EVALUATION OF MEASUREMENT METHODS
  AND INSTRUMENTATION FOR ODOROUS
  COMPOUNDS IN STATIONARY SOURCES
                 VOLUME I -
                 STATE OF THE ART
       U.S. ENVIRONMENTAL PROTECTION AGENCY

-------
EVALUATION  OF MEASUREMENT METHODS
          AND INSTRUMENTATION
        FOR ODOROUS  COMPOUNDS
         IN  STATIONARY SOURCES

     VOLUME I - STATE OF THE ART
                     by

             H. J. Hall and R. H. Salvesen

          Esso Research and Engineering Company
             Government Research Laboratory
          P.O. Box 8, Linden, New Jersey 07036
               Contract No. 68-02-0219
                  Prepared for
            ENVIRONMENTAL PROTECTION AGENCY
          Research Triangle Park, North Carolina
                  July 1972

-------
The APTD  (Air Pollution Technical Data) series of reports is issued by
the Environmental Protection Agency to report technical data of interest
to a limited number of readers.  Copies of APTD reports are available
free of charge to Federal employees, current contractors and grantees,
and non-profit organizations - as supplies permit - from the Air Pollution
Technical Information Center, Environmental Protection Agency, Research
Triangle Park, North Carolina 27711 or may be obtained, for a nominal cost,
from the National Technical Information Service, 5285 Port Royal Road,
Springfield, Virginia 22151.
This report was furnished to the Environmental Protection Agency by
Esso Research and Engineering Company Government Research Laboratory,
Linden, New Jersey in fulfillment of Contract No. 68-02-0219.  The
contents of this report are reproduced herein as received from the
Esso Research and Engineering Company.  The opinions, findings, and
conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency.
                       Publication No,  APTD-1180
                                   11

-------
                            TABLE  OF  CONTENTS
1.   Introduction ..........................    1

2.   Background ...........................    4

    2.1  Odors Vs.  Od .rants .....    ..............    4

         2.1.1  Characteristic  Odorants ..............    5

    2.2  Problems in Emission Measurements .............    7

         2.2.1  Range and Sampling .................    7
         2.2.2  Sulfide Interferences  ...............    8
         2.2.3  Point Source and Diffuse Source Effects  ......    8
         2.2.4  Effects of Odor Controls ..............    9

    2.3  Coulometric Titrations for  Total Reduced Sulfides .....    10

         2.3.1  Automatic Instrumentation .............    10
         2.3.2  Major Advantages and Disadvantages .........    11
         2.3.3  Types of Coulometric Cells .............    12

    2.4  Methods for Detailed Analysis ...............    14

         2.4.1  Infrared and Mass Spectroscopy ...........    14
         2. A. 2  Gas Chromatography .................    15

    2.5  Continuous Sensors ....................    15

3.   Odorant Sources Considered ...................    17

    3.1  Kraft Pulping .......................    17

         3.1.1  Sulfur Odorants Present ..............    17
         3.1.2  Sources of Emissions ................    17
         3.1.3  Sulfur Disposal ..................    19
         3.1.4  Measurement of  Odorants ..............    19
         3.1.5  Requirements for Monitoring ............    21

    3.2  Petroleum Refining/Petrochemical Operations ........    21

         3.2.1  Sulfur Odorants Present ..............    21
         3.2.2  Sources of Emission ................    22
         3.2.3  Petrochemical Operations ..............    23
         3.2.4  Sulfur Disposal ..................    23
         3.2.5  Measurement of  Sulfur Odorants ...........    26
         3.2.6  Instrumentation for Monitoring ...........    28
                                   ill

-------
                      TABLE OF CONTENTS (Continued)
    3.3  Animal Waste Rendering Plants ...............   30

         3.3.1  Odor ant Emissions .................   30
         3.3.2  Sources of Emission ................   30
         3.3.3  Methods of Measurement ...............   34

                3.3.3.1  Odor Panels ................   36
                3.3.3.2  Wet Chemical Methods ...........   36
                3.3.3.3  Instrument Methods ............   37
                3.3.3.4  Spectroscopy .... ...........   38
                3.3.3.5  Gas Chromatography ............   38

4.  Specific Instrumentation for S Odo rants   ...........   42

    4.1  Preferred Commercial ...................   43

         Coulometric Titration ...................   43
         Gas Chromatography ....................   53
         Tape Samplers .......................   60
    4.2  Commercial Alternates
         Conduct imetry ....... .  ......... «  .....   64
         Ultraviolet ........................   66
         Color imetry ........................   68
         Electrochemical ......................   72

    4.3  Potential .....................  ....   73

         Bioluminescence ........  .  .............   73
         Chemiluminescence ...........  .  .........   74
         Fast Scan Infrared ....................   75
         Fluorescence .............  .  ..........   73
         Infrared .  . .......................   81
         Infrared Laser Radiation .................   83
         Metallic Silver Filters ............  .!!..!   84
         Plasma Chromatograph ....  ...............   86
    BIBLIOGRAPHY	    37
                                 IV

-------
                            1.   INTRODUCTION
          This is a state-of-the-art  review made  to  collect  and evaluate
information on the instrumentation available to measure  the  quantities of
specific odorants in industrial plant emissions.   Its  object  is to examine
the possibility of substituting a quantitative measurement of Q dorants for
the subjective measurement  of odors as a guide for the control of odorous
emissions.  This report completes Phase I of EPA Contract No. 68-02-0219
on "Evaluation of Measurement Methods and Instrumentation for Odorous
Compounds in Stationary Sources."  Phase II of this  program will  consist
of the laboratory and field evaluation of instruments selected on the
basis of Phase I.

          The instrumental  measurement of odorants rather than odors is  a
valid approach only when there is prior agreement as to what odorants
are primarily responsible for odor. No reliable  correlations exist  in  the
general case between odor and the physical or chemical properties of
odorants, and the extent to which satisfactory correlations  have been
made to date varies greatly from one industry to  another.  This  is  examined
here for three specific industries which have been the subject of major
odor complaints:  kraft pulp mills, petroleum/petrochemical  refineries,  and
animal rendering plants.

          The degree to which instrumentation is  available for field use in
emission measurements can be related directly to  the complexity  of  the
odorant mixture to be measured.  The nature of the odorant mixture  in
kraft mill emissions is well defined, and instrumentation to measure it  is
commercially available.  The situation in petroleum refineries is more com-
plex:  a part of the odor problem is in stack emissions which can be
instrumented directly, but  another part is due to small amounts  of  material
from diffuse sources which  can only be measured  at present by ambient
methods.  The odor problem in animal rendering plants is even less  well
defined: there  is no agreement yet as to the identity of the specific
odorants most responsible for the odors observed, and instruments  for  their
detection and measurement are just now being developed.

          The odorants emitted in kraft pulping  consist primarily of four
reduced sulfides:  t^S, methyl mercaptan, dimethyl sulfide,  and dimethyl
disulfide.  The I^S and methyl derivatives are considered equally objection-
able in odor complaints, and they are frequently measured together  as
total reduced sulfide (TRS).  They are present in varying proportions
in different streams, together with moderate amounts of S02  which  is not
malodorous.  The literature provides ample evidence  that these four are
the key components both for odor emissions and for process changes  to
control them.  The much smaller amounts of higher sulfides which are
present have a secondary effect and they respond similarly to the methyl
homologs in odor controls.

-------
                                  - 2 -
          Petroleum/petrochemical refinery emissions also contain ^5 and
mercaptans, as key components.  The problem of odorant measurements is
much different from kraft mills, however, since a primary object of
petroleum refining is the removal and disposal of large amounts of sulfur
present in the original feed stock.  The refining process handles many
streams rich in H2S and other organic sulfides.  These are converted to
non-odorous products, chiefly elemental sulfur and SC>2.

          The principle emission point for residual sulfides in the refinery
is the burner stack of the sulfur plant.   This normally consists chiefly
of SC>2 and other combustion gases.  It may contain H2S or other sulfides
during periods of upset conditions.  Small amounts of carbonyl sulfide are
also present, and they normally exceed the amount of l^S.  The sulfur
plant in a modern refinery processes all collectible gas streams to remove
residual H2S and sulfides, before the gas is vented to the atmosphere.
The odorants emitted from small diffuse sources throughout the plant are more
varied in composition.  Key components include I^S, alkyl and aromatic
mercaptans, all of which will be included in an ambient measurement of total
reduced sulfur.  The massive amounts of S02 compared to reduced sulfur in
refinery emissions are a complicating factor, and the contribution of COS
requires further study.

          The odorant picture for animal rendering plants is much less
clear.  The composition of the odorants released is directly related to
the nature and quality of the material being processed.   Odorants such as
trimethylamine which are highly characteristic of the emissions from a
plant processing feathers, hair, or fish meal are minor components or
missing in plants processing animal flesh.  Other objectionable amines
such as skatole, putrescine and cadaverine are present only if the feed
to the plant is not properly washed, or if it contains badly decayed
flesh, which can be avoided.  The odorants recognized as generally
characteristic of animal waste rendering include aliphatic aldehydes,
free fatty acids, ammonia, and variable amounts of amines and H.2S
depending on the specific feed stock.   The laboratory instruments available
for the analysis of these recognized odorants are reviewed herein,  to
indicate preferred approaches for a prototype for field  use.

          The continuous automatic instrumentation available for the direct
measurement of total reduced sulfide depends upon two types of sensors:
bromine coulometric titration, and the flame photometric detector.
Coulometric titration can be carried out continuously in a flowing stream,
or in a microcoulometer which accepts very small discrete samples such as
the fractions eluted from a gas chromatographic column.   Both approaches
are very useful for differential measurements, but they are subject to
difficulties with zero stability which require frequent calibration if
absolute values are desired.  Coulometric cells can also be based on titra-
tion with iodine or with silver ions,  which respond to different portions
of the total sulfur present.

-------
                                    - 3 -
          Flame photometry responds equally to all sulfur atoms.   It  can
be used as a total measurement, or more desirably in a GC/FPD combination.
This permits a distinction between S types such as I^S, SC>2 ,  total S, or
some other selected component such as methyl mercaptan.  Automatic
programmed equipment for such measurements is available.

