using a standard gas mixture purchased from Matheson
Company. The standard gas contains 5 (five) parts per
million by volume of ethylene gas in dry nitrogen gas.
This standard gas is diluted with dry nitrogen gas to
obtain standard concentrations of ethylenn in concentra-
tion ranges of interest, i. e., comparable to fhe samples
being measured.
No satisfactory independent test for ammonia concentra-
tion at the levels of interest was found.
32
-------
2. Urban Site
a. Laser Measurements
The urban measurements procedure was identical to the
procedure used at the rural site except for two differences.
The test cell was six inches long instead of twenty-seven feet
and an absolute zero reference calibration was made to
isolate any spectral effects other than those caused by
absorbing materials in the air.
The six-inch test cell was calibrated against the twenty -
seven-foot test cell at the rural site to insure that no unfore-
seen effects take place such as signal variations caused by
changes in concentration for the same optical xhickness of
the gas. This six-inch-long cell also had polyethylene
windows and a rubber partition for sample injection.
Absolute calibration of the system was established using
a portable retrotelescope at approximately five hundred
feet (a short path) and comparing the long-path and short-
path signals. By making the short path optically similar
to the long path, the only differences in the two conditions
is the thickness of the air sample measured (see page 47).
By extrapolating the two readings obtained to zero air
thickness, a zero baseline is obtained. This calibration
technique was tested at the rural site before moving into
the urban area.
At the urban site, measurements were made of both
ammonia and ethylene; however, only the latter was used
for analysis and system performance evaluation. For
ammonia we correlated readings with known tost cell
concentrations.
b. Supplementary Measurements
Once the baseline was established, calibration and meas-
urements proceeded exactly as at the rural test site. The
supplementary measurements were taken at the end points
of the test path and at several points along the path.
D. DATA RECORDING
1. Data Format
A standard data sheet for recording measurement is shown in
Table IX. In addition to the data sheet, an 8-channel strip chart recording
was made of the variables vs. time listed on the page following Table IX.
33
-------
TABLE DC. TANDARD DATA RECORDING SHEET
LASER AIR CONTAMINANT MONITOt
Test Number
Date
Location
General Weather Conditions
Precipitation
Visibility
% Cloud Cover
Local Weather Conditions
Transmitter Retroreflectoi
Time
GC Ethylene Measurements
Wind Speed/Direction
Temperature
Humidity
Calibration and Measurement
3
Cubic Centimeters of Gaseous Ethylene Injected _ cm
3
Cubic Centimeters of Gaseous Ammonia Injected cm
Synchronous Detector Relative Transmission
9.505u
10.532n>
10.675^
10.7 3 -1^
Absorption Meter
Ethylene
Ammonia
Comments
Before
Peaking
After
Peakina
34
-------
Synchronous Detector Output Voltages:
a. 9. 505)i wavelength
b. 10. 532/i wavelength
c. 10. 675/i wavelength
d. 10. 719jj wavelength
Absorption Meter Voltages:
e. Ethylene
f. Ammonia
Reference Voltage-
g. Average received laser power
h. Spare channel (seldom used)
2. Ambient Conditions
Consistent with available weather conditions during the test periods,
data runs were made during daytime, dusk and nighttime for as
wide a variety of ambient weather conditions as possible. It was
not possible to obtain data for all the combinations of weather
conditions.
At least 2 hours of continuous data were taken for each data run
and a minimum of 24 hours total of data were taken, ^or each
of the conditions listed below, data were taken at midday, dusk,
and after dark.
a. Clear sky - less than 25% cloud cover. Wind zero to five mph
b. Heavy overcast but no observable precipitation. Wind zero to
five mph
c. Clear sky - less than 25% cloud cover. Moderate to high
winds - fifteen mph or greater
d. Heavy overcast with moderate to high winds
e. Precipitation - light snow flurries and high winds
f. Precipitation with low wind conditions
I
35
-------
E. DATA ANALYSIS
The objective of the data analysis was to evaluate the performance of the
laser air contamination monitor under a variety of environmental conditions
and to determine which of those conditions influenced the behavior of the laser
system. Two kinds of error were measured in this evaluation. One was
short-term signal fluctuation or noise such as is produced by system noise
and atmospheric scintillation. The other was long-term drift or apparent
absorption signal which is undistinguishable from absorption sipnal. These
two noise phenomena are different only in the length of time involved, but the
distinction is convenient. The data is presented in graphical form, and also
tabular form, to show the correlation between the following parameters:
Signal-to-noise vs. atmospheric scintillation
Signal-to-noise vs. signal level returned (atmospheric transmission)
Signal-to-noise vs. each of an assortment of weather conditions
Long-term drift or signal unbalance vs. weather conditions
Long-term drift or signal unbalance vs. scintillation
Long-term drift or signal unbalance vs. signal level
Noise and drift vs. time between measurements
36
-------
VI. RURAL TEST SITE MEASUREMENTS PROGRAM
A. SUMMARY OF RURAL TEST RESULTS
At Cazenovia, New York, the laser site and the retroreflector site were
each located on top of a ridge 9, 800 feet apart and separated by a valley ap-
proximately 500 feet deep. Data were taken from November 1970 through
April 1971 under the full range of weather conditions available during this
time.
Over the five-month period, more than sixty hours were spent recording
or attempting to record data. These data were taken under both clear sky and
overcast sky, with and without precipitation. Atmospheric scintillation was
measured by fluctuations in the return signal level, both on the oscilloscope
that monitors the return signal directly, and also on the signal level indicator
that is recorded on one strip chart. The signal level indicator recorded
variations in the average signal level with a ten-second time constant. Scintil-
lation was highest during the conditions of clear sky, at mid-day and, on two
occasions, around midnight; whiJe at dusk, the scintillation was minimum.
The correlation between the scintillation magnitude and errors or pseudo-
absorption signals on the signal processor output was high.
During conditions of maximum scintillation, we detected five carts
per million of ethylene and seven parts per million ammonia in the wenty-
seven-foot test cell. The measured absorption signals for these concentrations
were ten times the RMS error occurring during times of maximum scintillation
and atmospheric turbulence. To produce the same absorption signal, the
average concentration over the entire ten-thousand-foot range over which we
were operating would require only fourteen parts per billion of ethylene and
nineteen parts per billion of ammonia.
The data taken at the rural site clearly indicate the following about the
limit of the detectability of the system:
2) The accuracy with which we could reliably measure the
average concentration of ethylene and ammonia over the path
was determined by the optical noise and scintillation produced
by atmospheric turbulence and temperature gradients in the
path.
2) All the recorded data taken with both test cells confirmed
the fact that the calibration of the system never varied.
The inside diameters of both test cells, the 6-inch cell and
the 27-foot cell, were 6 inches. One milliliter of gas, either
37
-------
ethylene or ammonia, injected at atmospheric pressure, produced
an optical thickness (concentration times path length) of 0.0055
atmosphere-cm. Each milliliter of ethylene always yielded
a CL = 0. 25 and each milliliter of ammonia always yielded
a CL = 0.15. Temperature effects due to change in density
with temperature and doppler broadening were too small to
measure.
3) The signal Level returned from the retroreflector two miles
away was more than two-hundred times larger than that needed
for good operation of the system. Attenuation of the signal by
rain, haze or snow, up to two-hundred to one, had absolutely
no effect on system performance.
4) When the attenuation was greater than 500-1, or rather when the
signal level at the output of the preamplifier dropped to less than
100 millivolts, the system became inoperable because of internal
system (detector) noise.
