EPA-650/2-75-041
May  1975
Environmental Protection Technology  Series

                                             $$^
                                             ^

-------
                                     EPA-650/2-75-041
         INFRARED  SENSOR
FOR  THE  REMOTE MONITORING
                OF  S02
                     by

                 E. R. Bartle

                   SAI, Inc.
                P. O. Box 1393
             La Jolla, California 92037
              Contract No. 68-02-1208
                ROAP No. 26AAP
            Program Element No. 1AA010
         EPA Project Officer:  Dr. H. M. Barnes

          Chemistry and Physics Laboratory
         National Environmental Research Center
      Research Triangle Park, North Carolina 27711
                 Prepared for

      U. S. ENVIRONMENTAL PROTECTION AGENCY
        OFFICE OF RESEARCH AND DEVELOPMENT
             WASHINGTON, D. C. 20460

                   May 1975

-------
                         EPA REVIEW NOTICE
 This report has been reviewed by the National Environmental Research
 Center - Research Triangle Park. Office of Research and Development,
 EPA, and approved for publication.  Approval does not signify  that the
 contents necessarily reflect the views and policies of the Environmental
 Protection Agency, nor does mention of trade  names or commercial
 products  constitute endorsement or recommendation for use.
                    RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environ-
 mental Protection Agency, have been grouped into series.  These broad
 categories were established to facilitate further development and applica-
 tion of environmental technology.  Elimination of traditional grouping was
 consciously planned to foster technology transfer and maximum interface
 in related fields. These series are:

           1.  ENVIRONMENTAL HEALTH EFFECTS RESEARCH
           2.  ENVIRONMENTAL PROTECTION TECHNOLOGY
           3.  ECOLOGICAL RESEARCH

           4.  ENVIRONMENTAL MONITORING
           5.  SOCIOECONOMIC ENVIRONMENTAL STUDIES

           6.  SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
           9.  MISCELLANEOUS

 This report has been assigned to the ENVIRONMENTAL PROTECTION
 TECHNOLOGY series.  This series describes research performed to
 develop and demonstrate instrumentation, equipment and methodology
 to repair or prevent environmental degradation  from point and non-
 point sources  of pollution.  This work provides  the new or improved
 technology required for the control and treatment of pollution sources
 to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.

                Publication No. EPA-650/2-75-041
                                 11

-------
                            ABSTRACT
A prototype passive infrared sensor for the measurement of sulfur dioxide
emissions from stationary sources is described.  The infrared radiation
emitted by gases in a plume originating from smokestacks may be detected,
and from this the SC>2 concentration in the plume may be determined.  In
general,  the radiation received by the sensor is a function of the intervening
and background atmosphere.  Thus, the problem of quantitative  measurements
is generally complex.  A technique is described, based upon the principle of
Gas Filter Correlation, which minimizes these effects.

This report presents a detailed description of the sensor,  it's specifications,
and performance characteristics.  The basic unit is battery operated and
weighs only 10 kgms; thus,  it is readily portable. It's sensitivity  is presently
limited to about 70 ppm-m for source plume temperatures of 270 C and about
290 ppm-m for temperatures  of 170 C,  but this can be improved.

The results of field testing at both oil and  coal-burning power plants are com-
pared with extractive sample  data.  In general, the remote measurements
agree with the extractive data within ± 25 percent over SC>2 concentrations
ranging from 150 ppm to 1300 ppm from slant ranges of 130 to 400 m.

This report is submitted in partial fulfillment of contract number 68-02-1208
by JRB Associates, a division of Science Applications, Inc. under the spon-
sorship of the Environmental  Protection Agency.
                                 iii

-------
                          CONTENTS
                                                             PAGE

ABSTRACT	Hi

TABLE OF CONTENTS	iv

ILLUSTRATIONS	v

TABLES	vi

ACKNOWLEDGEMENT	vii


SECTION

    1     INTRODUCTION AND SUMMARY	1
         Statement of the Problem	1
         Brief Description of the Sensor	1
         Definitions, Symbols, and Units	2

    2     DESCRIPTION OF THE RGFC TECHNIQUE
         FOR REMOTELY DETECTING SO0	4
                                       £i
         Phenomenology	4
         Signal Calculations	6
         Theory of the  GFC Technique	9

    3     SENSOR DESIGN AND PERFORMANCE	14
         Mechanical-Optical Design	14
         Electronics Design	 16
         Theoretical Performance	18
         Laboratory  Test Results and Calibration	21
         Field Testing	29
         Discussion of  Field Test Data Reduction	32
         Conclusions	34

   4     REFERENCES	36


APPENDDC  - Optimization of Gas Cell Parameters	37
         Theoretical	37
         Experimental	42
                               iv

-------
                         ILLUSTRATIONS
                                                                  PAGE

FIGURE
   1      Schematic of viewing geometry                               5
   2      Theoretically computed radiances of hot N« -diluted SOg        8

   3      Schematic diagram of remote gas filter correlation sensor     10

   4      Optical system schematic                                   15

   5      Signal processing block diagram                            17

   6      Radiance difference detected by sensor  (assumes chopper's
         radiance  is a 300 K blackbody)                              20

   7      Radiometric mode calibration for V2R channel                22

   8      Radiometric data compared with theoretical  values            24

   9      Data obtained with sensor operating in GFC mode
         (T = 270 C)                                                25

  10      Data obtained with sensor operating in GFC mode
         (T = 170 C)                                                26

  11      Data showing sensitivity of ratio to SO2  temperature
         and concentration                                           27

  12      RGFC ratio mode calibration                                28

  13      Limiting  sensitivity of the sensor for  t  = 90  seconds           30

  14      Photograph of SO« remote sensor during initial field testing    31

  15      Comparison of remote sensing data with extractive data
         obtained from the DuPont analyzer and from EPA Method 6     33

  A-l    Effect of specifying and reference cell optical thickness
         on sensor sensitivity                                        41

  A-2    Theoretical calculation and measurement of  Ng-diluted
         SO« pressurized to one atm for SOg sensor                   44

-------
                          TABLES

                                                             PAGE
TABLE


  1     Summary of instrument parameters                        19

  2     Summary of remote SCL field measurements made at two
        oil-burning power plants                                 29

  3     Summary of. remote SO2 field measurements made at a
        coal-burning power plant                                 32
                             vi

-------
                       ACKNOWLEDGEMENT
 This report documents research and development performed by JRB
 Associates, a wholly-owned subsidiary of Science Applications,
 Incorporated,  under Contract 68-02-1208 between 25 June 1973 and
 28 February 1975.  The work was sponsored by the Environmental
 Protection Agency,  Research Triangle Park,  North Carolina.  The
 technical monitor was Dr. H. M.  Barnes, Jr.

 During this program valuable contributions were made by G. Houghton
 (mechanical-optical design), L. Acton (optical analysis),  E. Meckstroth
 (electronics design, instrument assembly, testing, calibrations,  and
field measurements),  G.  Hall (electronics consultation), Dr. W.  Malkmus
(theoretical programming), and,  especially, Dr. C. B. Ludwig (theoretical
calculations, field testing, data analysis, and reporting).
                                vii

-------
                            SECTION 1

                  INTRODUCTION AND SUMMARY
 STATEMENT OF THE PROBLEM

        The goal of the program was to design, fabricate, calibrate, field
 test, and reduce the data for an infrared sensor that remotely measures
 sulfur dioxide emissions from stationary sources.  The design goals were
 for the instrument to achieve the following performances:

        Remote Range: 100-1000 meters (slant range)
        Weight: less than 35 pounds
        Volume: less than one cubic foot
        SO2 Concentration Range:  100-1000 ppm
        Accuracy:  ± 5% for stack diameters of 1-10 meters.