          Continuous measurements can also be made using lead acetate
tape samplers or similar automatic instruments which are sensitive only to
     or to H2S and mercaptans as key components.
          An alternate approach is to convert all sulfur compounds present
to SC>2, with or without a preliminary separation of the SC>2 initially
present, and use this as an analysis of sulfides in the sample.   This
combination has been applied particularly with a conductivity cell or a
microcoulometer for the detection of S02-  It could also be used with
any of the very large number of S02 detection systems available.  The
analogous approach of catalytically or thermally converting all sulfur
compounds to I^S has also been used, particularly with the silver cell
microcoulometer or lead acetate tape as the sensor for H2S.

          Wet chemical methods which consume reagents are considered
intrinsically less desirable for field measurements than the use of
electrical or optical sensors, but they are much more flexible when
analyses of varying composition are required.  Most of the present standard
methods for product quality measurements in petroleum refining are based
on wet chemistry.  Many of these measurements can be automated, if enough
samples are to be measured in a given time.  Nevertheless, they have been
excluded as far as kraft mills and refineries are concerned in the
present review,  which is aimed primarily at the evaluation of physical methods
of measurement.   Automated wet chemistry may well be a preferred approach
to the analysis for odorants in animal rendering plants, where both
qualitative and quantitative base data are still required.

          The review which follows presents background information in
Section 2 for the basic definitions of the problem of odorant emission
measurements, and the types of instrumentation available.  Section 3 gives
details of the three industries considered as odorant sources.  Section 4
summarizes the information on specific instruments which are commercially
preferred, on available alternates, or potential alternates where further
development is required.

-------
                          2.   BACKGROUND
  2.1  Odors vs. 0dorants

            The measurement of odors is the one problem in air pollution
  In which the subjective response of human sensors has continued to be
  most important.  At present, there are no known physical or chemical
  properties of odorous materials which correlate with their odor, and the
  only accepted method of measuring odors as such is a panel of human observers,

            Odor is measured on a hedonistic scale, in terms of degrees
  of pleasure or obnoxiousness.   The usual scale has five units,  from zero
  or undetectable to  intolerable or overpowering.   This scale,  like  many
  biological reactions,  is related exponentially to the amount  of the
  stimulus,  so that  it is  linear by orders of  magnitude.   A  typical  curvp
  is that shown by Wright (26) for ethyl mercaptan, in Figure 1:
                                Figure 1
J
I







X




1
X*




A



4
^^
Of



>




x'



- *«
* I




)








t



0
^;



^
1 0



<




-»




*



e
*'



rf




^




,•>
i




                •V     IV     *V     10     '**     ^^     «W     W     rv
                              ETHYL MERCAPTAN
                              MK1K>««AM« KK LITER

 The  relationship  between  odor  intensity and  odorant  concentration  is
 considered as  a classic example  of  the Weber-Fechner Equation  (I):

                             P  =  K log S

 where P = odor intensity, K =  constant,  S =  odorant  concentration.   Values
 of K for different materials range from  0.3  to 0.6;  at 0.5,  one unit  on
 the  odor intensity scale  corresponds  to  two  orders of magnitude in
 concentration.

           The  statistical reliability of  odor panel  measurements involving
 several individuals  is considered no  better  than  + 1 unit  on the intensity
 scale.   Recent  data  on a  biological response to chemical stimulus have  a
 bearing on such measurements,  from  results obtained  with a bio-luminescence
 detector.   This device measures  the change in light  emission of a  luminescent
 bacterial  culture, on exposure to a chemical to which the  particular  strain
 has been sensitized.  This is a biological response which  is free of subjective
 influences.  The  detector can distinguish clearly between  S02 concentrations of
0.2,  2.0 and 20 ppb (ICT^), with about two to three levels of discrimination
between each order of magnitude.

-------
                                   - 5 -
          The number of odor levels between odorant orders of magnitude
corresponds to the constant in the Weber-Fechner equation.  The analytical
significance of this coefficient is that it takes a shift or change of
some -50% to +100% in a measured concentration to be biologically
detectable as a difference.  This means that odorant measurements within
+_ 20 to 50% may be quite acceptable for odor correlations, even though
they would be considered marginal or worthless from the viewpoint of
conventional analytical chemistry.

          If odorous compounds are to be measured for purposes of pollution
control, it is necessary to determine first what compounds to measure,
and where, and in what amounts.   The determination of what to look for
is a major problem.  It has been estimated that the human nose is sensitive
enough to determine concentrations of the range of 10^ molecules per ml.
of air, and the stated limit of detectability is continually going down
with improved methods of test.  The odor threshold for H2S was stated as
0.13 ppm in 1930 (2), 20 ppb in 1962 (3), and 0.5 ppb in 1968 (4).  The
total effect is usually due to a mixture of many odorants, at extremely
low concentrations.  Dravnieks has estimated (5) that- the odors emanating
from a human being consist of from 150 to 500 different compounds or
groups of compounds, but that as few as 15 groups are enough to distinguish
types of humans such as Caucasian male, Caucasian female, and Asian Indian
(vegetarians).  If only a  limited number of compounds are to be measured
they must be chosen on the base of experience, or the statistical analysis
of large amounts of data which indicate that a useful correlation between
odor and odorant does exist.  The odorant approach is valid only when
there is prior knowledge that the particular compound to be measured is
significant (45).

     2.1.1  Characteristic Odorants^

          One class of odorants for which the connection between odor
and chemical composition is clearly known are the volatile sulfides,
including H2S (rotten egg  odor), the mercaptans  (skunk odors), and the
organic sulfides which are the thioethers of the mercaptan thio-alcohols•
The human odor response is highest for methyl and ethyl mercaptan, a
little higher for ethyl (lower threshold), and in each case the odor
response is somewhat lower (higher threshold) for the sulfide than it
is for the corresponding mercaptan.  These relationships are shown in
Figure 2.

          Hydrogen sulfide, mercaptans and organic sulfides as a group
are easily recognized by their odor, and known to be present in the
uncontrolled emissions from pulp and paper mills and from petroleum
refineries.  Further studies have shown that H2S and methyl mercaptan
and its derivatives are usually by far the major compounds of this series
present, and that they may be considered as key components for the
measurement of sulfides as a whole.  The recognition that they are
important odorants has led to the development of automated industrial
chemical and instrumental methods for their measurement, which are
reviewed further herein.

-------
                   - 6 -

                   Figure 2


Variations In Odor Threshold with Number of C Atoms

    in Several Series of Non-branched Homologs (5)
   8


   I
   ex
   JJ
   QJ
   H

   8
  T3
  1-1
   O
   (0
  o
  g
          -2
-3
       -- 4
        "5
       O
          -7
          -8
          -9
                             CHLORO-
                            PARAFFINS
               I            5             10    12

                   NUMBER OF  C ATOMS

-------
                                - 7 -
          In many industries there is no such body of information as to
which specific odorants are most important, either in terms of quantity
or in terms of contribution to the odor effect.  This is the case for
the odor emissions from animal waste rendering plants.  The total odor
effect from these sources is strongly affected by the presence of
particulates or moisture droplets, which can adsorb enough odorant near
the source to carry a strong odor stimulus in individual particles to a
distance.  Under these conditions, the usual relationships governing
dilution effects by atmospheric dispersion do not apply.  Qualitative
data are available to indicate that aliphatic aldehydes, free fatty acids,
and alkyl amines are present as major odorants, but very little is known
as to their amounts, or what changes these amounts during plant operations.
In this situation, the collection of data on selected odorous materials
in terms of their amount and relationships to odor panel measurement
on the one hand and plant operating conditions on the other are a necessary
first step.  Instrumentation suitable for measurements of the mixed
aldehydes, free acids and amines characteristic of animal waste rendering
odors is available, but only at the research level.  It is still to be
assembled into a prototype and tested before it is available for use in
the field.

2.2  Problems in Emission Measurements

          The measurement of odorants at the emission level rather than
at ambient dilutions presents special problems of sensor range, sampling
procedure, and interference effects.  The specific nature of these problems
depends upon the source being measured.

     2.2.1  Range and Sampling

          In the case of kraft pulp mills, the total emission of l^S and
organic sulfides ranges from about 300 to 600 ppra in the older uncontrolled
operations, to less than 3 ppm in the best modern mills.  This is commonly
referred to as total reduced sulfur (TRS).  Regulations now in force or
set in various jurisdictions specify a total allowable sulfide emission
of 17.5 ppm.  TRS measurements in the emission range have been standard
practice in the industry for purpose of process control, and the sensors
in commercial equipment are designed for this range   (6)

          The problem of sampling in kraft mill odorant measurements is
serious, and it has not been entirely solved.  The stack gases which are
the principal emission source are hot, wet, and dirty, and subject to
rapid and wide fluctuations in concentration and amount.  Major efforts
to improve this situation have been made by individual industries and
the industry's trade association NCASI (National Council for Air and
Stream Improvement).  The standard Barton Instrument probe includes
integral filters and a timed blow-back cycle designed to minimize these
problems.  All measuring instruments, however, are subject in some degree
to the complications caused by the inability or failure of the sampling
system to provide a clean gas sample for analysis.  Thus, the ability
of a given instrument to handle or withstand poor samples is an important
characteristic for field evaluation tests   (7).

-------
                                 - 8 -
      2.2.2  julfide  Interferences

          The  reaction  of  SC^ as an oxidizing agent  toward  reduced sulfur
 compounds and  particularly toward I^S  can cause serious interferences in
 the  emissions  concentration range.  This is a major  problem in stack
 sampling.  The reaction is  promoted by heat, moisture and reactive
 particulate surfaces, all  of which may be present.   The result is to
 form water and elemental sulfur which drop out in  the sampling line, and
 can  easily clog  the  apparatus while vitiating the  analysis.

          The  sulfur compounds of interest in kraft  mill and petroleum
 refinery emissions have many chemical and physical properties in common.
 An instrument  based such  a non-fcpeciflc  property will  respond to  any sulfide
 present.  With such an  instrument, all other compounds in the group will
 count as interferences  when only one of them is of interest.  This makes
 it relatively  easy to analyze for total sulfur, but  correspondingly
 difficult to measure say H-S alone, or total reduced sulfur in the
 presence of large amounts  of SO^-

          The  ability to discriminate between individual sulfur compounds
 or types when  the sensor responds to all depends on  a preliminary separation
 into fractions, which may  then be separately analyzed by the same sensor to
 measure their  sulfur content.  Separations of this type can be made by
 a series of chemical absorption solutions (8), by a  similar series of
 dry  absorption filters  (9)  or by gas chromatography   (10)  Any of these methods
 can  be applied to a sensor which does not discriminate between sulfur types,
 such as a bromine coulometer or a flame photometric  detector.