B. RURAL TEST SITE
The rural test site is located at General Electrzc's Antenna Test Facility
on U.S. Route 20, about 15 miles southeast of Syracuse, New York, near the
town of Cazenovia. It consists of a transmit site and a retroreflector site,
each located on top of a ridge, and separated by a valley approximately 500 feet
deep. Figure 14 shows the profile of the terrain separating the two locations.
The 9, 800-foot separation between the transmit site and the receiver site is
an excellent range for evaluating the ILAMS performance. The sparse rural
population, remote from any major sources of air contamination, provides
a relatively clean atmosphere in which to measure data independent of the inter-
fering effects of pollutants. All the ethylene point measurement made at this
Cazenovia test site registered zero ethylene. This indicates that the concentra-
tion was below 5 ppb. Since there were few roads and little traffic in the area,
this result is not surprising. Figure 15 shows an aerial view across the valley
as seen from above the ILAMS transceiver.
The ILAMS transceiver was located in a heated, insulated room in a pole
barn with a concrete floor; see Figure 16. The beam was transmitted through
a 0.0006" polyethylene window, and then through the twenty-seven-foot gas
cell and across the valley to the retroreflector. The gas cell can be seen in
Figure 17.
The retroreflector was mounted on the eastern side of the valley at the
same altitude as the transceiver, in a small unheated building with an open
window. A shield was placed over the window to keep snow fi om blowing into
the retrotelescope.
C. RURAL TEST RESULTS
The goal of the rural test program was to evaluate the system performance
in a clean-air environment. The objectives were to study and measure the be-
havior of the ILAMS while it was isolated from the effects of airborne contaminants
and particulates, and thereby provide a baseline from which to judge the data
taken in the urban environment. There are two major differences between the
rural and urban environments that would influence the system performance.
38
-------
Transmit
Site
Retroreflector
Site -|
1400
1300
1200
Q
5- 1100
d
55 1000
I
.5 900
<
& 800
700
Figure 14. Rural Test Facility Profile
Figure 15. Photo of Cazenovia Valley
This page is reproduced at the
back of the report by a different
reproduction method to provide
better detail
39
-------
Figure 16. ILAMS Transceiver and Recording Equipment at Rural
Test Site
Figure 17. 27-foot Gas Cell at Rural Test Site
This page is reproduced at the
back of the report by a different
reproduction method to provide
better detail.
-------
First, the urban airborne contaminants might have unknown spectral absorption
patterns that would produce false signals. Secondly, the atmosphe -ic turbulence
that produces scintillation of the return signal is more severe in i/>e urban
environment, particularly when the path of the monitoring beam passes over
large buildings and industrial sites.
The emphasis in the rural measurements program was tn me^sare the noise
and drift in the system output. Noise and drift represent the measurement error
and establish the detectivity or minimum detectible signals 'hat c?n be used
reliably as an indication of average pollutant concentration alon^ t>, e monitored
path. These errors were measured and correlated, where posi.ib e, to the
environmental effects that produced them. In some cases, th-3 mechanism
by which such noise and drift were produced was studied so as to ( ) improve our
confidence in the validity of the results and (2) project improvable ts in the
system and estimate the potential detectivity of the post-breadboavd system.
The limit of the detectability of the system, that is, the accuracy with
which we could reliably measure the average concentration of ethj ene and
ammonia over the path, was determined by the optical noise and scintillation
produced by atmospheric turbulence and temperature gradients in he pattu
Spectral attenuation — attenutation at some wavelengths more thai, others --
is the information used by the ILAMS to measure the amounts 01 p rticular
gasses in the air. Attenuation due to absorption by a known gas o' curs in
predictable spectral patterns and is used to identify or discrimina e against
the gas. Random attenuation, such as that caused by atmospheric scintillation
or by internal system noise, represents an error and limits the sensitivity of
the system.
The spectral attenuation can be produced by the atmosphere r several
ways. The measured intensity distribution of a cross-section of t' e laser
beam at the output of the beam expander shows that the pattern is '>ot as
symmetrical as one would expect for a TEM00 mode. Contours d awn through
areas of constant power density, in Figure 18, show difference? i~. the beam
patterns in the near field for each of the four wavelengths; consequently, the
far-field patterns would also exhibit differences. Therefore, amplitude modulation
of the signal is produced by translation or distortion of the beam pattern at the
retroreflector, as might be caused by atmospheric effects. If the patterns are
different for the different wavelengths as is indicated by the measured difference
at the beam expander -- then the modulation at each wavelength is different
and such modulation is an error -- that is, it cannot be distinguis ed from
an absorption pattern. Atmospheric scintillation due to air in the mmediate
vicinity of the retroreflector can have similar adverse results on lie illumination
of the receiver. The data that have been accumulated indicate that it is these
atmospheric effects, particularly those near the beam expander and near the
retroreflector, that are the dominant effects responsible for the noise and
drift recorded on our data.
In order to reduce the scintillation to reasonable levels, it whs necessary
to isolate the test cell from sunlight and reflected sunlight and lor ite the cell
away from the building so that the cell window did not come into c. ntact with
room air at a temperature different from that of the air inside the cell. A
separate polyethylene window was installed across the window of Hie room
41
-------
Figure 18. Contours of Equal Power Density at the Beam
Expander Output
42
-------
through which we transmitted and received, to prevent mixing of room air and
outside air in the path of the laser.
Two types of retro reflectors were used with grossly different results. A
mosaic of thirty small cube corners with a total capture area of one square foot
was used initially. This reflector resulted in one-hundred-percent modulation
of the energy returned to the system at each wavelength on a clear but windy
day, and the modulation (or scintillation) at each wavelength was apparently
independent, producing large errors. In other words, the scintilla ion was
highly spectral. The term "scintillation" as used here refers only to
fluctuations in the peak amplitude of the signal at the output ot the pre-
amplifier as observed on an oscilloscope. The dual-beam oscilloscope is
synchronized to the chopper wheel so that one sweep is produced for each
rotation of the chopper. The reference (transmitted) energy and the return
signal were presented together on the separate traces, so that the two signals
could be compared rapidly on a wavelength-to-wavelength basis. Tne output
of the reference detector (as seen at the output of preamplifier B) remained
constant, while the signal return fluctuated in amplitude. When a nearby
retroreflector, mounted fifty feet from the transceiver, was used, no amplitude
fluctuation occurred. The retrotelescope across the valley was mounted on a
concrete pedestal in the ground. Therefore, the amplitude fluctuations were
attributed to atmospheric turbulence and are called (for convenience) scintillation.
Two types of scintillation effects, spectral and non-spectral, w^re noted.
Non-spectral amplitude fluctuations affected all wavelengths together, like a
rapidly changing neutral attenuation. This type of scintillation is not recorded
by the system because of the action of the AGC and also because the weights are
chosen to discriminate against neutral attenuation. Therefore, the non-
spectral scintillation produced no error*. The spectral type of scintillation did
produce system error. The short-term spectral fluctuations were averaged
out by the three-second integration time of the instrument, while tYe longer-
term fluctuations appeared as false absorptions. When both kinds of scintillation
occurred together, as they usually did, a rough estimate could oe made as to
their relative magnitudes, by observing the oscilloscope.
It was found that, by shaking the cube-corner mosaic back ar.d forth either
vertically or horizontally across the beam, the scintillation magnitude could be
reduced by a factor of three to one, and the spectral component only accounted
for about fifty percent of that modulation. In other words, while shaking the
cube corners, the modulation of the signal due to scintillation was about 35%
rather than 100% and only partly spectral. Moving the cube-corner mosaic
towards and away from the transceiver at the same r?te yielded no improvement
over holding it still. The improvement is attributed to time averaging of the
interference fringes as they change rapidly with cube-corner motion.