        The phenomenology for remote detection of pollutants emitted from
 stacks is dependent upon specific spectral measurements of infrared radiation.
 The infrared radiation emitted by gases in a plume originating from smoke-
 stacks may be detected, and from this the SO2 concentration in the plume may
 be determined. In general,  the radiation received by the sensor is a function
 of the intervening and background atmosphere. Thus,  the problem of quanti-
 tative measurements is generally complex.  A technique has been developed,
 based upon the principle of Gas Filter Correlation, which minimizes these
 effects.

 BRIEF DESCRIPTION OF THE SENSOR

       The infrared sensor for the remote monitoring of SOo is based upon
 the technique of Gas Filter  Correlation described in Section 2 and elsewhere^   '
 in detail.  Briefly, it is a modification of the Non-Dispersive Infra-Red
 (NDIR) technique that has been known for some time.  It is based on the
 concept that a sample of gas provides a selective filter  for radiation absorbed
 by a polluted mixture of atmospheric species.  The radiation at the sensor
 is chopped so that it alternately passes through two optical paths, one through
 a cell containing the specific gas and one transparent.   Thus, the radiation
 is modulated only at the wavelengths at which the pollutant absorbs and high
 specificity results.

       It has been shown ^   that the signal generated by chopping between
two cells is a non-linear function depending upon the SO2 in the plume and
fixed instrument parameters and the difference between the radiance emitted
by the plume and background atmosphere.  By ratioing two GFC signals

-------
  obtained using different amounts of SO£ in the specifying cells, the effects
  of plume and atmospheric radiance are greatly minimized.  As is shown,
  it is desirable to operate the sensor as a pure radiometer also.   This is
  accomplished by blanking off the radiance from passing through the specifying
  gas cells.
        Extensive laboratory testing was conducted and the gas cell para-
  meters optimized.  The sensitivity limits of the sensor were determined.
  Field tests were conducted at two oil -burning power plants and one coal-
  burning power plant under a variety of weather conditions.  Extractive
  samples were taken and analyzed using EPA Method 6; these results were
  compared with the remote data and the two sets of data agreed within 25
  percent.
  DEFINITIONS,  SYMBOLS, AND UNITS
  a         ratio of spectral line half-width to line spacing
 aQ        "a" divided by the equivalent pressure of  the gas  (atm   )
 f         aperture  adjustment parameter,  see Equations (15) and (16)
 Af        noise bandwidth of the sensor (Hz)
 f(u), F(u)   defined by Equations  (A-8) and (A-12)
 k         monochromatic  absorption coefficient (atm  -cm  )
 k        mean absorption coefficient over a prescribed spectral interval
          AX
 *        integration limits, see Equation (1) [length]
 p        pressure (atm)
 t         integration time (seconds)
 u        optical thickness (atm -cm or ppm-m)
 x        integration length
 A, B      defined by Equation (A-6)
                           o
 Ad       detector area (cm )
 AQ       sensor entrance aperture (cm^)
 C        gas volumetric  concentration
 C(to)      normalized spectral definition function of the sensor
D*        detector detectivity (cm-Hz^/W)

-------
                                                 o
E        spectral energy arriving at sensor  (W/cm -(/m-sr)
L        optical pathlength (cm or m)
N        blackbody radiance (W/cm^-jjm-sr)
R        instrument responsivity (V/W cm'^-sr'1)
T        temperature (°C or °K)
V        voltage generated at the detector by the energy passing through
          either the reference or specifying gas cell (V)
AV       ac voltage generated by alternately passing the energy through
          the two cells (V)
W X
"'£'     defined by Equation (A- 17)
i, f->t
c         emissivity of a particular radiator
77         sensor overall efficiency
X         wavelength (jim)
AX       spectral interval
r         monochromatic transmittance of a particular instrument
          component or gaseous species
o>         wavenumber
v         vibration transition parameter
&Q        solid angle entrance to the sensor (sr)
       The sub and superscripts used are self-explanatory in the text.
       A bar over a parameter indicates  that it is the mean value over a
spectral interval AX.

-------
                            SECTION 2

                   DESCRIPTION OF THE REMOTE
              GAS FILTER CORRELATION TECHNIQUE
                  FOR REMOTELY DETECTING SO0
                                                ti

 PHENOMENOLOGY

        To illustrate the problem of remotely detecting SO2 in stack plumes,
 consider a sensor receiving energy from a plume with the sky as the back-
 ground (see Figure 1).  From the basic theory of radiative transfer^4),  an
 expression is developed which describes the monochromatic radiation re-
 ceived by the  sensor for the background, the effluent plume, and the inter-
 vening atmosphere:
E<*> = TPra/
                                (Tb«)
Ox
                                              dx
                                                       (1)
where the first term is the emission of the atmosphere from infinity to the
far edge of the plume, the second term is the emission of the plume, and
the third term is the emission of the atmosphere between the plume and the
sensor.

       N (T(x)) represents the blackbody function at temperature T, which
is,  in general, a function of x along the line of sight and, of course,  is also
a function of wavelength X.  The atmospheric transmission is indicated by
the terms ra and T^; it consists of the transmissivities of all of the  normal
atmospheric species, i. e.,
                                        x T
                                                       (2)

-------
                                             Plume /Background
                     Intervening
                     Atmosphere
                           i  i  i  i  i  'i  >• i   j
X«0
          FIG.  1.  Schematic of viewing geometry.

-------
       The transmissivity of the plume is similarly formulated

       T  = r (S02) x r (CO2) x r (H2O) x r (NgO) x f (CH4)x ...    (3)

This may be written as Tp= TTJ, where T is the transmissivity due to S02
and TJ is the transmissivity of all interfering species. The SO2 transmissivity
T is given  by

                           H
                 T = exp-/  k(X)C«pt (x)dx                 (4)
where k(X) is the spectral absorption coefficient of SO2,  C is the unknown SO2
concentration in the plume and pt is the total pressure of the plume.

       Consideration of Equations (1-4) shows that the task of quantifying the
SO2 concentration in the plume is complex.  The radiation received by the
sensor is a function of the atmospheric and plume temperatures as well as
the emissivity of SO2 and interfering species in the atmosphere and plume.
In principle, it appears that a number of simple radiometric measurements
over carefully chosen wavelength intervals and a suitable computer program
may be used to obtain quantitative results.  However, this is a very complex
procedure and generally not sufficiently accurate.  Therefore,  we have
devised a technique which is independent of the plume and atmospheric tem-
peratures,  and which does not require a computer to reduce the data.

SIGNAL CALCULATIONS

       Sulfur dioxide possesses many infrared-active bands, two of which
are the most promising ones.  These are the v± + v% combination band
centered at about 4 urn and the v-± fundamental band centered at about 8. 6 ym.
Although the band at 8. 6 urn is about 4 times stronger,  it is heavily inter-
fered by water vapor and— to a lesser extent— by CH4,  N2O and 03.   On
the other hand, only very weak interference occurs at 4  ym due to the presence
of N2O, CO2 and CH4 and the H2O continuum.  In addition, more sensitive
IR detectors are available for operation at 4 um than at 8. 6 ym.  Thus,
measurements at 4 nm are preferred.
       In order to gain insight into the spectral emission levels of hot
smokestack plumes containing SO2, we have calculated the radiance and
emissivity of plumes,  having various optical depths and temperatures.
These calculations were performed with our line-by-line  computer  program,
whose input parameters were  taken from the AFCRL atlas of atmospheric
lines and from data we have generated under different contracts^1).
                                 6

-------
        From the listing of the SC^ line parameters, it is determined that
 this band consists of many thousand lines.  Besides SOj lines, there are
 those of N2O and a few CO2 and CH4 lines.  Water is present through the
 continuum, which is composed of the tails of strong lines originating in
 the 2. 7 um and 6. 3 urn band systems.

        In order to gain a better overview about the distribution and strength
 of the SO2 band,  we have generated the band model parameters,  k and an
 averaged over 5 cm-1 intervals.  They are based on the statistical band
 model with exponential line strength distribution, viz. ,

                 r(S02) = exp[-ku/(l+ku/4a)'1/2J               (5)


 where u is the optical thickness (= CptL) and a = aope where pe is the equi-
 valent pressure for N2 broadened SO2.