      2.2.3  Point Source^and Diffuse Source Effects

          While  measurements at  the emission levels  of 1-100  ppm  correspond
 to standard industrial  practice in kraft pulp mills, this is not  true in
 many other industries.  In petroleum refineries, instrumentation  for sulfide
 measurements has been primarily at the percent level, for process streams
 rich in H2S or sulfur-containing hydrocarbons, or  at ambient  levels near
 or below 1-10  ppm for toxicity alarms and air pollution effects.  No
 instrumentation  has been regularly applied to the  measurement of  odorous
 emissions from animal rendering plants, and new proposals for such
 measurements are just now  appearing in the literature.

          Emissions measurements apply only to effluents from a discrete
 source such as a stack  or  vent line, while the total odor effect  of a
 plant may include many  diffuse sources.  Both are  important in plants
 such as a petroleum refinery, or animal waste rendering, where process
 flow streams include odorants in high concentration.  The odor effects
 of many small  leaks, each  in low amount, can only be evaluated by ambient
 measurements.

          Ambient measurements are of more interest  than emissions
measurements as far as  odor thresholds and esthetic  effects are concerned.
No direct health effects effects are expected below  I ppm,  however,  in the
 emissions range, even with materials as toxic as H2S (16).

-------
                                - 9 -
     2.2.4  Effects of Odor  Controls

          Emissions measurements frequently apply to the analysis of
effluents before and after a treatment to reduce their amount.  Odor
controls for this purpose may remove the odorant by oxidation or other
rhemical conversion, combined with dilution by ventilation in rrwfi~*A
areas or dispersal in open air, or by absorption or scrubbing.  Chemical
conversion by oxidation may be effective only if the oxidation is complete,
resulting in final products which are either odorless or much less
odorous than their precursors   (11,  12,  13,  14),  Partial oxidation can
increase odor.  Direct flame incinerators are used for air oxidation
employing temperatures of 1,100 to 1,500°F, while catalytical decomposition
systems operate at temperatures 500-800°F lower   (11, 13, 14),  Chemical
conversion by oxidizing agents such as ozone, permanganates, and chlorine
compounds does not result in complete oxidation>and it is important to
appraise the actual degree of odor reduction obtained.  Proper incineration
is capable of reducing odor discharges by more than 99.99% in some instances,
and only rarely by less than 99%   (12, 13).  Control by absorbents is
practically synonymous with  the use of activated carbon,  and has usually
been found effective, though expensive if applied to  large amounts of
odorant.  In contrast to these results,  the effectiveness of odor control
by time honored water-spray scrubbing techniques has  been found  to be
quite variable—from less than 50% in some installations  to about 90-95%
at best   (16)

          Partial oxidation  or oxidation/reduction reactions  frequently
introduce new compounds into the  odorant mixture.  An important  constituent
of this type is carbonyl sulfide.  This  appears in the stack  gases from
a kraft mill when the recovery furnace is overloaded, and is  a regular
constituent of  the emissions from the Claus suxfur plants used instead
of burning for  the disposal  of refinery  streams rich  in I^S.  The  odor
response of COS can vary markedly from one individual to  another,  and  it
is not  included in  the usual definition  of TRS  as  H2S and total  organic
sulfides.  The  response  of a given  instrument  to  COS  is  important either
way,  since  it may be included with  H2S  (by FPD),  excluded (Barton or  Philips)
or give a partial response  (Dohrmann silver  reductive cell).  Further
research is required on  the  significance of  this  compound in the measurement
of odorants.

-------
                                 - 10 -
 2.3  Coulometric  Titrations  for Total  Sulfide

      2.3.1   Automatic  Instrumentation

           Automatic instrumentation  for the  measurement  of  organic  sulfide
 emissions was an  industrial  outgrowth  of U.S.  government research at  the
 end of World War  II.   The  original problem was  to  detect and  record very
 minute concentrations  of mustard  gas in air.   A highly sensitive coulometric
 cell capable of making this  measurement was  based  on  the chemical oxidation
 ot  the organic sulfide with  electrolytically generated bromine:

              RSR  +  Br2 + H20  	=*  R2SO + 2H+ + 2Br~


           The measurement  can  be  made  continuously on a  flowing stream, in
 terms  of  the bromine required  Lo  react  with  various forms of oxidizable
 sulfur present in the  sample.  In the  coulometric  cell,  the potentiotnetric
 end-point and electrolytic generation  are  coupled through a feed-back
 amplifying  system.  This controls the  rate of reagent generation so that
 it  is  at  all times  equal to  the rate of reagent disappearance in the gas
 titration.   The current  required to  generate the bromine  a  function of the
sulfide concentration.

          The commercial instrument  developed on this basis by Austin  at
 the  Consolidated  Engineering Corp. in Pasadena was known as the Titrilog
 (16).   It was originally offered for the monitoring of natural gas  odorants,
 air  pollution studies,  and a variety of uses involving the detection of
 hydrogen  sulfide.   The  instrument has  gone through several generations of
 improvements,  and is the predecessor of the Barton Titrator now available
 as Models 286  and 400.

          The  coulometric cell proved  to be very useful  in the measurement
 of odorous  emissions in  the pulp and paper industry, which  consist  primarily
 of sulfur compounds oxidizable by bromine.  Early  demonstrations of its
 success showed that the  measurement  of  total reduced  sulfur in these emis-
 sions  was a powerful tool to follow  the results of process changes  on  the
 amounts of  odorous  emission.   Its speed of response and  sensitivity were
 used to advantage in the study of new  methods of odor control, particularly
 in Kraft mills.   Improvements were made  in the  Barton Titrator to increase
 its  sensitivity and range to both lower and higher limits, so that  the same
 instrument  can be used for direct measurements  from 0.01 to 1000 ppm of
 total  reduced sulfur.   This covers both the  ambient and  emission ranges.
 Its  use in  Kraft  mills  has expanded  rapidly  in  the last  five years, using
 improved ^methods  of sampling and sample  pretreatment  developed by workers
 at Crown  Zellerbach (17) and Weyerhaeuser (8)  in Washington,  and in a  series
of research  and field projects sponsored by NCASI, the pulp and paper
 industry's National Council for-Air and  Stream  Improvement.

-------
                                   - 11 -
            Bromine  coulometry  can be used to measure hydrogen sulfide,
   sulfur dioxide,  mercaptans, thioethers, organic  disulfides and thiophenes.
   This  includes  all  of  the  organic sulfur compounds which are recognized as
   characteristic odorants in Kraft mills, and the  same compounds as character-
   istic odorants in  petroleum and petrochemical  refining.   It also includes
   the two major inorganic  sulfur pollutant gases,  H2S and  S02.   Sulfur
   trioxide,  carbonyl sulfide and carbon disulfide  are inorganic  gases  which
   do not titrate with bromine.


             Iodine coulometric  cells  have also been developed commercially by
   Austin (a Barton prototype),  and Dohrmann  (oxidative microcoulometer).
   Iodine titrates S02,  H2S, and mercaptans,  but  does not react with other
   organic sulfides or disulfides.  Dohrmann  oxidative cells of this
   type are widely used.  Theoretically, the difference between analyses with
   an iodine cell and a bromine cell could be used to calculate the total of
   these sulfides and disulfides, but  this approach over-emphasizes small
   differences  between the  two cells  and it has not proved  practical  (18).

   2.3.2  Major Advantages  and Disadvantages

            1.   All  reduced sulfur compounds react:

            The fact that   one instrument responds  to such  a variety  of
   hydrogen sulfide and reduced  organic  sulfur compounds has been a distinct
   advantage since all of these  compounds appear  together,  in  varying
   proportions,  in the odorous emissions which are  to be measured.  However,
   the lack of  discrimination between different types of sulfur  compounds  is
   also a disadvantage.

            2.   Conversion  factors differ for different compounds.

            The amount  of  halogen  required for the titration  is  not  the same
   for different sulfur types, so that the conversion factor between  cell
   output and  sulfide concentration varies with a varying  ratio of sulfur
   compound types. The  exact course  of  this  oxidation  is not known in
   detail, and  it does not  correspond stoichiometrically to  any single  react-
   ion.   The following reactions are  typical (19):


                                          S Compound   Conversion Factors  Observed
   Theoretical Reaction with Bromine       Mol. Wt.   (x millivolts • ppm.  IQx  range)

                S + 2HBr                      34                   0.86

CH3SH -I- 2Br2 + 2H2<>	*  CH.SO.H + 4HBf      46                    1.3

                                     2HBr     66                    3.5


                                              64                    2*°

-------
                                      - 12 -
           The problem which  this  creates  can be ignored by using the
 factor  for H2S  and  reporting the  total  calculated  "as H2S".  The value
 thus  obtained is  called  "total  reduced  sulfur".  This is justified as a
 simple  method to  get  a useful result, because the  compounds measured are
 usually present together,  and they  are  removed together by the usual
 methods of control.   If  it is known, however, that another compound type
 such  as mercaptan predominates  in a mixed stream,  more accurate data may
 be obtained by  using  its factor instead in the data conversion.  For any
 higher  accuracy It  is necessary to  carry  out a preliminary separation into
 individual compounds  or  compound  types, by subtractive chemical absorption
 or otherwise, and measure  each  fraction using its  appropriate conversion
 factor.  All three  of these  approaches  are used.

           3.  Prefilter  separations are not exact.

          Unfortunately, the  separation by sulfur  compounds types which is
 obtained by chemical  absorption filters is only approximately quantitative.
 Automatic equipment is available  to make  such measurements in a programmed
 cycle,  but it is not widely used  except for research.  The TRS measurement
 is much simpler to make.   Even  though it  is admittedly a compromise, it
 is considered a reliable basis  for the measurement of differences* to indicate
 the direction and magnitude of  changes  in odor ant  concentration as a result
 of process changes  in the  plant.