A substantial improvement (20% modulation under the same weather
conditions) in the magnitude of the scintillation noise was achieved by using
a 12" aperture telescope with a mirror at the image of the laser transmitter as
a retroreflector. One such retroreflector was described, in Section III, as a
"cat's eye" retroreflector. Two other telescopes have been used i.s retro-
reflectors in the course of these experiments. A 12"-aperture Dail-Kirkham
* This statement is not precisely true. If some of the scintillation noise occurs
at the chopping frequency, 75 Hz ± 0. 2 Hz, it will produce an error. This
effect is relatively small.
43
-------
witn 3. tnree-incn-aiameier bt?cunud.i y
-------
Such detectors are available in thermistor bolometers as well in cooled photo-
conductors. Reduction in signal was often caused by rain, tog, or snow. The
retrorefLector was mounted in a small white building across fhe valley. We
found that, in rain or snow, the system would operate at optimum detectivity
until the building was no longer visible with the naked eye. If we could not see
the two miles across the valley, the return signal from the retioreflector
was usually too weak for reliable operation of the system. Under conditions
of fog, however, the infrared transmission held up until the visibi'ity was
less than one mile, indicating that a large percentage of the fog particles
were small enough (below ten micrometers diameter) to produce less scattering
at ten microns than raindrops or snowflakes.
The dynamic range of this system can be expressed in terms of the
response to concentrations of the target gases in the test cell. Tf 3 system
responds linearly to ethylene gas from a minimum detectable concentration of
one part per million in the twenty-seven-foot test cell (2. 7 ppb over the two
mile range) to 53 parts per million in the test cell (144 ppb over t"ie two mile
range). The corresponding concentration range for linear detection for ammonia
is 1. 5 ppm to 80 ppm in the test cell, or 4 ppb to 216 ppb over a two-mile
range. The linear range is limited by the electronics and not by saturation of
the absorbing medium. The maximum signal that the electronics can tolerate
for linear operation is 12 volts.
It is significant that, except for loss of signal because of wea'her conditions
heavy enough to obscure good "seeing" of the retrorefleclor site, scintillation,
and those atmospheric conditions known to contribute to scintillate w\, were the
only factors that correlated with the signal-to-noise of the system. The data
recorded indicates that the signal-to-noise was aggravated by sun loading of
anything along the active path between the laser system and the reiroreflector.
Windows, building walls and the test cell were particularly susceptible. During
periods of high wind, overcast sky or dusk, the scintillation magnitude was
low. The improvement in system signal-to-noise, from times of high scintillation
to times of low scintillation, was approximately five to one. Ac one point, we
attempted to stop occasional water condensation on the retrotelescope by wrapping
heating tape around the primary mirror. This had disastrous results on the
magnitude of the scintillation noise and had to be discontinued. Temperature
and humidity appeared to have no effect on signal-to-noise, and wind velocity
had only a small effect (that of reducing the scintillation) under cloar weather
conditions during the day or late at night.
To present a clearer understanding of exactly what data wtie recorded
and how, a reproduction copy of one of the chart recordings is included as
Figure 19. This recording was made in April, during conditions of poor
visibility. A light drizzling rain was falling, and the valley was sufficiently
foggy for visibility to be a maximum of one mile and sometimes ynly one-
half mile. The eight lines each represent a separate recording cnannel, as
follows:
1. Average signal level returned from the retroreflector.
The left-hand edge represents zero signal.
2. The log of the transmission of the path at 10. 532 ^m.
45
-------
3. The log of the transmission of the path at 10. 719 /im.
4. The log of the transmission of the path at 10. 675 /im.
5. A weighted sum of the outputs of the four wavelength
channels, designed to respond to ethylene and discriminate
against ammonia, water vapor, and ozone.
6. A weighted sum of the outputs of the four wavelength channels,
designed to respond to ammonia and discriminate against ethylene,
water vapor, and ozone.
7. A weighted sum of the outputs of the four wavelength channels,
designed to respond to ethylene and discriminate against ammonia
and the small measured spectral unbalance associated with changes
in the fog density across the valley.
8. The log of the transmission of the path at 9. 505 p
Time proceeds from the bottom of the chart to the top. The horizontal grid
lines each represent twenty seconds. The elapsed time of this chart recording
is fifty-five minutes. At somewhat irregular intervals of approximately three
minutes, we injected one-half of a milliliter of ethylene gas at atmospheric
pressure into the test cell up to a maximum of eight milliliters. Each one-
half of a milliliter represents a concentration of 3. 33 ppm in the 27-foot test
cell, or 9 ppb over the two-mile range. Alternately, we injected one milli-
liter of ammonia gas at a time, also at atmospheric pressure,
-------
midnight under clear skies; while, at dusk, the scintillation was minimum
both before and after dark. The correlation between the scintillation magnitude
and error or pseudo-absorption signals on the recorded outputs from the signal
processor output was high.
During conditions of maximum scintillation, we were consistently able to
detect five parts per million of ethylene and seven parts per million of ammonia
in the twenty-seven-foot test cell. The measured absorption signals for these
concentrations were ten times the RMS error due to scintillation and atmo-
spheric turbulence. To produce the same absorption signal, the average
cr/jcentration over the entire ten-thousand-foot range over which we are
operating would be fourteen parts per billion for ethylene and nmeteen parts per
billion of ammonia.
During the day- and night-time measurements, thermometers were set np
along a support structure shielded from thermal radiation. The purpose of tl.^s
experiment was to measure the vertical temperature gradient across the
monitoring beam during conditions of both high and low scintillation. Four
thermometers were located at one foot, four feet, seven feet and ten feet from
the ground. The thermometers are bulb/capillary alcohol thermometers.
Because relative temperature between thermometers is of primary interest,
they were calibrated against each other, but not against an absolute standard.
The data taken indicated less than one-half degree difference over the ten-foot
path, which La low compared to temporal fluctuations in temperature over short
periods of time. The experiment was discarded as fruitless.
Two types of portable retro-telescopes (a single cube corner and a small
retro-telescope) were set up at approximately five-hundred feet from the
transmitter, at the rural site, in order to test the absolute calibration tech-
nique. The cube-corner retroreflector was shown to have pufficiently good
optical and diffraction properties for use at this close range. The interference
phenomenon that we observed with a cube-corner mosaic at long range was not
present to a measurable degree. Because the cube corner is much more
tolerant of angular misalignment than is the retro-telescope, it was chosen
for use in absolute calibration of the instrument. This cube cotiier, which
has a two-inch capture area, would be usable for absolute calibration out to
one-half mile. Under optimum weather conditions, we were able to operate
the system over the two-mile range with the same cube corner. The return
from the cube corner was steady, almost scintillation-free, and repeatable
from day to day. It was consequently judged adequate for calibration purposes.
The response of the ILAMS system to ethylene and ammonia in the gas
cells was always the same, to the limit of our measuring ability. The res-
ponse was linear with optical thickness (concentration times path length) and
independent of the size of the cell. The 27-foot gas cell was calibrated against
the 6-inch cell. Because they were both the same cross-section - - 6 inches
inside diameter -- one milliliter of ethylene always produced the same result,
(18% absorption — a CL = 0.20) in either cell. Similarly, one milliliter of
ammonia always produced 14% absorption ( a CL = 0.15). ^hese results
were independent of weather conditions, signal level returned, time of
day, etc. These environmental factors influenced the system noise, but not
its sensitivity nor its calibration.