        It is found that the SO2 band at 4 jum is quite weak.   It's band strength
 is only 22 cm^atm'1 at 300 K.  In comparison,  the f i - SO2 band at 8  7 Wm
 is about 4 times stronger, the CO fundamental band at 4. 6 urn is about 10 times
 stronger and the CO2 band at 4. 3 ym is over 100 times stronger.  Thus, the
 emission of the hot smokestack gas is relatively weak at low SO2 concentrations.

       By multiplying the emissivities  with the blackbody functions, the ap-
propriate radiances are obtained.  These calculations have been made for the
       We have also used our computer program to calculate the typical
transmission between a smokestack and an observer on the ground.  For the
conditions

                     Height  of smokestack: 54 m
                     Horizontal Distance: 122 m
                     Relative Humidity:   85%
                     Temperature:  16 C
                     Concentration of N2O: 0. 3 ppm
                     Concentration of CO2: 320 ppm
                     Concentration of CH4: 1. 4 ppm

the transmission becomes 0. 9979, which indicates an insignificant loss due
to atmospheric absorption. Even if the atmosphere were heavily polluted by
S02, the transmission is decreased only slightly.  As an example, 100 pnb
of SO2 reduces the transmission to only 0. 9977.

-------
 10
10"
                                Optical Thickness, ppm-m
 FIG.  2.  Theoretically computed radiances of hot N0-diluted SO0.
                                                   t>           £i
                                    8

-------
 THEORY OF THE GAS FILTER CORRELATION TECHNIQUE

        Since 1969, we have been developing the Gas Filter Correlation
 (GFC) technique^1-3) which is a modification of the Non-Dispersive Infra-
 Red (NDIR) technique  that has been known for some time(5X  In contrast
 to pure radiometry or dispersive spectroscopy, a GFC (non-dispersive)
 device uses the gas itself to obtain the ultimate high-spectral resolution
 filter (provided by the natural line-width of the gas).  High spectral re-
 solution is the  most important parameter in obtaining specificity and
 accuracy in pollutant analysis.

        Even quite narrow spectral bandpass radiometers are low-resolution
 instruments  and specificity is difficult to obtain.

       High-resolution instruments  depend upon finding a  single line of
 the pollutant to obtain  specificity,  but this necessitates the use of narrow
 apertures.  Thus, sensitivity is difficult to obtain.

       GFC combines the high energy throughput feature of radiometers
 and the high-resolution features of dispersive instruments.  It makes use
 of the contributions of all spectral lines of a band system of a particular
 species to obtain sensitivity.  Specificity is obtained by making use of
 random correlation between spectra arising from the particular species
 and interfering species; the principle of random correlation has been
 establishedU-o; for most pollutant species and interfering species  occurring
 naturally and in polluted atmosphere.  In addition,  a ratioing technique may
 be employed  that eliminates effects of changes in source intensity,  back-
 ground radiation, and  continuum absorption due to aerosols,  water vapor,
 or other molecular species.

       A remote GFC  sensor consists basically of  a single detector, a
 light chopper, a lens,  a gas cell containing  SOo and a reference cell  (see
 Figure 3).

       The chopper alternately passes the entering radiation through the
gas cell and the reference cell.  When the chopper  is in the position indicated
 in Figure 3, the signal generated at the detector is



        Vl  =   /{Er2ToT3 + NcT2TrT3 + €lN?T2roT3 + ins} RdX
               AX


 where T, N  and e refer to transmissivity,  blackbody function and emis-
 sivity respectively; the numerical subscripts are indicated in Figure 3
 and ins refers to the instrument which is maintained at a constant

-------
                                    Chopper
   Detector
To

Tr
*
f
*
e
Tt^t n«ii
\


•
EM \
vA;
Front Window
                                     2 C
FIG. 3.  Schematic diagram of remote gas filter correlation sensor.
                                10

-------
temperature; R is the overall responsivity of the detector, optical effi-
ciency and electronics.  All symbols are,  of course, a function of X ,
but this is not noted for the sake of brevity.

       Similarly, when the chopper blocks off the gas cell,  the signal
generated at the detector is
       2  =   /{EVoT3 + NcT2TrT3 + ClN?T2ToT3 + ins} RdX
       The signal generated at the frequency of the forward chopper is the
difference between V^ and Vj:
AV  =  /T2T3 tTl E + el Nr N
                                       c
                AX
       The instrument is balanced by replacing the incoming energy E by
two calibration blackbody sources of temperatures TBB  and TBB  and r
is adjusted such that AVT    =AV*T    •  This implies  fr « f0   where
                       iBB1     1BB2                      °
the bars denote mean values over the interval AX .  But, since the values
        anc* ^ TQO  are not electronically zero, in general, the value AV
                 662
given in Equation (8) is always referenced against level AV T-.-. or AV
Since R, T% ,  and 73 are only slowly varying functions of X , an overall
effective responsitivity Ro may be defined by
                          Ro =  T2 T3 R

where the bar denotes the mean value over the interval AX .  Thus,

            AV
                       l E(X) + C]L Nj(X) - N^(X)] [T - T 00] dX
                  AX

 where E(X) is given by Equation (1).  e.,N°(X) and N°(X) are effectively
 eliminated through the balancing procedure; thus,  c


                   g =  /E(X)[rr-ro(X)]dX
                     0   AX
                                11

-------
When E(X) given by Equation (1) is introduced into Equation (10) and the
integration over AX carried out, one obtains, using the mean value theorem
AV - 7  N^ (f -
  --   i   b  v a
                            a  a
                                 )  (f T -
                                 '  v    r
                              r -7ft)
                               r   o
                                                                  (11)
       Since by balancing the instrument,  irosrrt Equation (11) may be
simplified:
  AV _ - r-
  R-     :
                                                                  (12)
       Thus, an expression results that shows the ac signal is effectively
a product of a modulation function that is only related to the SO2 transmissivity
and fixed instrument transmissivities and of the difference between the ra-
diance emitted by the plume and background atmosphere.

       If we now consider a second cell pair with TQ" t r"p and chopped at a
different frequency,  but using the same detector and optical components,  a
similar expression is derived:
f = VaNp-
-------
 temperature if the gas cell and plume temperatures are the same.  Any
 mis-match in these temperatures will cause only relatively small errors
 in determining SO2 concentration due to the second order effects  of tem-
 perature on the absorption coefficient of SO2 and changes in spectral slope
 of the blackbody function with temperature.

       The signal as a function of SO2 concentration in the plume can be
 adjusted depending upon the amounts of SO2 in the specifying cells.  In
 addition, we  have observed that the addition of -a small amount of SC>2 to
 the reference cell will greatly enhance the sensitivity. In this case the
 signal is given by

                          TTn9 - FT  f0
                    c _      v* _ r  2
where fj and f2 are the dimensionless aperture attenuators used to balance
or zero the sensor; viz.
                       = frf2  and
       The optimum values for rJJ, r~02, and r  were studied in detail.
These results are presented in the Appendix.
                                 13

-------
                              SECTION 3

                SENSOR DESIGN AND PERFORMANCE
 MECHANICAL-OPTICAL DESIGN

        The two-cell GFC pairs are contained in a single optical system as
 illustrated in Figure 4.  Separate tuning fork choppers are used for each cell
 pair.  The cells are 10 cm in length.   The image of the stack plume is
 focused by an f/3 lens on a 1. 0 x 1. 0  mm detector which defines the field-
 of-view to be about 8 milliradians (8 m at 1 km). An ambient temperature
 operation (ATO) PbSe detector is used; but, a single-stage thermoelectric
 cooler  is used to provide temperature control.   A passband optical filter is
 located in front of the detector.  This filter is centered at 4. 00 microns
 with a half-transmission width of about 0.1 microns.

        Sapphire optics, antireflection-coated, are used.  A transmission
 of greater than 0. 9 is obtained.  The objective lens has a diameter of 5 cm
 and a focal length of 15 cm.