          4.   Interaction  effects are possible.

          A further limitation  of the coulometric  cell in the discrimina-
 tion  of individual  compounds  or types in  a mixed sample is the possibility
 of secondary interaction effects  on the stoichiometry of the oxidation
 reaction, when  there  are marked changes in the ratio of sulfur types
 present.  When  this occurs, the results obtained- in a mixed sample will
 not exactly equal the sum  of  the  results obtained on separate fractions.

2.3.3  Types  of Coulometric Cells

          Coulometry is based upon the principle of electrically generating
 a selected titrant  (ion or element)  in a titration cell.   This may be a
 free halogen (bromine or iodine)  in  aqueous solution, as an oxidizing agent,
 or a metal ion  (silver)  as a  reducing agent.   The electrolytic current re-
quired to generate the titrant as  it Is consumed is a linear measure of
the concentration of reactable compounds in the  gas sample.   In practice,
the titration cell has two sets  of electrodes—a generating pair and a
reference pair.   The electrical  relationship  between these  differs in
different types of cells.

          1.   In the Barton bromine  cell and the Philips  S02  Monitor bromine
 cell a constant low level of titrant is continuously ijalntafned  to produce a
 "zero level" on the instrument.    When reactable compounds enter the cell,
 the available titrant is consumed.  The associated feedback amplifier
 responds to the change in  sensed  titrant level and generates sufficient
 additional titrant to maintain the "zero level" concentration.  In this
type of cell, the generation  rate of free halogen  as titrant Is governed
by a control  circuit which senses  an excess of halogen,  at  a level fixed
by the depolarization of and flow of current  through the  reference electrode.
In this case  the generating current  required to  maintain  that excess is
measured, and is proportional to the measured component.

-------
                                  -  13 -
          2.  In a second type of titration, the cell is driven by a constant
 current source.  The consumption of free halogen prevents reaction at the
 cathode and current flows instead through the reference electrodes.  The
 reference current is measured and is proportional to the decrease in halogen
 concentration and thus to the concentration of reactable pollutant in the
 sample.  (Dohrmann Microcoulometer, Beckman 906A) .

          3.  In another type of cell, hydrogen is allowed to build up on the
electrodes.   Current flow,  limited by  diffusion to  the electrode surface,
is proportional directly to halogen concentration  and inversely to the
 concentration of reactable pollutant.   (Atlas 1203).

          The  Barton Titrator Model 400 for  pulp and paper plant use is  a
 modification of  the earlier Titrilog  (19), which adds an optional SC>2 filter
 cycle  to  the basic  total sulfur Model  286.  A high  degree of selectivity
 in S02 removal  in obtained by absorption in  3% aqueous  potassium acid
 phthalate.  Total sulfur includes S02> and  total reduced sulfur includes H2S,
 mercaptan,  sulfide, and disulfide.  These sulfur compounds can be separated
 by type if  desired  and measured separately,  using an auxiliary  system of
 absorption  filter cells which can be automated (Model 327).  The Titrator
 measures  bromine oxidizable sulfur gases through the entire  range from
 10 ppb to  1000  ppm.  This  is  the widest range  for direct measurements of
 H2$ and TRS of  any  instrument which has been available  for purchase, and
 the convenience  of  using the  same instrument for both ambient and emission
 measurements has been one  factor  in its acceptance  in the  industry.

          Microcoulometric  cells which operate on small discrete  samples
 rather on  a flowing stream have become widely  accepted  during the past
 two or three years  as preferred methods for  the research laboratory
 determination  of low concentrations of sulfur, in either Kraft  mill  gases
 or petroleum products.  Their major advantages are  sensitivity  and  speed,
 and the fact that they  require only microliter amounts  of  liquid  sample
 or milliliter  quantities of gas.

           There are two principal  types of microcoulometer cells.   The
 oxidative cell was  originally developed as a specific gas  chromatographic
  detector,  and  linked directly to the  column outlet through a combustion
 unit which converts all S  compounds to S02 for the  titration.   An alternate
 is the reductive microtitration  cell  utilizing silver-silver chloride
 electrodes, used for  the  coulometric  analysis  of hydrogen  sulfide  and
 raercaptans  (2jD) . The  reductive  cell  will  not  respond to  S02,  or  to  alkyl
 sulfides  and  disulfides.   The reductive method is  usually  combined  with  a
 preliminary hydrogenolysis by which  the  sulfur compounds  present  in  the
 sample are reduced  to hydrogen  sulfide, over a Pt/alumina  catalyst  in a
 stream of hydrogen, and titrated with silver ions.   The silver cell can
 also be used  for the  determination of ammonia, or ammonia formed from
 nitrogen  compounds  by catalytic  reduction  with hydrogen.

           The microcoulometer,  originally  combined with a preliminary
 decomposition by combustion to S02 or hydrogenolysis to H2S  prior to
 titration, may also be  used to  advantage  for the  direct titration of the
 sulfur species which  react.   These are SC>2 in  the oxidation  mode, and
 H2S or RSH in the  reductive mode.   t^S and RSH also react ia the oxida-
 tive mode, but with less  favorable conversion  factors.

-------
                                 - 14 -
           Commercial models of the microcoulometer are marketed by the
 Dohrmann Instrument Co. (now Envirotech Corp.)-  The original model was
 an iodine cell, and this has been used extensively in the petroleum industry.
 Its effective range covers three orders of magnitudes, from about 0.2 to
 1000 ppm of sulfur.  The addition of sodium azide to the electrolyte over-
 comes the interference effects of small amounts of chlorides or nitrogen
 oxides  (21), which are often responsible for short electrolyte life.  The
 original design of the iodine cell was modified and converted to bromine
 coulometry by Adams (22) with improved sensitivity for ambient measurements
 at concentrations down to 3 ppb of hydrogen sulfide.  Both the iodine and
 bromine cells are available from Dohrmann (Envirotech).

           The zero adjustment in the Barton and Dohrmann cells is an approxi-
 mation which is never exact, since it is a dynamic balance involving the
 small amount of bromine gas physically swept out of the cell by the constant
 flow of sample carrier gas, in the absence of reactive sulfur compounds in
 the sample.  The effect has been minimized in the Philips cell by changes
 in cell geometry and improved mixing within the cell.  This zero and the
 cell readings above it also respond sensitively and promptly to any change
 in the flow rate, pressure, or molecular weight of the gas flow through
 the cell.

 2.4  Methods for Detailed Analysis

      2 ..4 ..1  IR and_ Mags Spectrosgopy

           The determination of individual compounds in a complex odorous
 plant emission is a laboratory research problem in which the most
 sophisticated methods available can be employed.  Basic studies of this
 type have been made using both infrared and mass spectroscopy, and such
 measurements are necessary to define the specific compounds which need to
 be measured for purposes of monitoring.  This includes both qualitative and
 quantitative data on composition.   In the general case of odorous compounds
 it is necessary sooner or later to make a further correlation by data or
 by analogy between the amount of a given odorant and its contribution to
 total  odor response.   However,  such  a correlation,  while  it  is required,
 is not  a part  of the  present project.   Once  the nature  and concentration
 range  of the odorous  components to be measured has  been  defined,  a much
 simpler method is desirable for field use  to measure  a  selected group  of
 key componerttf.'

           The  complete  analysis of individual  components  by  infrared,  mass
 spectrometry or  other  sophisticated  spectroscopic methods gives definitive
 results, but these  methods  are  time  consuming,  expensive, and not  suited
 for routine use  in the field (6).   The odorants to be measured by instru-
 mentation  for  H2S and  TRS in Kraft mills and in petroleum refining are
 known,  as  the  result of prior studies»  and it  is necessary to control
 them all if complaints  from obnoxious odors  are to  be avoided.   The research
 methods are still useful and used  to great  advantage  for  special purposes
 such as the study of trace  impurities or the mechanism of reactions, but
 the large  amount of additional  information which they supply is neither
-accessary  nor  desirable for routine  use in the monitoring of emissions.

-------
                                 -  15  -
          It is also possible to make simplified equipment  using either
infrared or mass spectroscopy which is suitable for field use  to measure
a single pollutant compound in air, but IR is not a good method for I^S
and the other odorants of interest are a complex mixture.   While further
developments may appear along these lines, other approaches have turned
out to be simpler, to date.

     2.4.2  Gas Chromatogr&

          Gas chromatograp hy (GC) has been used to great advantage  in the
analysis of odorous emissions (9).  It was one of the early industrial uses
of this new technique, starting in 1957 with studies by the British Columbia
Research Council (23) and in parallel studies since 1958 by Adams and co-
workers at the University of Washington (24).  The GC technique can be
used either for complete analysis or, by a proper choice of substrate and
elution procedures,  it may be used for a short-cut analysis in which a
few compounds of particular interest are measured first.  In a preferred
technique,  the rest  of  the sample  can  then be back-flushed out  of the column
and discarded, to restore the column to its original state  for another run
in minimum elapsed time.

          One limitation on the sensitivity and discrimination of the GC
technique was originally the type of sensor available to detect changes
in composition, for the minute amounts of material in the succeeding fractions
eluted from the column.  The original analysis using thermal conductivity
cells which responded to about 500 ppm of sulfur compounds on direct
injection was  improved by substituting the flame ionization detector (FID)
which responds to about 5 ppm of organic sulfide, and still further improved
by the flame photometric detector  (FPD) which is sensitive to 3 ppm of
sulfur compound in the GC  concentrates.  The FPD responds equally to sulfur
in S(>2, sulf ide, or any sulfur compound which is combustible in a hydrogen-
rich flame.  The combination of GC/FPD is just now becoming available in
commercial instruments which are offered as Bendix Model 8700 and Tracer
Model 250H, and Varian Aerograph Model 1490.

2.5  Continuous Sensors?

          Over the past two decades, instrumentation researchers and
manufacturers have put most of their efforts into the measurement step
of the analytical process.  Sensors that measure all sorts of properties,
for both  identification and quantification, have been developed to a high
degree of analytical  sophistication.   Some of these sensors have high
sensitivity, low holdup, and rapid response, and are suited with little
or no modification to continuous analysis.  Developments along  these lines
have been summarized  in detail by Blaedel (25). The methods listed in
Table 1 have been used or  recommended  for the continuous analysis of organic
sulfides  or their conversion products, H2S or

-------
                                    - 16 -
                                   Table 1

               Continuous Sensors for Sulfur Compounds


   Electrical sensors	   Optical sensors
   Conduetometrie                            Colorimetric (speetrophotometrlc,
                                             visible region)
   Coulometric  titration:  Bromine          Fluorimetric
                            Iodine           Infrared spectrophotometrie
                            Silver ion
   Flame lonization detectors                Turbidimetric
   Polarographic                             Ultraviolet spectrophotometric
   Potentiometric:  specific ion electrode
   Potentiometric titrations                 Mass spectrotnetric
   Thermistor:  thermal conductivity
          The ability of an instrument to make continuous measurements rather
than cumulative measurements in the field is a desirable characteristic, and
it is essential when the readings obtained are to be related directly to
process controls.  Continuous sensors of the types listed are discussed further
in Section 4 of the present report.