47
-------
Table X summarizes the data taken at the rural test site in Cazenovia,
N. Y. , during the last twenty-four hours of measurement. More than fifty hours
of data were taken previously and are recorded. All charts are stored and
indexed. All the data confirm the conclusions that may be drawn from
this table. The data taken before March 18 were recorded on three small
recorders -- two 2-channel brush recorders and one 2-channel Sanborn
corder. Only six channels were available, and two of these were noisy because
of recorder malfunction. No signal level channel was recorded on these data,
so we had to depend upon operator comments to indicate low signal. Because of
these facts, and also because of the other considerations explained under the
section dealing with "measurement problems", the data taken after March 18
were considered to be the most accurate. The noise and drift data were read
directly from 8-channel recorder chart records such as that shewn in Figure
19. Because of the log amplifier, the meters read directly in terras of a CL.
However, for such small values as those recorded here, one may multiply by
one hundred and read them as percent absorption. Thus, an a CL of 0. 04 may
be read a6 a 4% absorption. The scintillation is the percent modulation as
observed on the oscilloscope. The correlation between weather conditions and
the observed scintillation bounce on the oscilloscope (Figure 20), presents no
surprises. On two occasions, the scintillation did peak up again, around ;
midnight, on a cold clear night with low wind. No ready explanation has been
offered to explain this phenomenon. No such effect was noted during the
urban testing. The correlation between the observed scintillation and the
noise and drift errors as recorded on the strip charts (Figure 21), is also
expected.
Where snow flurries are indicated on the data, one should not conclude that
the sky was overcast. The winter weather in the lee of Lake Ontario often
includes snow flurries together with clear sky, hazy sky, or scattered clouds.
Thus, we were often faced with loss of signal due to a snow squall in the valley
between the transmitter and the retroreflector while the sun was still shining.
Table X indicates that the noise and drift correlate very v.'ell. One may
see, by looking at the chart record data, that the noiBe and drift are just the low-
and high-frequency ends of a noise spectrum extending from dc to 0.2 Hz. The
peak amplitude of the low-frequency noise, or dc drift, is not measurably
different from the peak amplitude of the noise at the high-frequency end of the
spectrum. Any lack of correlation between {1) the noise error and the drift
error, or (2) between the scintillation as measured on the oscilloscope
(Figure 21) and the noise and drift error, is attributable to the aieasurement
inaccuracies -- that is, the accuracy with which we can measure noise from
the oscilloscope and from the chart record.
48
-------
TABLE X
LAST 24 HOURS OF DATA TAKEN AT RURAL TEST SITE IN CAZENOVIA, N. Y.
Run #
Temp
°F
RH
Wind
MPH
Scintil-
Lation
%
Modul-
ation
Sky/
Precip-
itation
Noise
(equivalent
a CL)
Drift
(equivalent
a CL)
Period of
Run
51
30-38°
60%
5-10
30%
clear
±0.04
±0.04
5:00 pm - 1:00 am
52
37°
37%
16
20%
hazy
(flurries)
±0.02
±0.02
1:00 pm - 2:00 am
53
39-31°
55%
15
20%
hazy
(flurries)
±0.03
±0.02
noon - 5:00 pm
54
39°
45%
5-10
40%
clear
±0.04
±0.05
3:00 pm - 4:00 pm
55
34°
50%
15-20
15%
flurries
±0.02
±0.03
1:30 - 4:30 pm
56
44°
31%
18-20
30%
clear
±0.02
±0.04
1:15- 3:00 pm
57
61°
31%
8-15
10%
cloudy
±0,01
±0.01
2:50 - 3:30 pm
58
53°
36%
25-35
20%
scattered
clouds
±0.02
±0.03
2:00 - 4:20 pm
59
41°
86%
12
10%
drizzle
±0.02
±0.02
2:25 - 3:30 pm
60
39°
96%
8-15
15%
light
rain
±0. 02
±0.03
2:10 - 4:35 pm
-------
1
Figure 19. Chart Record
51/52
3
-------
50
UJ
& 40
o
CO
o
o
<2 30
z
o
z
o
H
<
-I
o
o
2
20
10
CLEAR
SUNNY
HAZY
0
CLOUDY OR
DUSK OR
RAIN OR
DRIZZLE
I
CLEAR
NIGHT
CLEAR
WINDY
WEATHER CONDITIONS
Figure 20. Relationship between Scintillation and Weather
at Rural Site
53
-------
0
0 10 20 30
MODULATION OF SIGNAL %i
Figure 21. Relationship between Scintillaticu and
Noise & Driff
54
-------
VII. TRANSFER TO URBAN SITE
The selected urban site at Skytop is shown on the geological survey maps
accompanying this report. Four locations were tried and two finally used; these
are shown at the ends of lines radiating from the Skytop site. The seven-and-
one-half-mile path between Skytop and General Electric's Eleciron^cs Lab-
oratory proved to be too long for summertime operation. The attenuation due
to water vapor alone should have been between 20 dB and 30 dB. (Refer back
to Table V). A return signal was received from the retroreflector in clear
but hazy weather, but only marginal operation was achieved. The reflector
was then moved to the top of the Effluent and Chemical building at the
Onondaga County sewage disposal plant on the southeast end of Onondaga Lake,
which forms the northwest border of Syracuse. This reflector site was four-
and-one-half miles from the laser system and the line between them runs over
the center of the city. Although a large signal was received back from the
retroreflector, the scintillation was also extremely large -- much larger than
we expected. At the time, we thought that heating of the roof caused the
scintillation. In the light of more recent data, it is now suspected that possibly
a slight misalignment of the retrotelescope was the major cause of trouble.
In mid-August, the retrotelescope was moved to the top of the State Office
Building in central Syracuse, three miles from Skytop. Most of the data were
taken over this range. The shorter two-mile range, also shown, will be dis-
cussed under Urban Measurements (Section VIII).
Skytop is a hill owned by the University of Syracuse, overlooking the City.
Before transfer to Skytop, the ILAMS was installed in a sixteen-foot office
trailer, along with the recording equipment. Then the system was aligned and
tested over the rural range at Cazenovia. This proved to be a wise precaution.
Severe instability problems with the new installation were caused by motion of
the trailer. Efforts to stiffen the structure and support it on the ground with
rigid supports failed. Wind loading and changes in the weight load of the trailer,
such as people getting in and out and moving around inside, produced sigmfidantly
large motions of the laser system. The effects of such motion wi)i be discussed
in Section IX. The trailer was mounted on three pedestals plus one large
screw-jack adjusted so that the force on the four mounting points en the
trailor's steel frame supported approximately equal weight. Mirrors were
mounted at several points on the trailer, including on its steel frarae, and
angular motion of these mirrors was measured by the displacement of the
reflected light from a small helium neon laser over a 70-foot path. No point
could be found on the trailer frame or structure that did not move several
milliradians as a 165-lb person walked about inside the trailer or Rot in and
out of it. Mirrors mounted on the pedastals and on boards set between the
pedestals showed that these points did not move. We concluded that the
trailer could not be sufficiently rigidized to prevent flexing and distorting under
these small load changes. This problem was solved by drilling 3" diameter
holes through the trailer floor and building a steel frame through these clearance
holes. At Skytop, a shallow pit was dug in the ground directly beneath the laser.
55
-------
Approximately four inches of mixed sand and gravel were poured into the pit
and concrete blocks were set on this bed. The laser sits on the steel frame
and the frame sits on the concrete blocks. The steel frame is clear of the
trailer floor, so that there is no mechanical connection between the laser and
the trailer, except through the ground.