        Section A-A of Figure 4 shows the configuration of the dual split cell.
 The two gas cells contain different partial pressures of SO2 and are pres-
 surized to 1 atm with pure N£ to pressure-broaden the SO2 lines.   The re-
 ference cell contain a lesser amount of 803 pressurized to one atm with pure
 N2.

        The tuning fork choppers,  made by American Time Products  (Bulova)
 are stable, reliable and low power.  Because the frequency is dependent only
 on the mechanical resonance of the fork, no stable-frequency AC power source
 is needed and battery operation is possible,  as is true of the entire electronics
 system.

        Section B-B of Figure 4 shows the dual tuning fork configuration.
 Fork-1  is shown closed,  allowing radiation to pass through the AV-1 reference
 cells; Fork-2 is shown open,  allowing radiation to pass through the AV-2 gas
 cell.  Fork-1 operates a frequency of 40 Hz and  Fork-2 at a frequency of
 100 Hz.  Because a single lens serves both AV-1 and AV-2 systems and super-
 imposes the image of each on the  same detector, both systems have exactly
 the same field-of-view at the stack plume, as is essential for proper can-
 cellation of the stack effluent temperature factor. The superimposed image
 signals  of the two AV systems are electronically separated by signal processing
described later.

       A third tuning fork chopper,  operating at 800 Hz,  is located immediately
ahead of the detector aperture.  This chopper obstructs the entire beam when
                                 14

-------
   Entrance Window

   B    » A
r
            Dual Split Cell   -
U\   L
       Windows

       Objective Lens
       Plane of AV Tuning
         Fork Choppers
  AV-1
                       Section B-B
                                     AV-2
                                                                             T. E.  Cooler
                                                                           PbSe Detector
4. 0 nm Filter
800 Hz Tuning Fork
Rel

Gas-1

Ref


•f


Rel

Gas-2

Ref
                                                                  Section A-A
                          FIG.  4.  Optical system schematic.

-------
 closed.  The purpose of this chopper is to eliminate the low frequency 1/f
 detector noise from the AV signals.

        Because of the small (8 mr) field-of-view, accurate aiming is
 essential.   A 4-power Bushnell riflescope,  accurately aligned with the
 GFC optics, is an integral part of the instrument.  An illuminated retical
 in the riflescope permits accurate  aiming at dusk or under other adverse
 lighting conditions.

 ELECTRONICS DESIGN

        An electronic signal processing technique that involves phasing,
 gating (sample-hold) and ratioing of different frequency signals has been
 developed.  A block diagram of the electronics is shown in Figure 5.

       In this instrument,  two AV signals are used resulting from the use
 of two gas/reference cells systems containing different partial pressures
 of SOo.  Each cell pair is chopped by a tuning fork chopper of different
 frequency.  Because of the common objective lens the image at the detector
 consists of both AV signals at their respective chopping frequencies.  These
 are later separated by electronic filtering.  The frequencies are 40 Hz and
 100 Hz.  These are not harmonically related and are far enough apart to
 allow good separation by filtering, yet are low compared with the 800 Hz
 "carrier" system created by the third chopper, described below.  This
 tuning fork chopper, operating at 800 Hz,  is located immediately in front
 of the aperture plate and obstructs the entire light beam when closed.  The
 detector signal is therefore the 800 Hz "carrier" signal amplitude modulated
 by the two AV chopper frequencies.

       The detector is followed by a preamp and a 800 Hz bandpass filter.
 The filter passes the 800 Hz signal  and its modulation side-bands at 700,
 760,  840 and 900  Hz while rejecting the DC  component of the detector output
 which is due to the bias current.  Thus, even though low frequency AV
 chopping is needed in order to obtain large-amplitude fork oscillations,
 only the detector  noise at the 800 Hz passband is processed,  and the large
 1/f detector noise at the low AV chopping frequencies is rejected and high
 effective detector D*'s are obtained.

       The filter output feeds a lock-in amplifier.  Receiving its reference
signal from  the 800 Hz fork driving circuit,  the lock-in synchronously detects
the 800 Hz wavetrain and provides a DC output on which are superimposed
the two AV signals.
                                 16

-------
Bias

 Preamp
0
PbSe
Detector
  Thermal
   Control
800 Hz


 40 Hz


 100 Hz
            800 Hz
       Tuning
       Fork
       Drivers
                  Filter
                               Lock-In
                                Amp
                   Phase &
                   Gating
                   Control
                                          40 Hz
                                          Filter
                                          100 Hz
                                          Filter
                                                     Lock-In
                                                      Amp
                                                   Lock-In
                                                    Amp
                                                             Low-Pass
                                                               Filter

                                                                 Divider
                                                             Low-Pass
                                                              Filter
                                                                                           JL
                                                                                           Visua
                                                                                         Read-Out
                                                                               AV2AV2AV1
                                                                                 Recorder
                                                                                 Output
                           FIG. 5.  Signal processing block diagram.

-------
THEORETICAL PERFORMANCE

       The signal-to-noise ratio for a GFC radiometer is given by


                        nr A O D*AE AX
                SNR =     ° °°	                         (17)
where

       77      is the instrument efficiency

       •F      is the transmissivity through the SO2 gas cell

       A      is the entrance aperture of the gas cell (cm )

       O      is the instrument acceptance solid angle (sr)

       D*     is the detector detectivity (cm-Hz1'2/W)
                                    2
       A,     is the detector area (cm )

       Af     is the electronic bandpass (Hz)

       AE    is the difference in radiance between the instrument
                                        2
              gas and vacuum cells (W/cm -y-sr)

       AX     is the bandpass of the optical filter

       A figure of merit for any radiometric instrument is the noise-
equivalent-spectral-radiance which is given by Equation (17),  setting
SNR = 1; thus,
                   NEN =  — - — - - ,  W/cm -sr             (18)
       A summary of the instrument parameters is given in Table 1.

       Calculations of the radiance integrated over the actual filter band-
pass were made.  The results are presented in Figure 2.  Operating as a
pure radiometer (channel 2), the difference between these radiances and
the radiance of the forward  chopper is indicative of the signal to be detected.
For example, if the chopper is emitting as a 300 K blackbody,  the radiance
difference is plotted in Figure 6.  Since the theoretical NEN for Channels 1
and 2 are 1. 22 x 10-7 and 8. 14 x 10-8W/cm2sr, respectively, the theo-
retical radiometric sensitivities are on the order of 150 ppm-m and 100 ppm-m
for a temperature of 350 K.
                                 18

-------
                       TABLE 1



       SUMMARY OF INSTRUMENT PARAMETERS
Parameter
Tl jfc
T **
O
f, Hz
A 2
AQ, cm
n , sr

D*, cm Hz /W
2
A,, cm
Af, Hz
2 ***
NEN,W/cm -sr
Channel 1
0.37
0.44

100
1.54
6. 4 x 10"5
q
2. 7 x 10y
io-2
2.8x 10"3 to 0.25
1.221x 10"7
Channel 2
0.37
0.66

40
1.54
6.4x 10"5
q
2. 7 x 10y
io-2
2. 8 x 10"3 to 0. 25
8. 143 x 10"8
    77    Twindow x Tgas cell window x Tlens x Tfilter x ^electronics



       = . 98 x (. 90)2 x . 95 x  . 5 x . 98  =  0. 37



 ** See Appendix.


                                                          -3
*** Assumes integration time = 90 seconds; Af = 1/4 t = 2. 8 x 10  Hz.
                          19

-------
                                                      330-81
CO
o
                                |*r,(S02', T S) - Nu(300 K)
                -ILL
                  10
                                                       SO, Optical Thickness, ppm-m
                  FIG. 6.  Radiance difference detected by sensor (assumes chopper's radiance is a 300 K blackbody).