-------
                                 - 17 -


                     3.   ODORANT SOURCES CONSIDERED

3.1  Kraft Pulping

     3.1.1  Sulfur Odorants Present

          The characteristic odor for which kraft pulp mills are known
as a major source of air pollution is due to a mixture of reduced sulfur
compounds consisting largely of hydrogen sulfide, methyl mercaptan, and
its derivatives dimethyl sulfide and dimethyl disulfide.  Very much
smaller amounts of higher sulfides are present (26).  The methyl deriva-
tives predominate because of the chemical structure of wood, for the same
reason that methyl alcohol is formed in wood distillation.

          The same four compounds appear in all of the gas streams pro-
duced in kraft pulping, in varying proportions.  Terpene hydrocarbons are
also released in the treatment of soft woods and may contribute to odor,
but they are not considered as objectionable as the sulfides and mercaptans,
Sulfur dioxide appears in the stack gases from the recovery furnace and
the lime kiln, in an amount which is normally low or moderate compared to
total reduced sulfur.

          The sulfur odorants emitted in the kraft process come from the
pulping reaction.  They are not part of the feed to the process (wood)
or the product (pulp), but they are chemical products of the reaction,
and the sulfur which they contain is reusable in the process.

          Sulfite pulping is an alternate to the kraft process which pro-
duces no H2S or organic sulfide emissions, only SC>2.  Its use is limited
and decreasing, however, since it is employed primarily in reprocessing
plants and on a limited variety of woods which are not those most readily
available.  The present review is based on kraft pulping.

     3.1.2  Sources of Emission

          The pulping process is a means of chemically dissolving the
lignin in wood away from cellulose fibers which are then made into paper.
The kraft pulping reagent is an aqueous solution of sodium sulfide and
hydroxide,  A major feature of the process, shown schematically in
Figure 3, is its ability to recover most of the chemicals from the spent
treating liquor and re-use them for the next batch  (26).  Sodium sulfide
is continuously regenerated from the l^S and organic sulfides released
and recovered in the cycle.

          The combined fresh and regenerated white  liquor is mixed with
wood chips and cooked for 3 hours in one of a battery of large upright
digesters, with steam at a pressure of approximately 110 psig.
Digester gases are relieved periodically during the cooking period to
reduce the build-up of pressure within the system.  At the end of the

-------
                                  -  18  -
 cooking,  the bottom  of  the digester  is  suddenly opened  and its contents
 forced  into a blow tank.   Here the major part of the  spent cooking liquor
 containing dissolved  lignin is drained  off.   The raw  pulp is filtered  free
 of undissolved knots, washed, bleached,  pressed, and  dried ready for
 further use.
                                  Figure  3
              ScheTT.atic Diagram of the  Kraft Pulping Process
             "FLIEF
                                                                             $
                                            WON
                 CH,SH.CMjSCMj.H,S

                    CONXNSABLLS
TURPEWTlNt
    /
                                          CONTAMINATED WATER
                            OONIAMPN'ATED WATER
                            AIR
                       OXIDATION
                             I  STEAM,CONTAMINATED WATER,KjS, 6. CHj
              13 X SOLIDS

       Sf'E'iT AIR . CHjSCHj f
i
i


EVA
           The spent black liquo*- from  the  blow tanks is  concentrated to
50-60% solids by evaporation,  and better results are realized by an
oxidation  of this black  liquor before  the  first stage of  evaporation.
This reduces corrosion in the  evaporators  and decreases  the loss of

-------
                                  - 19 -
sulfides in the gases by converting them to partially or completely
oxidized compounds which are less volatile.  It thereby conserves
reagent sulfur, while decreasing the emission of sulfur-containing
odorants into the atmosphere.  The black liquor may be further con-
centrated in a direct contact evaporator, by heat exchange with the
recovery furnace flue gases.  Much lower emissions can be realized
by replacing this direct contact evaporation with an indirect heat
exchange system for the final concentration, which avoids re-evaporating
odorants into the stack gases.  Several designs of this type are available.

          The concentrated black liquor is combustible because of its
lignin content.  This is burned in the recovery furnace under lean air
conditions, so that the inorganic chemicals present are reduced to a
melted mixture which is mainly sodium sulfide and sodium carbonate.
This smelt is removed from the bottom of the furnace, dissolved to form
a green liquor, and the carbonate is reconverted to sodium hydroxide
by slaked lime in the causticizer tank.  The spent mud from this tank
is sent to a lime kiln, where it is burned to regenerate calcium oxide
to complete the cycle.

     3.1.3  Means of Sulfur Disposal

          Sulfur containing gases collected from the digesters, blow
tanks, evaporators and miscellaneous washers and filters throughout the
plant are disposed of by burning in the lime kiln or the recovery  furnace.
The control of emissions by this means requires a careful adjustment of
furnace loading and operating conditions to keep the system in balance.
A summary of the principal emission sources to be controlled and typical
analyses of their odorant composition is given in Table 2 (27).

     3.1.4  Measurement of Odorants

           The  odor problems inherent in  the kraft process were well
recognized by  its developers in 1891,  and  technical papers evaluating
the problem began to appear around 1900  (26).  Quantitative measurements
of the amounts of me reaptan, methyl sulfide, and dimethyl disulfide emitted
per ton of pulp were published as early as 1922 (28) , and the picture as
to what compounds are responsible has not  changed since.

           The early measurements of the amounts of these compounds were
made by conventional chemical means (29),  usually colorimetric reactions
for H2S and  the  formation  of mercury salts from mercaptans  (mercaptan  *
"mercurium cap tans ").   The development of  black liquor  oxidation in
Scandinavian mills, shortly after World War II  (30) , was a process  for  im-
proving the  economics  of kraft pulping by  reducing the  emissions  of sulfur
 odorants.  It coincided with the development of the bromine coulometer as
a rapid and  convenient  automatic instrument for measuring total  reduced  sul-
fur (18).  This gave a  strong incentive  for the production of a  commercial
instrument to operate  in the normal range  of uncontrolled emissions, up
to about 1000 ppm.

-------
Table 2
Average and Range of Odorous Compounds
in Various Kraft Mill Emissions
Digester
Blow Gases
Odorant
Hydrogen Sulfide

Methyl Mercaptan

Dimethvl Sulfide
Ditnethvl Disulfide
Acetone
Methanol
Alpha-pinene
so2
Gas Flow
during tests
Av.
80

4400

5100
1100
350
1100
1350

170

Range
10-200

1300-8700

2300-7000
600-2200
100-600
500-2400
950-2200

CFM at 60 °F

Multiple
(PPM by Volume)
Effect Contact
Evaporators
Av.
9000

100 ,000

4000
5000
—
110
960

10 CFM

Range
3000-
30 ,000
50,000-
300,000
500-6000
—
—
50-160
300-1800

at 60°F

Evaporator
Av.
<10

45

18
35
60
1000
30

6,500
60
Range

-------
                                  - 21 -
          The Barton Titrator has been used widely as a process tool,
to follow the course of plant design and operating changes made to re-
duce the amount of sulfur lost as emissions.  Initial models were
improved to provide multiple ranges in a single instrument, from ambient
through stack emissions (16) .   The most serious problems encountered in
the field have been related to difficulties in sampling the gas stream
which is hot, wet, and dirty.   Methods to handle this difficulty have
been developed by many workers (_7 ) , but the problem has not been completely
solved.

          Apart from sampling problems, one disadvantage of the Barton has
been its inability to distinguish directly between any of the major
odorants present.  This directed attention to gas chroma tography and
physical sensors  to go wit1  it, to avoid the use of tedious and time-
consuming chemical filtrati -n systems  for preliminary separation.


     3.1.5  Requirements for Monitoring

          Experience has shown that the complete analysis for individual
compounds and the changes in their concentration as a result of process
controls is primarily a problem for laboratory research.  Once these
determinations have been made and suitably confirmed, there ^ ™ *£*-f
taae in making a  complete analysis for routine monitoring.  The number or
suth compounds which need to be measured is limited to a few key components,
whose  changes are enough to indicate satisfactory or unsatisfactory  opera-
tions  in the system or source concerned.

          Present indications are  that the  distinction between H^S,  total
reduced  S and  So  gives  useful  data  for the control of kraft mill emissions,
and that  this  may be  enough information for routine monitoring.

3.2  Petroleum Refining/Petrochemical Operations

     3.2.1  Sulfur Odorants Present

          Hydrogen sulfide and mercaptans are the major sulfur-containing
odorants in oil refinery emissions.  Some of the I^S comes from crude oil,
and more of it is formed from other sulfur compounds during the refining
process.  Alkyl mercaptans  are present in crude oil in a wide range  of
molecular weights, with relatively larger amounts of the lighter compounds
which have more noticeable  odors ,
           The alkyl mercaptans are  readily  decomposed  to  split  out
by  thermal cracking or  catalytic  treatment.   This  type of processing  is
usually preferred  in modern  oil refining  to the  older  processes of
chemically or catalytically  converting  the  mercaptans  to  the corresponding
dialkyl sulfides or disulfides, which are less odorous.   Sweetening pro-
cesses which double the molecular weight  of the  original  sulfide are  still
used  in the deodorization of fuel oils, but they are no longer  preferred
methods in the  treatment  of  lighter distillates, partly because they  de-
crease the yield of product  in the  original boiling racge.

-------
                                    - 22 -
             Aromatic mercaptans such as phenyl mercaptan are also present.
   They are mostly formed by catalytic processes such as catalytic cracking,
   where they may result from the ring-splitting of polycyclic S-containing
   compounds such as benzothiophene.  The amount of these aromatic mercaptans
   produced is directly related to the ring sulfur content of the crude.