Figure 22a shows the nominal 16-foot office trailer in which we have
installed our Laser Air Pollution Monitoring system. A large fl?.t mirror is
shown mounted on a tripod in front of the trailer, which serves to reflect
sunlight into the remote retrotelescope for alignment purposes. Figure 22b
shows the equipment installed inside the trailer. From left to right, we have
the eight-channel Sanborn chart recorder, the four-wavelength signal processor,
and the laser breadboard system mounted on sand-filled boxes that are supported
above the floor from below the trailer. The photograph in Figure 22c shows
the laser breadboard directed through the test cell and through the window to
the retroreflector. Figure 22d shows the area of Syracuse, N. Y., being
monitored. The retrotelescope is located in the center of the picture just above
the tree line, three miles from the laser monitoring system. The tree line is
approximately 11/4 miles distant.
56
-------
b.
Figure 22. Urban Test Site Photos (continued)
_ This pug<3 is reproduced at the
^ ' back o( the repoi b\ a different
reproduction met lod to provide
better detail.
-------
¦4 -
& aH-
< rt,i •
r1
* **
tij s.v, :
# »'
d.
Figure 22. Urban Test Site Photos (concluded) This page is reproduced at the
back of the report by a different
leproduction method to provide
better detail.
58
-------
Vin. URBAN TEST SITE MEASUREMENTS PROGRAM
A. SUMMARY OF URBAN TEST RESULTS
Based on the excellent results with the test cells -- both the 27-foot cell
and the 6-inch cell -- in measuring ammonia and ethylene, and the poor success
in finding sufficient concentration of ethylene in Syracuse, the contract officers,
Mr. John Nader and Dr.H. M. Barnes,recommended that we go ahead with the
plans to monitor the air over Syracuse, N. Y., and use the test cells to measure
system sensitivity. Injection of measured amounts of ethylene and ammonia
gas into a 6-inch-long, 6-mch-diameter cell was used as the measure of the
ILAMS sensitivity. As expected, the sensitivity of the system was the same
as at the rural site.
More than twenty-four hours of data were recorded on chart paper over the
period of August through December, 1971.
Noise and scintillation were slightly greater (approximately 30%) over the
city than at the rural site. Drift was about the same on some days and unusually
large on others. Whether this drift was caused by a spectral absorber like
ethylene was not confirmed. This was the only departure in the benavior of the
laser system from that at the rural site.
B. URBAN TEST RESULTS
The majority of the effort on this program was spent on the rural measure-
ments program. It was felt that once the behavior of the system w*s well under-
stood, system performance at the urban site would match that at the rural site.
The emphasis in the urban measurements program was to look for differences in
the behavior of the ILAMS, if any, in the urban as compared to the rural environ-
ment. The only difference noted was that the scintillation was larger with an
attendant increase in the noise and drift error. This fact caused i 3 to look
more closely at the system, the monitored path, and the reflectory to isolate
the causes of the measured scintillation noise and determine how much of it
was inherent in the measuring technique.
The first retrotelescope site was the top of the Electronics Laboratory
building on Electronics Parkway north of Syracuse. This site is approximately
seven-and-one-half miles from the Skytop site, as may be seen on the enclosed
topographical maps. The reflector used was the 12"-diameter Tinsiey telescope,
which is a Dall-Kirkham optical configuration with a 12" elliptical primary
mirror, a 3" spherical secondary mirror located three feet in front of it, and
has an effective focal length of 16 feet. The focal plane of the telescope is
approximately four inches behind the hole in the primary mirror. A two-
mch-diameter plane mirror was placed near the focal point on a s" ding trans-
lation mount for focus adjustment. The focus was adjusted for peat signal
using citizen-band walkie-talkies for communication between the Laboratory
and Skytop. Sufficient return signal was received at Skytop for peaking, and
59
-------
we believe that an optimum focus was achieved in spite of the marginal operation
of the walkie-talkies at this distance. The signal return was, however, too
small for good operation of the system. The combination of the high scintillation
and poor atmospheric transmission regularly ran the signal level below 100
millivolts, causing large errors, as was discussed in Section VI. These
measurements were made in July, when the absolute humidity was high and the
power level return was pretty much as would be predicted from the extinction
coefficients of Table V. The 7. 5-mile test provided a good measurement of
the range limit of the system.
The same telescope was used for the installation on top of the Chemical
and Effluent building at 4. 5 miles range. This site was discontinued because
of excessive scintillation noise rather than weak signal. The reason for the
very high scintillation noise was never discovered. With hindsight, we now
believe that the retrotelescope was improperly focussed. At the time, it was
not realized that this was a critical adjustment with regard to scintillation
magnitude. Since the signal return was adequate, no attempt was made to
improve the focus. In fact, it was speculated that perhaps deliberate defocussing
of the return beam might improve scintillation, since slight defocussing of the
transmitted signal did improve the scintillation noise at the Cazenovia site.
The Tmsley retrotelescope was moved, in mid-August, to the top of the
State Office Building in Syracuse. All the recorded data, with the exception of
run #110, were taken between this site and Skytop. The path length is 3 miles.
Very frequently, we had difficulty maintaining the retrotelescope in
alignment and focus. Personnel working in the State Office Building would
visit it and it would be found pointed in the wrong direction and/or out of
focus. Although a great deal of time was lost in tracking down the source of
weak signal return, one benefit did accrue. It was noted that the defocussed
telescope often produced greatly magnified scintillation. A reexamination of
the optics showed clearly that the secondary mirror, which obscures less than
ten percent of the energy, produces a "shadow image" on the beam expander
at the transceiver unless the plane mirror on the retrotelescope is set so that
the return energy is focussed onto the beam expander. This focus point for
the flat mirror is the plane of the real image of the beam expander in the
telescope. With the flat mirror located at this point, the aperture of the beam
expander is re-imaged back on itself by the retrotelescope. This focus is very
critical. For example, if the plane mirror is moved 0.1 inches forward, then
the return energy will exit via the retrotelescope collimated with a doughnut-
shaped cross-section. This is the very condition we are trying to avoid.
A second retrotelescope reflector was mounted on a tripod, aligned, and
calibrated against the reflector now mounted on top of the State Office Building
in Syracuse. Calibration was accomplished by bringing the new reflector to the
top of the state off :e building and taking data from the two telescopes
separately and together. There was no spectral difference between
the two reflectors, which was our main concern. The new telescope was then
moved to the top of the Syracuse University mens1 dormitory which lies almost
exactly on a line with the transceiver on Skytop and the first reflector. The
new reflector is just two miles from Skytop. This enabled us to obtain a
60
-------
separate sampling of the first two miles and third mile. A second advantage
of comparing the two-mile and three-mile systems is that they are optically
identical.
During the calibration of the two telescopes against one another, the
two were placed together -- one directly beneath the other -- and the trans-
mit beam from the beam expander was defocussed slightly to accommodate
both reflectors together. The result was no apparent increase in the return
signal, but a large increase in scintillation, similar to the results we obtained
at Cazenovia when using the cube-comer mosaic. A tentative explanation for
this phenomenon is coherent interference between the two returns.
The second retrotelescope is a Criterion with a twelve-inch f/6 parabola
for a primary mirror, with a two-inch-diameter 45 degree mirror mounted on
a spider five feet in front of the primary. A one-lnch-diamete^ flat mirror was
used at the focus point to return the energy. The focussing tolorance must be
better than 0.02 inches with this telescope at two miles, as compared to 0.1
inches with the Tinsley at three miles range. Similar scintillation problems
were experienced with the focus of the Criterion retrotelescope, in spite of the
fact that the obscuring secondary mirror is only 2 inches in d.ameter.
A single cube corner, masked down to l/4 inch aperture, waB mounted on
a 14-foot mast, 150 feet in front of the laser transceiver, for calibration
purposes and also to measure performance with a very nearby target. In
terms of angle, the cube corner was halfway between the Criterion telescope
on top of the mens' dormatory and the Tinsley on the State Office Building.