-------
 LABORATORY TEST RESULTS AND CALIBRATION

        Testing of the sensor was performed to determine the optimum
 specifying and reference cell SO2 concentration.   The optimum concen-
 trations and corresponding transmissivities are:

              uQ1 = 5. 0 atm-cm         TQ- = 0. 44

              uQ2 = 2. 0 atm-cm         f™ - 0. 66

              u  = 0. 5 atm-cm          7  = 0. 88
               r                        r

        Measurements were made of the signal and noise of the sensor,
 while operating in the radiometric mode and in the GFC mode.

        A Barnes Model 1140T Field Source was used to calibrate channel 2
 (V2R) operating as a pure radiometer.  The results,  covering a temperature
 range from 50 to 225 C,  are presented in Figure 7.  Note, these results are
 for the  sensor normalized to—- Gain = 10 and Attenuation = 10.

        In order to compare the experimental data with theoretical pre-
 dictions,  we have to determine the instrument responsivity.  The radio-
 inetric  responsivity RQ is defined through
              V2R = R02   C(w)[N(w, T) - N(W, TQ)] dco           (19)
                          Aw

where V is the radiometer signal in volts,  Ro is the responsivity in V/
Wcm-^sr'1, C(w) is the normalized filter  function, and N°(w, T) and
N°(w, TC) are the blackbody functions of the calibration source and chopper
blade, respectively.

       From the measurements  of V2R, the radiometric responsivity was
determined —

                  RQ2 = 7. Ox 103V/Wcm"2sr"1                  (20)

       The measured peak-to-peak noise level with an integration time of
90 seconds is 5 x 10~6 V.  This is  equivalent to 5 x 10~3/5 (=  10-3 Vrms),
assuming a factor  of 5 to convert peak-to-peak random noise to rms noise;
i. e. ,  99 percent of the noise energy exists within a voltage range of 5 x
Vrms-  T1\f theoretically predicted noise,  given by the NEN in Table 1, is
8. 14 x 10-° x 7. 0 x 10J = 5. 7 x 10'4 Vrms.  Thus, the sensor's noise limit
is about two times higher than predicted.  The actual rms noise measure-
ments converted into radiance is 7. 1 x 10~8W/cm2-sr.   This  defines the
minimum detectable SO2 optical thickness.
                                21

-------



         FR/+FH
O.Ot-
        50
     FIG.  7.  Radiometric mode calibration for Vor, channel.
                                              ZJtx
                           22

-------
       Measurements were also made of N2-diluted SC>2 mixtures heated
to 170 and 270 C in a 50 cm calibration cell with 7. 5 cm dia. sapphire
windows (transmissivity = 0. 88).  The experimental data corrected for
window  emission are plotted in Figure 8; also shown are the theoretically
predicted values obtained by multiplying Ro(= 7 x 10^ V/Wcm'^-sr"*)
with the radiance differences given in Figure 6.   Excellent agreement is
obtained.

       Thus, by using the curves presented in Figure 2, the radiance of
the choppers, and the empirical  value of Ro  (Equation (20)), the SC>2 optical
thickness may be determined from field radiometric measurements, if the
plume temperature is known, or vice-versa.

       Operating in the GFC mode,  the sensor is first balanced by adjusting
the reference cell apertures such that the same  AV signal levels are obtained
when viewing the field source at temperatures of 350 and 500 K.  Note,  this
is not necessarily electronic zero.  After balancing, measurements were
made of N2-diluted SC>2 mixtures heated to 170 and 270  C in the 50 cm cali-
bration  cell.

       Preliminary measurements were made with a small (0. 25 x 0. 25 mm)
detector which provides a sensor fov of 2 mrads.  The initial measurements
were made with pure N2 in the reference cells and with UQI » 4. 3 atm-cm
and UQ2  w 1. 3 atm-cm.  Tests were conducted by using  the laboratory atmo-
sphere as the background and by placing a dry ice block behind the cell
simulating a cold sky background; no discernible differences in the data
were observed in the two test procedures.

       Experimental data of AV^ and AV2 were obtained at  temperatures of
270 C and 170 C and are reproduced in Figures  9 and 10.  The abscissa is
the optical thickness of SO2 in ppm-m.   In dividing AV2 by  AVj, the ratio
which is independent of the gas temperature, is  obtained.  The results are
shown in Figure 12.  As expected, the results are independent of the tem-
perature.  These experiments demonstrate the viability of  the ratio technique
for the measurement of hot SO2 emission without knowing the gas temperature.
The resulting calibration data (Figure 11) were  used to  interpret field mea-
surements at two oil-burning plants,  as described in the following section.

       Latter field measurements and laboratory experiments showed that
the AV2/AVj ratio could be made more sensitive by optimizing the gas-
reference cells' SO2. optical thicknesses. Using near optimum values,  a
calibration curve was obtained (see Figure 12) that was applicable to field
measurements at a coal-burning power plant.
                                23

-------
N .
si

ii
I

                                             ^ rheoretical Data '••
       FIG. 8.  Radiometric data compared with theoretical values.
                                24

-------
                     10       20       SO
                     SO Optical Thickness, ppm-m x 10"
50
FIG.  9.  Data obtained with sensor operating in GFC mode (T = 270 C).
                                 25

-------
    0.20
    0.16
CO
?   0.12
•a
c
>   0.08
    0.04
      0  *
                       Gain =  10
                       Att. =  10
                   10
20
30
40
                     SO2 Optical Thickness, ppm-rm x 10
                                                      -3
                                     AV
                                                                   AV.
                                                              50
 FIG. 10.  Data obtained with sensor operating in GFC mode (T = 170 C).
                                 26

-------
                                      O   T  = 270 C
                                      A   T  = 170 C
FIG. 11.  Data showing sensitivity of ratio to SCL temperature and concentration.
                                 27

-------
                                      K-S
•CMI-COOAHtTHMte  40 4973

         CO.
                                                                                        r    r
                                                                                                       . S
CO
CO
                                                           ; ; : :    ;-p	rr
                                                            -;  ;ii!ii i>i>  il;
                                                             10*
                                                        U (SO2), ppm-m
                                     FIG. 12.  RGFC ratio mode calibration.

-------
       From the measurements of signal and rms noise of the sensor, its
limiting sensitivities as a function of plume temperature have been estab-
lished.  The results are given in Figure 13.  These results are based on
an integration time equal to 90 seconds and, for the radiometric mode,
assume the chopper radiates as a blackbody at a temperature of 300 K; they
are based upon the condition when the signal-to-rms noise ratio equals one.

FIELD TESTING

       Preliminary field measurements were made using the 0. 25 x 0. 25 mm
detector and non-optimized gas cell  parameters at two oil-burning power
plants near San Diego,  CA. A photograph of the first site is shown in
Figure 14. A summary of these measurements is presented in Table 2.
Extractive measurements analyzed using EPA Method 6 were made during
all tests except the first one.

                             TABLE 2

        SUMMARY OF REMOTE SO2 FIELD MEASUREMENTS
          MADE AT TWO OIL-BURNING POWER PLANTS
Date
3/25/74
5/20/74
5/21/74
5/21/74
5/23/74
Range
130 m
185 m
185 m
130 m
130 m
Stack
Dia.
4.25 m
4.0m
4.0m
4.0m
4.0m
Fuel
oil
oil
oil
oil
oil
Temp.
650 K
410 K
420 K
420 K
440 K
SO2 Cone.
Range
~ 200 ppm
350-450 ppm
260-340 ppm
170-230 ppm
260-340 ppm
Weather Conditions, Time
20 C, calm, partial clouds, p. m.
21 C, light wind, clear, p.m.
26 C, light wind, partial clouds, p. m.
26 C, light wind, partial clouds, p. m.
22 C, calm, clear, dusk
The data taken on 3/23/74 are consistent with a nearly steady SO2 concen-
tration in the stack of about 200 ppm, which was a calculated value based
upon the known sulfur content of the fuel oil'''. (No extractive data were
available during these first tests.) However,  the accuracy at these low SO2
concentrations is low (see calibration curve in Figure 11).  Radiometric
mode data were also taken and reduced, using laboratory calibration curves
and the measured in-stack temperature; the results indicate a concentration
range from 150 to 210 ppm SO«.