        3.2.2   Sources of Einisjsions

            Hydrogen sulfide  comes from three major sources in a modern
   refinery.   A schematic processing plan is shown in Figure 4.   I^S in the
  original  crude oil comes  overhead with the wet  gas from primary crude
  distillation.  Minor  amounts  are also released  into the air space of
  crude storage  tanks  (not  shown), which can be vented to a gas  collection
  system to avoid odorous emissions.  The second  major source is the dilute
  H2S-containing wet gas  produced  by a  variety of thermal treatments which
  include catalytic  cracking,  catalytic reforming,  and thermal  or fluid bed
  coking.   The wet gas  from all of these sources,  rich in light  hydrocarbons,
  is combined  and sent  to a gas plant for recovery.   The gas in  this plant
  is scrubbed with a suitable  solution,  such as an  ethanolamine, to remove
  H2S and the  low molecular weight mercaptans present.   The desulfurized gas
  and light hydrocarbons  produced  are available as  finished products or they
  may be used  as fuel and as raw materials  for further refining  or petro-
  chemical processes.   The  third major  source is  a  rich stream of H2S in
  hydrogen, ranging  from  about  15  to 70  mol %, produced by  treating middle
  distillate  fractions  with a  limited amount of hydrogen for sulfur removal (31)
                                   Figure 4
                                                                    ^Fu«l Got
                                                                           Gotolmi
Crudt Oil
        Crudi
           R«tiduwm

-------
                                   - 23 -
          A listing of points recognized as potential sources of specific
emissions in the refinery is given in Table 4  (32).


                                  Table  4

                          Potential  Sources of
                       Specific Malodorous  Emissions
      Emission             	Refinery  Sources	

  Hydrogen Sulfide         Untreated  gas  stream leaks,  vapor  from crude
                           oil  and  raw distillates,  process condensate
                           sewers;  traces  in  treated gas streams.

  Mercaptans                Cracking units, caustic regeneration units,
                           some asphalt plants.
  Organic Sulfides         Movement and storage of the  acid solutions
                           used in  scrubbing  organic sulfides and mixed
                           sulfur-nitrogen bases,  if they  are present,
                           from straight  run  or cracked distillates or
                           lubricating oil fractions.
  Sulfur Dioxide           Combustion of  fuels containing  sulfur, flares,
                           catalytic  cracking unit regenerators, treating
                           units,  decoking operations.

      3.2.3 Petrochemical Operations

          Petrochemical operations are based mainly on  the chemical reactions
of olefins, aromatics, or other light  hydrocarbons derived from petroleum.
These reactions are frequently  sensitive  to very small  amounts of sulfur
compounds  in  the  feed stocks.   The  processes  used  to desulfurize  these
chemical  intermediates are essentially  those  shown in Figure  2,  involving
catalytic or  thermal  cracking,  catalytic  reforming,  further hydrogen
treatment if  required, and an amine extraction of  the product to  remove
the H2S  and mercaptans present.  The  odorants present and  the measurement
problems  which  they present  are basically  the same.

      3.2.4 Means of  Sulfur  Disposal

           The H2S from hydrotreating  and  the  H2S/RSH concentrate from the
gas plant are too rich in H2S to be vented to the  atmosphere  as  such.  They
may be  burned to  S02, which  gets rid  of the toxicity problem  but creates
an equal amount of  S02 which is frequently undesirable. The  preferred
method  which  is becoming standard practice is to send all  of  this H2S  to
 a Glaus  process  sulfur plant.  A third of the ELS  is burned to SC>2, and
the two gases are combined at a slightly  elevated  temperature to produce
elemental sulfur:

                      2HS +  S0   	>•  2H0 + 3 S°

-------
                                 - 24 -
 The small amounts of alkyl RSH present with this l^S are ordinarily
 consumed in the sulfur plant.  If they are present in unusually large
 amounts, however, they may be recovered from the amine extract in the
 gas plant by a simple distillation and re-refined to recover their hydro-
 carbon content.

           The vent gas from the sulfur plant contains chiefly combustion
 gases, S02, carbonyl sulfide, carbon disulfide, and trace amounts of un-
 converted organic sulfides.   A typical composition  is shown in  Table 3.
 This stream was originally vented to the atmosphere,  after suitable dilu-
 tion.   It is usually incinerated first, which converts almost all of the
 remaining sulfides present to S02 and dilutes the whole with combustion
 air.


                                Table  3
                     Typical Tail Gas to and from
                       Sulfur Plant Incinerator
                    From Glaus Plant             From Plant
                 to Incinerator, Vol. %     Incinerator, Vol. %

         S02             3.0                        1.2

         COS             0.25                       0.008

         CSj             0.25                      <;10 ppm

         H2S           200-1000 ppm                <10 ppm

         CO              2.0                        0.6

         C02            13.0                        3.2



           The stack gas  from the sulfur plant incinerator and the vent
 gas  from the sulfur plant,  if any,  are thus the chief  points of source
 emissions  in a modern refinery.   This system provides  sulfur disposal for
 every collectable  sulfur-containing gas in the refinery.   A variety of
 means may be employed to dispose of collectable odorants  outside of the
 sulfur plant,  usually by burning or  air  oxidation  (33).

           Hydrogen sulfide  and mercaptans may escape  from process steam
 condensates,  drain  liquids  from  the  various processes, barometric condenser
pumps, sour volatile product tankage, and spent caustic solutions  from
 treating operations.  The malodorous  components in steam  condensates  can
be removed by  stripping with air, flue gas or  steam, and  the malodorous
gases burned in a furnace.  Drain liquids can be routed to a closed

-------
                                  - 25 -
storage vessel and returned to the process.  Barometric condensers are re-
placed by the more modern surface condensers; the non-condensables are
burned in process heaters or in a separate incinerator.  Spent caustic can
be  degasified, neutralized with  flue  gas  or  stripped before  disposal»
Sulfides can also be removed from sour process water and spent caustic
solutions by air oxidation to thiosulfates and sulfates.

          Another I^S problem may be caused when suddenly large quantities
of hydrogen sulfide must be burned in the flare.  It is difficult to burn
these large quantities completely to SOo.  The problem must be overcome by
redressing the upset causing the increased l^S production, usually by re-
ducing the charge to the hydro desulfurizers.  This situation can be
minimized by suitably sizing the CLaus sulfur plant  to accomodate expected
overloads.  Concentrated organic sulfides are rendered harmless by conver-
sion to H2S in a hydro desulfurizer unit, or by combustion in a process
heater.

-------
                                     - 26 -
      3.2.5  Measurement of Sulfur Odorants

          Analytical methods and procedures using wet chemistry for the
 determination of sulfur-containing streams in the refinery have been highly
 standardized.  Referee methods formalized as ASTM Procedures by the American
 Society  for Testing Materials have been adopted throughout the industry,
 and  continuously annotated or revised to include the results of further
 experience and comparative testing between laboratories.  The commercial
 use  of sulfur specifications as a measure of product quality and relative
 market value between competing manufacturers has given the strongest
 incentive to the development of reliable methods.

          The primary emphasis in these methods has been on the determination
 of sulfur in the products.  The analytical range of interest for liquid
 fuels was originally in the percent range with stated limits of accuracy
 of the order  of +  0.1  for the Lamp Sulfur  Method, which has been in use
 for  over fifty years.  Sensitivity has been continually improved, for the
 analysis of lighter fuels and gaseous hydrocarbon products ,anrl for the up-
 grading  of product quality to lower levels of sulfur content.

          The old Lamp Sulfur Method ASTM D-90 oxidizes the sulfur in a
 liquid hydrocarbon sample to sulfur dioxide in a wick-fed flame, and
 gravimetrically determines its amount.  The same principle of oxidation
 to S02 and measurement as such is applied to light liquid hydrocarbons and
 gases by burning in a pressure-fed flame (Wickbold, ASTM D-2785).  Higher
 sensitivity for very low  sulfur contents can be achieved by turbidimetric
 measurements after further oxidation and precipitation as BaSO^.  The
 general  approach here is  to oxidize all sulfur compounds present to S02,
 and measure this as an indication of sulfur content.  When the material
 being analyzed is a hydrocarbon liquid or gas, there is no problem of
 interference from SC>2 in  the original sample.

          Analyses for S02 as such have been developed primarily for the
 monitoring of combustion  stacks.   The prior art and evaluation of instruments
 for this purpose have been covered elsewhere and are not the object  of the
 present  review.   Particularly useful references include discussions by Palmer
 and Rodes (34),  Tokiwa (35),  and the current EPA Project EHSD 71-23.

          The oxidation of all S compounds to SC>2 as a preliminary step to
 instrumental measurement was  recommended by Thomas (36) using a conductimeteric
method,  titrating the acidity developed after absorbing the 862 in   dilute
hydrogep peroxide.   The Thomas Autometer was used in industry but did not
 widely displace the standard wet chemical methods, partly because of
problems in instrument maintenance.   Preliminary oxidation to SC>2 has also
been combined with the Dohrmann Microcoulometer, as a research tool for the
 measurement of very small samples.  Analyses by this procedure have been
 made  in  the field, but it has not yet been developed sufficiently for
 routine  use.

-------
                                    - 27 -
          The comparable principle of converting all S compounds present
 to  H2$ by catalytic hydrogenolysis and measurement as H2S has also been
 applied to refinery products.  This can be accomplished using a lead
 acetate tape recorder, as commercialized by Houston Atlas,  It can also
be done using the silver reductive micro coulometric cell offered for labora-
tory use  by  Dohrmann ,  based on developmental  work by  Burchf ield
           The use of instrumental methods for sulfur determinations has
 been  slow  to displace wet  chemistry  in refinery product analyses.  This
 is partly  because  the  standard wet  chemical  methods  have  been  so highly
 developed,  in a market situation where the direct competition between
 manufacturers on product quality allows no leeway for the adoption of  less
 reliable or accurate procedures.  Another factor has been the problem  of
 instrument  maintenance, which  is particularly important when there are
 relatively  few measurements over a period of time.   It must also be
 recognized that once the basic correlations  between  product quality and
 operating  conditions of temperature, pressure,  flow  rate  and feed com-
 position have been  established for a given process,  by research  measure-
 ments and  pilot studies, very  few complete analyses  are required on any
 plant stream except the feed and final product.  Research type measure-
 ments may  be made once a year  or once every  few months, for a plant test
 to confirm previous correlations, but this amount of use  does not justify
 automatic  instrumentation.