The cube corner proved to be an extremely stable reference. No measurable
scintillation or drift occurred using this nearby reflector. However, the optics
of the system, using the nearby cube corner as a retroreflector, are different
from the optics with the retrotelescope in two important ways. The output
beam from the beam expander is normally focussed at a point halfway
between the retrotelescope and the transceiver. Since the diameter of the
retrotelescope is 2. 5 times the exit aperture of the beam expander, the beam
still does not overfill the retrotelescope. Minimum scintillation was observed
at the rural site for this focus. Because the beam expander is so much larger
than the cube corner, we cannot use the cube corner to simulate this con-
dition. Secondly, in order to focus the energy from the beam expander at a
point between the beam expander and the cube corner, it is necessary to back
off the germanium focussing lens approximately \ inch. Since the germanium
lens is operated off-axis in order to prevent reflections from the lens surfaces
getting back to the detector, backing off the lens changes the centering of the
output beam on the beam expander. This change in centering introduced the
type of masking error described in Section VI. C.
Table XI is a summary of the urban test site data recorded on chart-
recorder paper. The charts are indexed according to the run numbers. All
these measurement runs were made between Skytop and the State Office Building,
with the exception of run #110, which was made between Sl.ytop and the mens'
dormatory. Figure 23 shows the relationship between the scintillation and
weather. Figure 24 shows the noise and drift error versus scintillation
magnitude. Because of the large variance in the data, we do not feel that the
differences between these results and the rural results are very significant.
One major difference that does not show on Table XI, nor on the figures, is
61
-------
TABLE XI. URBAN TEST SITE DATA
Run#
Temperature
°F
RH
Wind
MPH
Scintil-
lation
%
Modulation
Sky
Precipi-
tation
Noise
(equivalent
a CL)
Drift
(equivalent
a CL)
Period at
Run
101
63
50%
10-15
25%
Hazy
Zero
±0.02
±0.02
3:20-4:00 pm
102
70°
40%
5
45%
Few
Clouds
Zero
±0.05
±0.04
2:00-3:30 pm
103
87°-74°
70%
5-10
25%
Cloudy
Zero
±0.04
±0.04
3:30-11:00 pm
104
75°
60%
8
40%
Clear
Zero
±0.04
±0.04
4:00-4:50 pm
105
65°-70°
75%
0-5
30%
Clear
Zero
—
±0.06
9:50 am-12:30
106
70°
55%
10-15
20%
Hazy
Zero
±0.02
±0.04
11:00 am-l:00 pm
107
60°
90%
10-15
15%
Cloudy
Light
Rain
±0.02
±0.04
10:00 am-2:00 pm
108
50°-40°
60%
5-10
35%
Clear
Zero
±0.04
±0.07
4:10-4:00 pm
109
35°-30°
60%
10
15%
Cloudy
Snow
Showers
±0.02
±0.03
1:15-4:15 pm
110
30°
85%
15
20%
Cloudy
Light
Snow
±0.03
±0.03
1:00- 5:00 pm
111
45°
60%
15-20
25%
C luudy
Zero
±0. C2
±0.04
ll:0C-nOon
112
39°
76%
10-15
35%
Clouuy
Drizzle
±0.04
±0.03
3-30 4:4o Dm
113
35°
40%
10
45%
Cloudy
None
±0.06
±0.03
2:50-3:30 pm
i
-------
50
ui
a.
o
o
in
O
u
(/}
o
z
o
z
o
40
30
< 20i
3
O
o
2
10
CLEAR
ANO
SUNNY
HAZY
OR
SCATTERED
CLOUDS
CLOUDY
WEATHER CONDITIONS
Figure 23. Relationship Between Scintillation and Weather at
Urban Site (Skytop)
63
-------
0.06
0.05
0.04
u. 0.03
0.02
0.01
X
•
~
r\
U 9
•
—
•
•
•
•
•
o
w %
•
f A
I KJ
•
•
- X
•
•
•
•
i
X
X
y* /
% y.
'— \J m \
•
J *
•
•
x DRIFT
— •
•
#
O NOISE
•
i
1
1
10 20 30 40 50 60
MODULATION OF SIGNAL ON SCOPE (%)
70
Figure 24. Relationship Between Scintillation and Noise & Drift
(Urban Site)
64
-------
that, on a few of the runs recorded (and several times not recorded on chart
paper), an apparent ethylene absorption came and went over a period of an
hour or two. This was either a local spectral absorption, or else a system
error. Neither hypothesis has been confirmed. The absorption signals
measured corresponded to aCh of from 0.1 to 0. 25. This would i eqmre
an average concentration of ethylene of from 13 ppb to 32 ppb over one mile,
or an average concentration of from 675 ppb to 1, 700 ppb, spread over 100
feet. We found no indications of such concentrations in open areas in Syracuse
C. SYSTEM IMPROVEMENT WITH A CLEANUP APERTURE
As was pointed out in Section VI., the lack of similarity in the spatial
power distribution of the exit beams at the four wavelengths, as shown in
Figure 18, is the main source of error in the ILAMS. The conventional
solution to this problem is to introduce a spatial filter in the form of a very
small circular aperture (pinhole) at a focus point of the beam before it is
transmitted. Unfortunately, the design of this laser system is such that the
introduction of such an aperture is not easy without some modification of the
external optics. The focus point of the germanium lens immediate'y pre-
ceeding the beam expander will not suffice for this purpose, fo~ three reasons.
First, a large percentage of the energy rejected by the aperture will be
reflected directly back to the signal detector as a false signal return:
secondly, any variation in the transmitted energy through the aperture caused
by laser beam wander will not be compensated for by the reference detector;
and thirdly, if it is desired to defocus the transmit beam slightly, hen the
return signal from the retrotelescope will not focus at the same po nt as the
aperture and considerable energy would be wasted. In ciear weather over
moderate ranges, this last consideration is not as important.
The ideal location for the spatial filter is between the coupling mirror and the
beam splitter. If a pair of small, short-focal-length lenses could :>e Tiiounted
back to back in a confocal configuration in the small space available, then the
pinhole could be placed at the common focal point of the two. If such an
aperture is smaller than two wavelengths in diameter, or smaller .han the
diffraction pattern of an ideal laser beam at the focus of the lens, Ifren >t will
clean up or filter out any spatial variations in the pattern in bolh the near
field and the far field, except for variations in size. The variations in size would
occur because the diameter of the far-field diffraction pattern is proportional
to wavelength. The size variation would be a source oi error if the output from
the beam expander were focussed or the retrotelescope. Then the beam
patterns for different wavelengths at the telescope would be coneeiv.ric and
similar, but of different sizes. Thus, if the beam moved, an unoa. anced or
false absorption would be produced. This error can be reduced bj defocussing
the transmit beam from the beam expander, so that angular displacements of
the beam caused by atmospheric turbulence would not easily resulf in per-
centage changes in reflected return that are different from wavelength to wave-
length. The focussing of the energy return from the retroreflecUr- presents
the same problem. One might argue that, because light travels the round
trip to the reflector and back in a short time compared with the speed of
atmospheric changes, if the energy found its way to the reflector by some
crooked path, then it would follow the same path back and there would be no
scintillation effects interfering with the return trip. The fact is that thermal
gradients in the vicinity of the retrotelescope would cause the return beam to
65
-------
wander, because the energy enters and exits on opposite sides of the telescope.
This effect has been confirmed by experiment. The conflict between the need
for focussed return from retroteJescopes with obscuring secondary mirrors and
the need for defocussed return to eliminate scintillation effects because of the
different size of the focussed beams at each wavelength may be resolved by
using another type of telescope as a retroreflector. Either an off-axis parabola
like the beam expander, or a telescope with a very small secondary mirror
would make good retrotelescopes.