       The data taken on 5/20, 5/21 and 5/23 were obtained at the second
site.  In these cases the plume temperatures were lower,  as seen by the
data in Table 2, and the signal levels were correspondingly low,  giving
the larger indicated uncertainties  in the data.   Extractive data were taken
                                29

-------
                         K-C SEW-IOOABITHMIC « CYCLES X 70 DIVISIONS
                         C. (turret, t CMC* eo. ..«..»>.
                        46 6013
650
600
300
                             10
        102"

SO2 Optical Thickness, ppm-m
            F'G. 13,   Limiting sensitivity of the sensor for t = 90 seconds.

-------
FIG. 14.  Photograph of SCX, remote sensor during initial field testing.
                                  31

-------
  and analyzed following EPA Method 6; these data indicated SO2 concen-
  trations ranging from 30 to 50 ppm.   However, the data were not believed
  reliable because of a leak in the sampling apparatus.  Furthermore, the
  power plant company assumed 200 to 300 ppm were being emitted^,
  which is consistent with the observed results.

         Field measurements were also made using the 1. 0 x 1. 0 mm detector
  and nearly-optimized gas cell parameters at  a coal-burning power plant near
  Charlotte,  N.  C.  A summary of the measurements is presented in Table 3.

                             TABLE 3

         SUMMARY OF REMOTE  SO2 FIELD MEASUREMENTS
             MADE AT A COAL-BURNING POWER PLANT
Date
••' ' •
6/25/74
6/26/74
6/27/74
6/27/74
6/27/74
—
Range
^MMM.U_
210
170
170
4CO
210

m
m
m
m
m
Stack
Dia.

2.
2.
2.
2.
2.

0
0
0
0
0

m
m
m
m
m
Fuel

coal
coal
coal
coal
coal
Temp.

425 K
425 K
425 K
425 K
425 K
SO2 Cone.
Range
700-1300 ppm
700-1300 ppm
500-700 ppm
500-700 ppm
500-700 ppm
Weather Conditions, Time'

32 C,
30 C,
24 C,
24 C,
24 C,

light wind, scattered
light wind, cloudy, a.
calm, foggy, a. m.
calm, foggy, a. m.
calm, rain, p.m.

— — — _
clouds, p.
ro. and p.



       Both radiometric and GFC mode of operation data were taken.  In
 addition, extractive data were taken and analyzed following EPA Method 6
 and continuously monitored by a DuPont Model 460-1 Analyzer (8).  All of
 the data are presented graphically in Figure 15.  As seen, all of the data
 are in agreement within + 100 ppm.

 DISCUSSION OF FIELD  TEST DATA REDUCTION

       The data presented in Figure 15 were reduced by measuring the
 individual AVj and AV2 signals because the signal levels were too low for
 the electronic divider module to function properly.   In addition, the sensor
was not perfectly balanced during the measurements.  This means the
terms (c. f. Equation (11))
         N°  f  (r -T )
          p   a v  r   o'
and
are not zero.  Of these two terms the first term is much larger than the
second term, and,  the data were corrected by subtracting the contribution
                               32

-------
  1200
  1100
  1000
  900
s
  800
  700
  600
  500
             6/25/74
    o
•
 0°
             •
             o
      13   14   15   16   17

              HOUR
                                           6/26/74
                9   10   11   12   13   14  15

                           HOUR
                                                      6/27/74


                                                 O  DuPONT

                                                 A  METHOD 6

                                                 O  REMOTE CfC
                                                                               o
9   10   11   12   13

     HOUR
                                                                                   .SO
                                                                                   .55
                                                                                   .60
                                                                                      .W2
                                                                                   .65
                                                                                   .70
                                                                                   .75
    FIG.  15.   Comparison of remote sensing data with extractive data obtained
    from the  DuPont Analyzer and from EPA Method 6.
                                         33

-------
 due to the first term.  This contribution was measured in the laboratory
 using a blackbody field source both before and after the field measurements.

        The majority of the measurements were averaged over about a 15
 minute time period.  In cases where the two AV signals were changing rapidly
 with time, the analyzed data was erratic and, thus, not used.

        In the GFC mode,  two  principle sources of error arise. The first is
 due to the basic sensor sensitivity (signal-to-noise ratio) and the second due
 to the temperature dependent correction applied because of sensor imbalance,
 as noted above.

       In the first case,  the SNR averaged about 13 and 25 for  channels AV2 and
 AVi, respectively.  Thus,  for u = 1600 ppm-m, AV^/AVj = 0. 55, and in terms
 of error,  this is equivalent to


                         AV2  _  0. 55(1 + 1/13)
                                  1.0(1 ±1/25)
                            A
 which gives


                         AV2
                        -g^-   =   0.55 ±0.095                     (22)


 Or,  u = 1600 ± 350 ppm-m and,  at  Location 2,  this gives an uncertainty of
 ± 220 ppm in SOg concentration.

       In the second case, it is believed that the plume temperature was
 known within + 25 C.  Note,  in the  cases where radiometric data were taken,
 it is believed to be known within ± 10 C.  For temperature uncertainties of
 + 10 and +25 C, the equivalent uncertainties in 563 concentration are about
 + 70 and + 175 ppm,  respectively (on the average).

       In summary,  since both  of these sources of uncertainties may be
 considered to be random, the maximum RSS uncertainty is about + 280 ppm.

 CONCLUSIONS

       The GFC dual channel technique has been proven that it can remotely
determine SO2 concentrations in hot plumes with only minimal effects due
to plume temperature and backgrounds.  However,  the prototype sensor
was found to have certain limitations.  These are:
                                34

-------
        1.     Limited sensitivity due to detector noise.

        2.     A susceptibility to becoming unbalanced, which
              necessitates rather large temperature corrections.

        3.     Lack of provision for in-field rebalance adjustment
              capability.

        4.     Subject to erratic behavior due to overheating when
              operating in direct sunlight on very warm days.

        5.     The electronics divider module used does not permit
              electronic ratioing of the AV signals because of low
              signal levels.

        6.     A fov (8 mrads) that is larger than desired; 2 mrads
              would be better.

       The deficiencies above were corrected by a modified design; this
was done under EPA Contract 68-02-1696.
                                 35

-------
                            SECTION 4

                          REFERENCES


 1.     NASA Contracts 12-2109, 1-10466,  1-11111, and 1-12048.

 2.     Ludwig, C. B. et al.,  "Remote Measurement of Air Pollution
       by Nondispersive Optical Correlation", AIAA Paper No.  71-1107,
       November 1971; AIAA Jr. 11, 899,  1973.

 3.     Bartle, E. R.  et al., "An In-Situ Monitor for HC1 and HF", AIAA
       Paper No. 71-1049, November 1971; J. Spacecraft and Rockets,
       11,  836, 1972.

 4.     Chandrasekhar, S., Radiative Transfer,  Dover Publications, Inc.
       New York (1950).

 5.     Considine, D.  M.,  Editor., Process Instruments and Controls
       Handbook,  McGraw-Hill, New York (1957).

 6.     "Determination of Sulfur Dioxide Emissions from Stationary Sources",
       Federal Register 36, 24890, December 1971.

 7.     Hardway, J.,  SDG&E (private communication),  March 1974.

8.     The data obtained by the DuPont Model 460-1 Analyzer were provided
       by Dr. W.  Herget of EPA.
                                36

-------
                            APPENDIX

            OPTIMIZATION OF GAS CELL PARAMETERS

 THEORETICAL

       A comparison of the two GFC calibration curves (Figures 11 and 12)
 shows a  large increase in sensitivity to changes in SO2 optical thickness
 when SO2 is added to the reference cell.  The signal-to-noise is actually
 decreased, but the sensitivity obtained when chopping between two cell
 pairs containing different optical thicknesses of SO2 and ratioing is much
 greater than when chopping between two cell pairs, one of each containing
 different optical thicknesses of SO2 and reference cells containing N«.