          Instrumentation  for emissions  control is  a new emphasis, and
automatic instrumentation  for this  purpose is just  now being developed.
An example is the exact control of  the Glaus process to give minimum sulfur
emissions in the product gas.   A general discussion of this system is
given in EPA pulbication AP-42, "Control Techniques for Sulfur Oxide
Pollutants."  The critical point for analysis is the affect of variations
in t^S to S02 ratio on the maximum theoretical conversion to S in a given
plant.  The effect  is illustrated in Figure  5.  Optimum conversion is
obtained at a mol ratio of exactly 2.0 and if this deviates by as much as
2.1 to 1.9, the maximum conversion obtainable decreases to an efficiency
of 96.8 to  98.3% conversion.   This corresponds  to an analytical  accuracy
of +  5% in  the reported value.  Monitors developed for this measurement
 include a  non-dispersive UV, sensitive to both  S02 and H2S.  While this
 difference  in conversion level may have relatively little effect on the
 tonnage of  sulfur produced, it is critical for  the amount of unconverted
material emitted from the  plant.

-------
                                      _ 28 -
                                    Figure 5

               Variation of I^S to S02 Ratio with Conversion and
           Maximum Theoretical Conversion Possible at Specified Ratio
100.0
                                CONVERSION, %
         3.2.6  Instrumentation for Monitoring

              The  sulfur  odorants produced and  removed  in oil  refining  come
    directly from the feed stock to the refinery.  Unlike the Kraft pulping
    process which seeks to conserve sulfur, the I^S and mercaptans produced
    in oil refining must be disposed of in production quantities.  They are
    handled in the refinery as process streams at a relatively high concen-
    tration, in the percent range or higher, and in systems which are com-
    pletely closed because of the toxic hazards involved at these concentra-
    tions.

              Odorant controls operate by converting the sulfur compounds to
    odorless or less odorous forms, primarily elemental sulfur or S02,  and to
    a limited extent to heavier sulfur compounds which can  be tolerated in some
    other petroleum product.

-------
                                   - 29 -
          Monitoring requirements are of two types:  emission measurements
on the principal stack gases, and ambient measurements to determine the
cumulative effects of miscellaneous small leaks from odorous process streams.
Emission measurements are of most interest for H^S, since most other sulfides
are now converted to H2? as an intermediate step in their final disposal.
Mercaptan emissions are more likely from diffuse sources, monitored by ambient
measurements, than from the stack gases which are  the major emission points.

-------
                                    - 30 -
 3.3  Animal Waste Rendering Plants

      3.3.1  Odorant Emissions

           Odor  emissions  from  animal waste  rendering (AWR)  plants  are
 derived from the processing of inedible  raw materials which are  unsuitable
 for other use.   The odors associated with these  operations  have  long been
 recognized as a severe  type of air  pollution and one of  the more frequent
 causes  of public complaint,  but they have been most  difficult  to measure
 or quantify.  Recent research  studies  combining  odor panels with physical
 separations and the chemical or spectroscopic identification of  constituents
 have been able  to identify specific malodorous compounds  and types of  com-
 pounds  as present in AWR  emissions, but  the data obtained by this  technique
 are only qualitative or semi-quantitative at best.   Instruments  to measure
 the compounds identified  are only in the  conceptual  stage or just  now  being
 developed.   Parts of this problem in odor measurement and instrumentation
 therefor  have  thus been  solved, but no  total system for  this  purpose  has
 been set up and tested.

           The odorants  released in  animal waste  rendering are  normal prod-
 ucts of animal  metabolism and  the concentrations of  these materials  de-
 tectable by smell are extremely low, of  the order of parts  per billion or
 even less.   This has been beyond the instrumental range  and in some  cases
 it is still marginal at best.   Concentrations at the source will be
 correspondingly higher, in the range of parts per million.

           The relationship between  source concentrations  and those de-
 tectable at a distance  from  the plant  are by no  means linear,  however, for
 different compound types.  The reasons for  this  are  only  partly  known.
 Aldehydes,  for  example, are  relatively easily peroxidized in air to  form
 the  physiologically reactive peracids, and  the concentration of  these
 might remain  constant or  even  increase with time and distance  from the
 source.   It is  entirely possible, although  not yet fully  established,  that
 it is this  or some similar mechanism which  explains  why aldehydes  are  con-
 sidered  among the  most  persistent odorous compounds  emitted.

      3.3.2   The Sources of Emission

           The animal rendering process consists  of reducing animal tissue
 from Inedible meat,  bone  scrap,  fish and  poultry wastes to  solids, fats
 and water.  The products  produced are tallow,  grease, fertilizer and animal
 feed.  Whole b^lood may  also  be  processed  to give solid blood meal which
 is also valuable as  a fertilizer and in the manufacture of  glue.

          A typical  continuous  rendering process such as  that  shown in
 Figure & consists  of a  receiving area for the  animal wastes, which are con-
veyed to  a  cooker where water  is removed, followed by additional cooking
 to remove tallow,, other liquids and gases,  drying, pressing  of the solids

-------
                                - 31 -
                                Figure 6

           Schematic Diagram of a  Continuous Rendering Process

Receivinf Pit
Animal Pastes


I
1
LiqUi
>
Cookers *'
i "•
solids &
liquids
Parcolacoi; g

| va,
AA *
Solids


1
Solids
si/
'tallows Expo Her
(Hydraulic press)
i
f
Steam-hen ted 	 HtiuitH
Separation
Tank
e
& 	 . Contact i spray lv atm03phct
mnm ' condensers f\ scrubber! 7
' 1 ;]
ascs
pors
gases
	 ^V 	 — ^T1-, i
\—
Liquids to
JL^ lagoon
vapors
solids x j Hananer- 	 v
"~^» mill 7


FinishcJ Product
Paclia^ed for sale

tallows
  i
 1'inished Product

  Tallov; storage
      Tank

-------
                                 - 32 -
 and grinding of the solids to form a finished product.   Gases and vapor
 from the cooker are condensed in a contact condenser.   If an air dryer
 is used, the vented drying air is passed first through a spray scrubber
 before being vented to the atmosphere.   The odorous effluent from such a
 plant will consist mainly of vented gases and odoriferous liquids.

           There are two main sources of odorants in animal rendering
 plants and these are discussed briefly  below:

       1.  Raw materials in transit and  in storage:   This source can emit
           powerful odorants due to the  natural and  bacterial decomposi-
           tion of tissue beginning with the death of the animal.   Such
           odors increase rapidly in intensity with  time.  To minimize
           such odors,  the handling time of the animal  material is usually
           kept to a minimum; however, it is sometimes  difficult to  avoid
           putrefaction and enzymatic decomposition  of  these materials,
           A variety of odorant materials may be formed  in these biologi-
           cal processes including hydrogen sulfide, volatile amines
           such as methyl and ethyl amines, putrescine,  cadaverine,  and
           skatole.

       2.  Odors arising from processing equipment:   The equipment in-
           volved consists mostly of boilers,  dryers, blood spray  dryers,
           crackling bins, rendering kettles,  mixers, holding tanks,
           storage tanks,  incinerators,  and catch basins.  The cooking
           process is the main part of the rendering operation in  which the
           proteinaceous material is decomposed at a temperature exceeding
           200°F.   The  effluent from the cooking process contains  mostly
           steam driven from the animal  tissue,  along with considerable
           amount of many  odorous compounds such as  amines,  fatty  acids,
           aldehydes, etc.   The gases from the cooker are usually  condensed
           (95% condensable)  and the remaining odorous uncondensable  gases
           are  vented to the  atmosphere  or passed through an afterburner.
           In  some plants,  an air dryer"is used to further concentrate the
           solids, and  the large volumes  of air used and emitted can  then
           add  to the odor problem.   Typical  levels  of air are 10,000  CFM
           per  dryer.   Attempts  to  clean  up the  exhaust  air  by condensation
           or water  scrubbing do not appreciably reduce  the  amount of
           odorants  contained in this air.  Most plants  do not incinerate
           the  dryer gas  effluent.

           Odor  emissions  from typical rendering plants  have been  identified
as to  their sources  and  types  of compounds.   Odor concentrations  and
emission rates  have been  determined by  several  workers  (37^381, and are sum-?
marized  in Table  4.  It is evident  from  these  data  that there are wide
variations in odorants produced from similar  equipment.   It may be noted
there  are  also  some  differences  in  the  types  of  compounds  emitted, e.g.,
the mercaptans  and sulfides  are  found mainly in  the processing  of feathers.

-------
                                                 Table 4
                        Odorants and Concentrations Emitted from Rendering Plants
 Operations
Dry Cooking
Blood Cooking
Feather Cooking
  & Drier
  Source of Air Pollution

  Noncondensibles in vapor
  Vapor leaks from cooker
  Dumping of hot fats
  Noncondensibles in vapor
  Vapor leaks from cooker

  Product dumping
  Product drying
  Loading docks
Compound
Classes

Amines
AIdehydes
Fats
Fatty Acids

Same as
 above

Aldehydes
Amines
Me reaptans
Sulfides
Fatty Acids
Fats
Odor Concentration
 Odor Units/SCF^ '
            Typical
  Range     Average

 5000 to    50,000
 500,000
10,000 to  100,000
1 million
600 to
25 ,-000
2,000
                      Exhaust      Odor Erais-
                      Products,    sion Rate
                       SCF/Ton   Odor Unit/Ton
                       of Feed      of Feed
                       20,000
           38,000
                        77,000
                       1000 x
                                    3800 x 10
153 x 10'
(1)  Odor Unit
     Odor Unit -
No. of volumes of odor free air required to dilute one volume of odorous air to the
minimum identifiable odor (MIO) level.

  Odor Free Air Volume
         MIO

-------
                                - 34 -
This is a somewhat oversimplified tabulation of the compounds released from
rendering plants,and actually the odorant materials are an extremely com-
plex mixture of organic compounds containing both odorous and nonodorous
components; however, many of the odorous compounds are obnoxious even at
very low concentrations.