The modifications needed to test the spatial filter approach properly were
deemed too extensive to undertake during this project. However, a compromise
was attempted as shown in Figure 25. Two apertures, 0. 044-inch-diameter
and 0. 081-inch-diameter, were tried between the coupling mirror and the beam
splitter without the addition of any focussing optics. This approach had the dis-
advantage that a great deal of energy was thrown away in order to get a small
enough aperture to get some spatial filtering. It was not known whether any
improvement at all would be realized, because some spatial zoning could still
occur within the aperture. To be effective, the spatial filter or aperture must
look like a point source. That is, the aperture must be small enough so that
the transmit optics cannot resolve the difference between it and a true point
source of power. The apertures we used were as small as could be tolerated
and still have a sufficient signal return. If it had been possible to install the
focussing lenses suggested above, we could have used a sufficiently small
aperture with only 75% loss in power.
The results of the test were very encouraging. The reduction in scintilla-
tion noise was approximately three to one. The reduction in drift was four to
one and, for the first time, we seemed to have close to an absolute calibration.
Absolute calibration means that the relative power returned from wavelength to
wavelength is the same as the relative power from wavelength to wavelength
seen by the reference detector. There was no need to correct for any fixed
absorption pattern related to the transmit receive optics, because such an ab-
sorption pattern was reduced by approximately ten to one. It is suspected that
the remaining drift and spectral scintillation noise was related to the focussed
beam size differences plus the remaining spectral zoning.
COUPLING
p—j—y MIRRPH
T REFERENCE
1 DETECTOR
*
/I
TO
» BEAM
EXPANDER
GOLD
MIRROR
EXPERIMENTAL
CLEANUP
APERTURE
BEAM
SPLITTER
Figure 25. Location of Cleanup Aperture
66
-------
IX. SUMMARY OF SYSTEM PERFORMANCE EVALUATION
Several sources of error in making measurements were identified during
the course of the measurements program. These problems and their resolu-
tion are outlined here.
1. Severe Scintillation Induced by Local Heat Sources
Heat loading from direct and reflected sunlight on the test cell, room
heat on the polyethylene windows separating the test cell from the room, sun
loading on the west wall of the building housing the retroreflector, cold air from
the polyethylene window in and near the beam expander, and at one point heating
tape near the retroreflector —these factors were all source.-; of scintillation
producing thermal gradients. The thermal problems were partially solved by
the use of sun shielding on the retroreflector building and the test cell, and
multiple-layer polyethylene windows at the room-air to outside-air interface.
The scintillation associated with the twenty-seven-foot test cell was never
completely eliminated until the cell was removed.
2. Long Term Unbalance or Drift Error Produced by Changing the
Attitude or Pointing Direction of the Laser Transceiver and then
Correcting for this by Translating the Germanium Focusing Lens
at the Entrance to the Beam Expander
This correction moves the beam slightly in the exit aperture. Because
of the asymmetry of the transmitted wavelengths (see Figure 18), this transla-
tion of the beams on the beam expander mirror produces a change in the masking
by the mirror and a resultant signal unbalance. The cause of this motion of the
laser system was a mystery for some time, until it was discovered that the floor
was heaving from ice formation under the building edges and under the unheated
portions of the building. As spring came, this problem solved itself. Accidental
bumping of the laser system, which pushed it out of alignment, would of course
produce the same result unless it was corrected by moving the laser itself
rather than the germanium lens.
3. Unbalance Caused by Aiming Error from Transmitter to
Retroreflector
If a misalignment such as described in (2.) occurred, and was not cor-
rected for, by either moving the germanium or moving the laser, then a more
serious unbalance or drift error occurred. The lack of similarity of the different
wavelengths shown in Figure 18 has a corresponding lack of similarity and con-
centricity in the far field at the retroreflector. If the laser output beam wanders
with respect to the support frame, then any error so caused will be compensated
for by the reference aperture, because this aperture sees the same portion of
far field pattern as the retroreflector does. But if the aiming error is caused
by motion of the ILAMS frame structure, or by displacement of the beam on the
67
-------
retroreflector by scintillation (thermal gradients in the air), then a wavelength-
to-wavelength unbalance or drift error occurs. The magnitude of this error is
much larger than the masking error we get if we correct by translating the
germanium lens. During the rural measurements, no solution for this problem
was found. The drift error caused by scintillation in this way was judged to be
the dominant source of noise and drift error in the system.
4. Direct Coupled Amplifier Drift and Paper Slippage in the
Chart Recorders
The meters on the signal processor permitted observation of both the
synchronous detector outputs at each wavelength and the weighted outputs. It
was noted that, often after many minutes of running time, the chart recorder
outputs did not correspond to the meters. There were two causes for this dis-
crepancy: (1) the chart paper would slip sideways slowly on the rollers so that
all channels would appear to drift off together and (2) the dc amplifiers in the
chart recorder would drift so that their output voltage would change for zero
change in input voltage. Both of these effects produced pseudo-errors of more
than 10% absorption (a CL = 0. 1) at times. Checks on the recorder drift were
made by disconnecting the output from the signal preamplifier and then con-
necting the reference preamplifier to both AGC inputs. This produced zero
output on all channels with no drift. After checking the recorder drift, the
operator would measure the false error, make the correction and note it on
the chart paper.
5. Errors Caused by System Noise
System noise had no effect upon the ILAMS performance, except under
the following conditions. When the total effect of (a) poor transmission in the
atmosphere, (b) absorption by a specific gas, (c) loss by the sample cell
windows, plus (d) any instantaneous drop in signal caused by scintillation,
caused any of the four wavelengths to drop in power level to the point were the
preamplifier output voltage was 100 millivolts or less, then the system noise
caused the log amplifier to see a negative voltage resulting in such large output
errors from the log amplifier that the system was considered inoperable. Over
the two-mile range in clear weather, with no test cell in the path, approximately
500 times this signal level was available. Six-inch circles of copper window
screen were used as neutral attenuators, so that we could work with such large
signal return without overdriving the preamplifiers. The result was that, by
using these attenuators and removing them as the available signal dropped, an
attenuation of up to 500:1 (including scintillation) could be tolerated.
6. Unusually Large Scintillation Caused by Defocusing the Retrotelescope
Unless the retrotelescope is properly focused (not aimed), the shadow
of the partially obs< '.ring secondary mirror on the telescope is imaged back on
the beam expander aperture of the ILAMS. Hence, the transceiver would receive
no return energy except when scintillation was present. This sort of alignment
tended to maximize scintillation, producing 100% modulation of the signal return.
The problem did not arise at the rural site, because the retrotelescope was in a
«8
-------
well-protected environment and because there was an excellent communications
link between the two sides of the valley for use in alignment and focusing. De-
focusing of the retroreflector was a constant problem m the urban tests because
of some combination of wind and peoples' curiosity. The problem was solved
simply by realigning- whenever the retrotelescope appeared to be out of focus.
The ultimate solution will be to provide better protection for retroreflectors.
7. Detection of 14 ppb Ethylene, 19 ppb Ammonia
The system sensitivity under the worst conditions of scintillation and
drift was never worse than ana CL of 0. 1. That is, we could always reliably
detect an absorption of 10% at one wavelength compared to the others. The
resultant sensitivity to ethylene over the two-mile path was 14 ppb and to ammonia
19 ppb.
8. Projected Performance Capability
With a cleanup aperture added to the ILAMS, the expected sensitivity
is from an a CL of 0. 01 to 0.001, which should greatly extend the usefulness
of the system in measuring trace constituents in the atmosphere, even over
short ranges.