       For the latter case, the ratio is given by Equation (15); viz.
                                rrn9 -   r  £„
                        _  .  _02  _r2

                           1    TTorTTr£l

Equation (A-l) reduces to Equation (14) when

                 V2  = f02   and   Vl  =  ?01 •

       The mean value  for transmissivity is given by
" dX =  ^  f±  .UfL [k(X)f u-
       -         e      dX  =             .       [k(X)  uOX     (A-2)
             AX                   AX  n=0

where u(cm-atm) is the optical thickness and k(X)(cm"1-atm"1) is the
spectrally dependent absorption coefficient.

       Since u is independent of wavelength,
                      =
                                n!
                         n=0
Using Equations (A-l) and (A-3) for the case where rr = 1 (ur = 0)
                                 37

-------
      = [l-k(u
 AV  =  l-k(u+u  ) +
                 O    ft      \j      u      O         I


                                                                     (A-4)
                                               IF
          - ku +   . u2 -   . u3
                                                   c\     f\



 Simplifying, and considering only second order terms in u,  gives



         AV  =  u[uo(7-k2)- n02
-------
        Considering now the case where there is gas in the reference cell
 of optical thickness, Uj., and transmissivity, rr, similar expressions can
 be derived.  For this case, the modulation function, AV', is given by


                     AV'  =  TT -TT(?o/f)                      (A-lO)


        Performing a similar series expansion to the second order in u gives

       [•^w ""I                   f*
        -KuI      /         s1  c>      o   o   1    ~o   5"      }
            r       I/      \ /, £• T \  i  £>   £\tl\f:-,4 . O»     I
       e    J  = u|(uo-ur)(k  -k )-(uQ -ur )(2-)(kk -k ) + .. .j

                                                                   (A-ll)
       Again, considering the second cell pair, and forming the ratio between
the two signals results in


               _  'Uo2 - V
        Differentiating Equation (A-12) with respect to u gives


             Flu)  =  (U°2 " "r )       (Uol " U02}
               W     (u  -u \  ' ~ - 7
                      %1  Ur}   [A/B - (uol + ur + u)]2

        Dividing Equation (A-13) by Equation (A -9) shows the relative effect
 in terms of sensitivity for the two cases; i. e. ,

               .  (U02 - V .    .
               -
       Equation (A- 14) can be examined to show the increase in sensitivity
to u in terms of the various cell optical thicknesses.  Also,  the fundamental
equations,  (A-7) (A-8),  (A-ll) and (A-12) can be examined to show the effect
on signal levels.

       The constant A/B was evaluated using the data presented in Figure  12.
The results are as follows.
                                 39

-------
            A/B (cm-atm)
               8.6
              14.6

               9.1
                                  Evaluated at
                             F(0) = 0. 95

                             F(u = 2000 ppm-m) = 0.47

                             F'(0) = -2. 5 cm"1-atm"1
        This evaluation indicates A/B is not a constant and that the sim-
 plified analysis inadequately describes the true performance of the sensor.
 Nevertheless,  calculations were made,  assuming A/B = 12 cm-atm.  The
 results are presented in Figure A-l.  These curves indicate the dramatic
 improvement in sensitivity as the reference cell's optical thickness is in-
 creased.  This can be seen by examination of Equation (A-14) since when
                 U01 + u + ur = A/B
                                                                 (A-15)
the ratio of the sensitivity with u * 0 and sensitivity with u_ = 0 (= F'(u)/f'(u))
becomes infinite.

       However,  as was pointed out, this simplified analysis does not
adequately describe the sensor's performance.  This is principally due
to not considering the higher order terms in the series expansion.

       A similar analysis has been performed to include  third order terms.
The resulting equation is

  s Uo2-ur   1+(u02+Vu)w+uo2urX + ^of^oZ"^^*  (uo2+ur)uZ
Uol"Ur   MuQl-Hir+u)W+ uQlurX +  (uQ1 +uolur-Hi
                            2  2;
                                                             (uol+ur)uZ
       The constants W,  X,  Y and Z have been evaluated by solving a 4 x 4
matrix using determinants developed from fitting the results presented in
Figure 12.   Their definitions and numerical values are given below.
            W =

            y  _


            Y  =


            Z  =
  —JT  —K  —K-   2
(kV -kd)/(k^-k )  = -0.3118

         °  ~5   J
          )/(k^ - k )  = -0. 0260


               k )  =  0.0343

              k ) = 0.0435
                                                                  (A-17)
                                                                         (A-l
                                 40

-------
too
                                              F'(l. 8) = oo
                      I  __-L
                 000 ppm-m
                          ., cm-atm
   FIG. A-l.  Effect of specifying and reference cell optical thickness

   on sensor sensitivity.
                              41

-------
 Using these values, the calibration data presented in Figure 12 is des-
 cribed within i 2 percent over the range of 100 to 5000 ppm-m of SOp.

        Calculations of AV for u  = 1, 2, 3,... 10 atm-cm and u = 0,  0. 5,
 1. 0, 1. 5,  and 2. 0 atm-cm have been made.  These results show how the
 addition of SC>2 to the reference cell affects the slope, but also reduces the
 signal  intensity. It is calculated that at large specifying cell optical thick-
 nesses, the signal levels are only reduced by about a factor of 2, even for
 large (2 atm-cm) amounts of SO£ in the reference cell.  It should also be
 noted that the results for intermediate  specifying cell optical thicknesses
 (4 and  6 atm-cm) and reference cell optical thicknesses greater than  1 atm-cm,
 do not  appear to be physically reasonable.  This indicates that the use of the
 analytical expression is not quantitative enough.  Apparently even higher order
 terms  in the series approximation should be included.  Nevertheless  approxi-
 mate optimum cell conditions have been calculated.

        The expression for F(u) = AV2/AVJ has been differentiated and
 numerically evaluated for u = 1000 ppm-m.  These results show the rapid
 increase in sensitivity as the reference cell's optical thickness is increased
 and the optimum value for the second specifying cell's optical thickness.
 Furthermore,  they indicate that the sensitivity,  F'(U), continues to increase
 as un-i  increases.
        Since the length of both specifying cells in the sensor is 10 cm and
 it is undesirable to pressurize the SOo beyond one atmosphere due to pressure
 broadening the 803 lines, the practical maximum optical thickness for UQI
 is 10 atm-cm.   For this condition,  the calculations indicate the optimum value
 for the second  specifying cell's optical thickness is about 4. 5 atm-cm.

        We have also calculated the values for the reference cell's optical
 thickness that forces F'(u) - » for u = 1000 ppm-m;  the result is u  =
 1. 8 atm-cm.

        In summary, calculations made using the third order series expansion
 analysis indicate the sensor would have optimum performance with

                            UQ- = 10 atm-cm

                            U02 = ^'
                            u   =  1. 8 atm-cm .
                             r
EXPERIMENTAL
       Measurements have been made using the sensor to determine SO2
transmissivity as a function of optical thickness.  The sensor was operated
in the radiometric mode by blocking off the reference cell apertures.  The
                                 42

-------
lower specifying cell was evacuated, and the signal (I ) recorded; then,
pre-mixed SOo diluted with N2 was admitted to the ceft at one atmosphere
pressure and the signal (I) recorded.  The ratio I/IO is,  of course, the
transmissivity.  The results of this experiment was presented in Figure
A-2.  Also shown are the results generated by a line-by-line computer
program using the actual filter function of the sensor.

       Laboratory calibrations of the sensor operating in the GFC mode
are difficult.  We used a 50 cm long cell with sapphire windows on both
ends providing a clear aperture of 6. 2 cm. The windows are uncoated
and have a transmissivity of 0. 88.   The maximum entrance aperture
dimension is (0. 775 + 0. 867) (2. 54)  = 4.17 cm.  Locating the sensor
100 cm away from the far end  window of the cell gives a projected maxi-
mum dimension of 4.17 + (. 008) (100) = 4. 97 cm and, thus,  no direct con-
tribution is given by the cell walls.  However,  three major  difficulties
arise in the calibration.

       One.  Multiple reflections internal to the cell and near field-of-view
thermal non-uniformities generally  give unreliable AV signals.

       Two.   The cell is heated by  external strip heaters, cold pre-mixed
SO2-N2 test gas mixtures are  admitted to the cell and the temperature
monitored by a thermocouple in the  cell's interior.  It is difficult to main-
tain the same temperature as different SO2-N2 concentrations are sequentially
admitted to the cell and maintained  at one atm pressure.  Since the tempera-
ture strongly affects the magnitude  of the AV signals and relatively long time
constants (30 seconds) are required to get adequate signal-to-noise, the data
tend to be erratic.

       Three.  Field data give AV signals that are larger than those simu-
lated in the laboratory and indicate  a greater sensitivity to changes in SC>2
optical thickness.

       The most desirable methods for calibrating the sensor appears to
be the use of field data or to use an  artificial stack.  For SC>2 optical thick-
nesses greater than 5000 ppm-m, laboratory measurements generally give
repeatable results if the temperatures and concentrations of test gas mix-
tures are carefully controlled.
     *
       In attempting to simulate the theoretically predicted optimum gas
cell parameters,  it was discovered  that for UQJ = 10 and ur = 1. 8 cm-atm
of SO2,  respectively,  the throughput (and signal) of the sensor was reduced
by 70 percent (see Figure A-2) and  the aperture adjustment did not have
enough travel to balance the sensor.
                                 43

-------
N .'
E-
5!
s      ~~T
       FIG.  A-2.  Theoretical calculation and measurement of N« -diluted
           pressurized to one atm for SO« sensor.
                                  44

-------
       Within the limits of the sensor's balancing adjustment and the rc-
pcatablc signals generated using the calibration cell,  the apparent optimum
cell parameters were determined and are summarized below.

                                                      Ref erence
   Parameter        Upper Cell, AV2  Lower Cell, AVj
 Chopping frequency       40 Hz         100 Hz       40 and 100 Hz
 Length                   10 cm          10 cm        23 cm

 SO2 Concentration        17. 2 %         47. 2 %        2. 2 %

 Optical thickness      2. 0 atm-cm    5. 0 atm-cm    0. 5 atm-cm

 Transmissivity           0. 66            0. 44          0. 88
       Note,  the sensor's construction is such that the two 10 cm long
specifying cells and the 23 cm long reference cell actually give effective
u0's that are  the sum of the SC«2 pressure in the specifying cell times 10 cm
plus the SO2 pressure in the reference cell times 13.
                                 45

-------
TECHNICAL REPORT DATA
(/'lease read /HSl/uctions on the reverse he jure completing)
1. RtPOHT NO. 2.
EPA-650-2-75-041
4. TITLE ANDSUBTITLE
Infrared Sensor for Remote Monitoring c
/. AUTMOH(S)
E. R. Bartle and E. A. Meckstroth
9. PERFORMING ORGANIZATION NAME AND ADDRESS
JRB Associates, Division of SAI
1200 Prospect Street
P. O. Box 2351
La Jolla, CA 92037
3. RECIPIENT'S ACCESSIOI*NO.
5. REPORT DATE
May 1975
'f SOn 6. PERFORMING ORGANIZATION CODE
fu
8. PERFORMING ORGANIZATION HEPOFU NO.
10. PROGRAM ELEMENT NO.
1AA010
11. CONTRACT/GRANT NO.
68-02-1208
12. SPONSORING AGENCY NAME AND ADDRESS 13. TYPE OF REPORT AND PERIOD COVERED
National Environmental Research Center Final Report
Office of Research and Development
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A prototype passive infrared sensor for the measurement of sulfur dioxide
emissions from stationary sources is described. The infrared radiation
emitted by gases in a plume originating from smokestacks may be detected,
and from this the 803 concentration in the plume may be determined. In
general, the radiation received by the sensor is a function of the intervening
and background atmosphere. Thus, the problem of quantitative measurements
Is generally complex. A technique is described, based upon the principle of
Gas Filter Correlation, which minimizes these effects.
This report presents a detailed description of the sensor, it's specifications,
and performance characteristics. The basic unit is battery operated and
weighs only 10 kgms; thus, it is readily portable. It's sensitivity is presently
limited to about 70 ppm-m for source plume temperatures of 270 C and about
290 ppm-m for temperatures of 170 C, but this can be improved.
The results of field testing at both oil and coal-burning power plants are com-
pared with extractive sample data. In general, the remote measurements
agree with the extractive data within + 25 percent over SO% concentrations
ranging from 150 ppm to 1300 ppm from slant ranges of 130 to 400 m.
This report is submitted in partial fulfillment of contract number 68-02-1208
by JRB Associates, a division of Science Applications, Inc. under the spon-
sorship of the Environmental Protection Agency.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
1. Remote Sensing
2. SCvj Measurements
3. Gas Filter Correlation
13. DISTRIBUTION STATEMENT
Release Unlimited
D.IDENTIFIERS/OPEN ENDED TERMS C. COSATl Hfltl/Urciup

19. SECURITY CLASS (This Report) 21. NO. OF PAGES
Unclassified 45
20. SECURITY CLASS (This page) 22. PRICE
Unclassified
EPA Form 2220-1 (9-73)

-------
                                                      INSTRUCTIONS

 1.   REPORT NUMBER
     Insert the I I'A report number as il appears on the cover of the publication.

 2.   LEAVE BLANK

 3.   RECIPIENTS ACCESSION NUMBER
     Reserved for  use by each report recipient.

 4.   TITLE AND SUBTITLE
     Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
     type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
     number and include subtitle for the specific title.

 5.   REPORT DATE
     Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, date of
     approval, date of preparation, etc.).

 6.   PERFORMING ORGANIZATION CODE
     Leave blank.

 7.   AUTHORtS)
     Give namc(s)  in conventional order (John K. Doe, J. Robert Doe. etc.).  List author's affiliation if it differs from the performing organi-
     zation.

 8.   PERFORMING ORGANIZATION REPORT NUMBER
     Insert if performing organization wishes to assign this number.

 9.   PERFORMING ORGANIZATION NAME AND ADDRESS
     Give name, street, city, state, and ZIP code.  List no more than two levels of an organizational hirearchy.

 10.  PROGRAM ELEMENT NUMBER
     Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.

 11.  CONTRACT/GRANT NUMBER
     Insert contract or grant number under which report was prepared.

 12.  SPONSORING AGENCY NAME AND ADDRESS
     Include ZIP code.

 13.  TYPE OF REPORT AND PERIOD COVERED
     Indicate interim final, etc., and if applicable, dates covered.

 14.  SPONSORING AGENCY CODE
     Leave blank.

 15.  SUPPLEMENTARY NOTES
     Enter information not included elsewhere but useful, such as:  Prepared in coopetation with. Translation of, Presented at conference of,
     To be published in. Supersedes, Supplements, etc.

 16.  ABSTRACT
     Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
     significant bibliography or literature survey, mention it here.

 17.  KEY WORDS AND DOCUMENT ANALYSIS
     (a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
     concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.

     (b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc.  Use open-
     ended terms written in descriptor form for those subjects for which no descriptor exists.

     (c) COSATl FIELD GROUP • Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
     jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
     endeavor, or type of physical object. The application(s) will be cross-referenced with  secondary Field/Group assignments that will follow
     the primary posting(s).

18.  DISTRIBUTION STATEMENT
     Denote reusability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
     the public, with address and price.

19. & 20.  SECURITY CLASSIFICATION
     DO NOT submit classified reports to the National Technical Information service.

21.  NUMBER OF PAGES
     Insert the total number of pages, including this one and unnumbered pages, but exclude distribution b'st, if any.

22.  PRICE
     Insert the price set by the National Technical Information Service or the Government  Printing Office, if known.
Form 2220-t (9-73) (R*v*ru)

-------