69
-------
REFERENCES
1. Philip L. Hanst, Private Communication, 1970
2. Yates and Taylor, NRL Report 5453, 1960
3. A. R. Curtis and R. M. Goody, "Spectral Line Shape and Its Effect On
Atmospheric Transmission," Quart. J. Roy. Meteorol. Soc., Vol. 80, •
p. 58 (1954).
4. Migeotte, M., Neven, L., and Swensson, J., "The Solnr Spectrum From
2. 8 to 23. 7 Microns, Part 1, Photometric Atlas, " Technical Final Report,
Phase A, Part 1, Contract AF61(514)-432 (1957).
5. "Water Vapor Continuum Absorption of Carbon Dioxide Laser Radiation
Near 10/li", John H. McCoy, David B. Rensh and Ronald Long; July
1969, Vol. 8, Ho. 7; Applied Optics, p. 1473.
70
-------
THE FOLLOWING PAGES ARE DUPLICATES OF
ILLUSTRATIONS APPEARING ELSEWHERE IN THIS
REPORT. THEY HAVE BEEN REPRODUCED HERE BY
A DIFFERENT METHOD TO PROVIDE BETTER DETAIL.
-------
Breadboard transmit-receive optics expand the laser beam u a diameter
of 5 inches which reduces transmitted power densities to values below 0.01 watt
per square centimeter while also contributing to beam collimation The "cat's
eye" style retroreflector is a 12-inch-diameter f/4 parabolic miri » with a 4b-
inch focal length and a one-inch-diameter,40-inch-radius-ol eurvai ire.concave,
spherical mirror at its focal point. The resolution of the retrorei lector is suf
ficient to return 90% of the incident laser energy to the receiver at one-mile
range. Such operational characteristics give the system an open r n>ge of
10 miles or more under conditions of "good visibility"
Figure 4 shows the present unit at the rural test site. While satisfactory
for current feasibility studies, these are breadboard components, with .subsiau
tial reduction in size possible. The V laser is mounted on an alu linuni channel,
6 feet in length. Repackaging the present analog functions will af >j d a con-
siderable reduction in size of the signal processing electronics, ^readiioaro
electronic components, excluding meters and substituting fixed r< ^lskir; and
trimpots for weighting potentiometers could be packaged ir a volu ne of about
2x4 inches. According to our studies, digital signal processing (cost
effective when the number of transmitted wavelengths exceeds six would not
have an adverse effect on processor size. A 10 x 40 inch laser breadboard
having comparable performance characteristics has also been dn k-|.ed by
under a separate program.
Figure 4. Breadboard System at Rural Site
7/8
Tins page js icpm-'"u-t'il at
liack of lid.' ri'juu I ,i i i, Jt i
reproduction rictlm to p.o>
better detail.
-------
1400
1300
1200
0)
> 1100
(D
J
rt
$ 1000
(1)
>
-a 900
<
-------
Figure 16. ILAMS Transceiver and Recording Equipment at Rural
Test Site
Figure 17. 27-foot Gas Cell at Rural Test Site
This page is reproduced at the
back of the report by a different
reproduction method to provide
better detail
-------
a.
b.
Figure 22. Urban Test Site Photos (continued)
This page is reproduced at the
back of the report by a different
reproduction method to provide
better detail.
-------
Figure 22. Urban Test Site Photos (concluded) Thjs page is reproduced at the
back of the report by a different
reproduction method to provide
better detail.
-------
NTIS Best Sellers
Just check your choices & return this announcement to NTIS.
Or call direct (703) 321-8543.
Mail to: N¦ Id
U S DEPARTMENT OF COMMERCE
PO Box 1553
Springfield, Va 22151
Please check one-
~ Charge my NTIS deposit account no
~ Here is my check to NTIS for $
~ Bill me (not applicable to foreign orders).
Add 50
-------
| | Slate of the Art for Controlling NOx Emissions, Part 1.
Utility Boilers
Catalytic, Inc , Charlotte, N C
Discusses emission control through combustion operating
modification, combustion equipment, design modifications,
and flue gas treatment Also gives information on the
sources and formation of nitrogen oxides 1972. 118 p.
$5.45 PB-213 297
Proposed Report of the National Water Commission.
Vol. 1
National Water Commission, Wash , D C
Presents tentative findings & recommendations for future
water policy on water & the economy, water pollution con-
trol, improving water related programs, and procedures
for resolving differences over environmental and develop-
mental values 1972. 478 p $6 PB-212 993
| [ Satellite Communications Reference Data Handbook
Computer Sciences Corp , Falls Church, Va
Surveys the background and present status of the satellite
communications field in general and the Defense Satellite
Communications system in particular with general and
technical data on the major subsystems of a satellite com-
munications network, contains number of monographs and
tables for analyzing satellite communications engineering
problems 1972. 279 p. $6.75 AD-746 165
j [ Fatigue and Fracture Mechanics
George Washington Univ , Wash , D C
Discusses different aspects of the fatigue process in rela-
tion to the basic concepts of fracture mechanics 1972. 33 p
$3.25 AD-746 122
[ I Afterburner Systems Study
from Shell Development Co for EPA
This handbook enables the user to decide if his particular
emission is amendable to afterburning and to obtain a
rough estimate of cost and size of the equipment needed
1972 512 p $6 PB-212 560
| | Guide to Technical & Financial Assistance for Air Pol-
lution Control
Gordian Associates, Inc , N.Y
Outlines Federal, state and local government financial
and tax assistance available to business and industry in
complying with abatement regulations and including a
catalog of state-by-state tax incentive rules. Nov. 1971.
147 p. $5.45. PB-210 670
| | Status of Advanced Waste Treatment
Environmental Protection Agency, Cincinnati, Ohio
Reviews the advanced waste treatment program and em-
phasizes developments that are ready for full-scale engi-
neering application May 1972, 80 p., $4 85 PB-213 819
[ | HUD Noise Assessment Guidelines Technical Background
Bolt Beranek and Newman, Inc , Cambridge, Mass
Discusses the need for noise abatement, various techniques
for measuring and describing noise and human responses
to it Gives technical background information for develop-
ing site noise assessment techniques 1971. 264 p. $4
PB-210 591
I I Proposed Report of the National Water Commission.
Vol. II
National Water Commission, Wash , D C
Presents remaining chapters on the following topics mak-
ing better use of existing supplies, intcrbasin transfer,
means of increasing water supply, better decision making
in water management, improving organizational arrange-
ments, water problems in metropolitan areas, federal-state
jurisdiction in the law of waters, sharing the costs of water
development projects, financing water programs, and basic
data and research for future programs 1972. 694 p. $9
PB-212 994
| | HUD Noise Assessment Guidelines
Bolt Beranek and Newman, Inc , Cambridge, Mass
Gives procedures for making assessments of present and
predicted noise exposure without technical training in
acoustics at sites proposed for new residential construc-
tion 1971. 36 p. $2.70 PB-210 590
| | A Guide for Developing Questionnaire Items
Human Resources Research Organization, Alexandria, Va
Presents a guide for the development of test questions for
questionnaire instruments other than achievement testing
Content is restricted to attitude and information type
questions 1972. 27 p. $3. AD-738 157
| [ Conservation and Better Utilization of Electric Power
by Means of Thermal Energy Storage and Solar Heating
University of Pennsylvania, Phila
Investigates the application of heat and coolness storage
for comfort heating and air conditioning 1971. 263 p.
$6.75 PB-210 359
| | Baling Solid Waste to Conserve Sanitary Landfill Space:
A Feasibility Study
San Diego City, Calif
Studies baling to reduce hauling distances, to fill small
canyons and to improve compaction Develops formulas
and techniques that could help other communities with
their solid waste collection and disposal problems 1973,
98 p , $4 85 PB-214 960
------- |