          Many efforts have been made over the years to identify the
odorous components in animal tissue as well as those emitted from render-
ing plants.  Ronald (39) identified H2S, NH3, di- and triethylamine, and C02
in the condensate from dry rendering of flesh by wet chemical methods prior
to the advent of GC techniques.  Chromatographic techniques were used in
1957 by Wood and Bender (40) to identify components extracted from ox muscle.
These workers separated the protein from the extract and identified 27
different materials by paper chromatography.  Most recent work has con-
centrated on the identification of components by a variety of gas chroma-
tographic techniques.  Studies at the IIT Research Institute (60) have shown
that decomposition of meat samples in the laboratory closely approximates the
odorant materials emitted from rendering plants.  A number of the organic
components released have been identified by GC and mass spectroscopy and
characterized by odor, in addition to determining their odor threshold
level.  The Kovats Index System was used to characterize their GC behavior.
Table 5 summarizes the major components identified and indicates the odor
threshold level and other pertinent information.  The materials identified
in Table 5 are the more volatile emissions from rendering plants and these
are believed to be the most important in determining odor..  A number of
other less volatile materials have been separated and many of them have
been identified by similar analyses.  These latter compounds consist mainly
of higher alcohols, aldehydes and acids; however, it is anticipated that
these heavier materials will make a smaller intrinsic contribution to odor
emissions, and that they would generally be trapped in the various con-
densers and scrubbing systems in the plant.

     3.3.3  Methods of Measurement

          As indicated in the above review, the volatile aldehydes and
organic acids appear to be major components in most AWR odorant emissions.
In plants where feathers, fish or decomposed materials are processed,
amines, mercaptans and I^S may make a significant contribution to odorant
emissions.  For these latter compounds, automated analytical techniques
developed for the Kraft Mill or petroleum industries may be used. However,
there are no instruments available for automated measurement of aldehydes
or organic acids and these need to be developed.

-------
                          Table 5


Characterization and Identification of Major Odor Compounds

  Commonly Found in Rendering Plant Emissions  (60)
Kovats
Index
889
925
969
989
1046
1078
Concentration Range
(Expressed in Peak Area %
from GC Analysis) Odor Note
0.34
0.22
0.01
1.19
0.06
1.41
- 23.60
- 36.62
- 2.21
- 55.19
- 2.32
- 46.00
Aldehydic
Rancid
Butter
Aldehydic
Butter
Aldehydic
Odor Threshold
Identity ng/cc
n-Butanal
3-Methyl Propanal
2,3 Butane dione
n-Pentanal
Probably 2,3-Pentanedione
n-Hexanal
2,000
--
•» w
3,000
~
—
                                                                                     u>
                                                                                     Ui

-------
                               - 36 -
           A wide variety of analytical research methods  have  been  used
 to determine the nature and composition of AWR odorant emissions.  At
 present, odor panels are the most widely used means  of measuring these
 objectionable odorants in the field.   Generally,  analytical techniques
 are applied to odorous materials after they have  been separated and
 concentrated from the ambient air.  Wet chemical  methods,  instrumental
 methods and combinations of these techniques have been used and are
 summarized below.

      3.3.3.1  Odor Panels

           While odor panels are used  mainly to determine whether  there is
 an objectionable odor, they may also  be used to determine qualitative and
 semi-quantitative concentrations of odorants in the  atmosphere.   Trained
 odor  panels can detect certain specific types of odorants and may  also
 determine threshhold levels of odor with surprising  accuracy.  Examples of
 the usefulness of odor panels in this regard have been given  by Hemeon (42) .
 This  technique and others described by Sullivan (43)  are generally based
 on a  dilution technique in which the  original sample is  diluted until it
 can just be detected by the human evaluator.  This vapor dilution  technique
 can be  a static method, a continuous  method,or a volatilization technique.
 A  number of instruments  have  been developed  to assist in using these techniques
 for both laboratory  and field evaluation.  The odor panel approach is excluded *
 by definition in  the present  review.

      3.3.3.2  Wet  Chemical  Methods
          Aldehydes in air may be determined directly by a number of
colorimetric techniques and titration methods which are. suitable for
automated wet chemistry.  These are generally for rather specific and
usually low molecular weight aldehydes, and with a number of  these
methods, ketones also interfere.   Methods which have been used  are given
below:

      1.  MBTH  (3-methyl-2-benzothiazolonehydrazone)   solution  in
          sulfuric acid produces a blue  cationlc  dye  that  can be
          measured at 628 n}ti(44).  The sensitivity  of this method
          is 2 ppb.

      2.  A modified Schiff procedure  using  rosaniline and dichloro
          sulfitomercurate may be used for  low molecular weight
          aldehydes with a reported sensitivity of  .01 ppm  (48)

      3.  NaHSO^ is used as a standard method  (48)  for the determina-
          tion of low molecular weight aldehydes.  The bisulfite
          solution is adjusted to a pH of 9.6 + 1 with Na2S03 or
          acetic acid as required which decomposes the bisulfite  ion
          equivalent to the aldehydes present in  the  sample.  The
          liberated bisulfite is then titrated with standard iodine.
          The sensitivity of this method is about 1 ppm, and interfer-
          ing compounds are methyl ketones.

-------
                            - 37 -
     4.  Formaldehyde may be determined using chromotropic acid or
         the modified Schiff method and both give a sensitivity
         of about .01 ppm.  Polarographic methods have also been
         evaluated but need further study
     5.  Acrolein forms a specific color compound with 4-hexyl-
         resorcinol in ethanoltrichloroacetic acid solution to
         yield a blue color product which can be measured spectro-
         photometrically at 605 m/j.  The sensitivity of this method
         is about .005 ppm (44) .

         Several methods have been developed for the direct determination
of organic acids which are .   c generally applicable to acetic acid and
other specific low molecular weight organic acids.  In one of these methods,
the acid is collected in a dilute solution of NaOH (57«) in an ice bath  (48) .
The acid is then liberated with concentrated H2S04 to a pH of 2 to 3 and
extracted by refluxing with ether for 8 hours.  The resultant extract is
then titrated with .IN NaOH.  A second method is a color indicator method
based on Fleisher methyl purple which has been developed by Miller and
co-workers (47) .  This is a sensitive method of detecting acetic acids
using a collection  medium  composed of 1:1 glycerol:water solution
containing a small amount of Dow Corning Antifoam A. This solution, con-
taining the indicator, is adjusted to a pH of about 6 and a color change
from Kelly green to purple occurs over a pH of about  1.  Thus, by measur-
ing the volume of air required to change this color,  it is possible to
calculate the concentration of acetic acid in air directly.

           A method  for  the determination  of  ammonia and  ammonium com-
 pounds has been  developed by  California workers  (fr.U)  for use where their
 concentration is generally very  low.   A series  of  impingers  containing
 water  and 5%  HC1 solution cooled with  an  ice  bath  is  used  for  collections
 of the nitrogen  compounds.   The  solutions  are then analyzed  by either the
 Kjeldahl method, or  for  more sensitive  determinations, by  the  Nessler  color
 test method.

     3.3.3.3   Instrument  Methods

           A few  instrumental methods have been used for the direct
 analysis of odorous  materials  including AWR emissions, but a preliminary
 separation is usually required.   Gas chromatographic  techniques have
 proven to be  well suited for  separation prior to  the  analysis  of  complex
 mixed  odorants.   Such separations  have been  combined  with various types
 of detection equipment  to differentiate and  identify  in detail the major
 components of effluent  gases.   A brief review of methods used  and their
 suitability for analysis of rendering  plant  emissions follows:

-------
                                - 38 -
      3.3.3.4  Spectroscopy

           UV spectrophotometers are useful mainly for analysis  of aromatic
 compounds and since these are not indicated to be of major importance  in
 rendering plant emissions, these techniques may not be appropriate for
 direct analysis.   However, UV may be useful  for the  determination of
 aromatic derivatives of  odorant compounds, such as the DNPH derivatives
 of aldehydes, either as  a group or as  individual  compounds.


            Infrared spectrometers have been used  for both qualitative
 and quantitative analysis of odorous emissions.   IR  techniques  are often
 used to complement the mass spectrometer,  since the  latter can  differ-
 entiate accurately between homologous  series and  structural isomers, but
 it cannot distinguish geometrical isomers  and identify  functional groups,
 which is often required.  This latter  capability  is  readily attainable
 with IR spectrometers .  Trief f and co-workers (58) tried  to determine
 aldehyde concentrations  and other organics  from rendering plant emissions
 directly using the IR Spectrophotometer.   However, they found that the
 concentration of  gases was too low to  be detected directly by this in-
 strument .  Many colorimetric methods are available to identify  aldehydes
 as a group,  and also specific aldehydes and  these may be  separately
 determined using IR detectors.  A good  summary of  the capability of
 such methods is shown in Table 6 from the work of  Stahl
          The mass spectrometer is a powerful analytical tool which can
give a positive identification of many different types of compounds.
However, this instrument cannot analyze complex mixtures.  Thus, odorous
materials, such as emissions from rendering plants, must be separated into
fairly simple mixtures of components before being subjected to analysis
on a mass spectrometer.  This instrument is also very big and expensive
and is thus most suitable for research studies.  An entirely new principle
which may be useful with polar compounds including many odorants is "drift
mass spectroscopy."  This is a laboratory development in which, an ion (H+)
is reacted with the molecules and a detector then measures the shift of
the protonated species in an electric field.

          In view of the requirements of the mass spectrometer,  it  is best
used in combination with some separation technique such as  a  chemical separa-
tion by compound type or a GC separation technique.

     3.3.3.5  Gas Chromatographv
     t
          Many workers have evaluated the use of GC  columns with and without
additional detectors for the analysis of odorant materials  In air (45 . 46) .
Several detailed studies have also been made on odorants  from rendering
plants.  These techniques have been outlined briefly below with  comments
on their limitations.

-------
                           - 39 -
                          Table 6
SUMMARY OF QUALITATIVE COLORIMETR1C DETERMINATION METHODS
1 	 • 	 	 	
Reaqent
Indole
Fuchsin (Schiff method)
4-Phenylazo-phenyl-
hydrazine sulfonic acid
2-Hydraz ino-benzothiazole
4- p-nitrobenzenediazoniunt
fluoborate
2-Hydraz ino-benzothiazole
(HBT)
3-Methyl-2-benzothiazolone
hydra zo n e ( MBTH )
(J-acid) 6-amino-l-
naphthol-3-sulfonic acid
Color
Orange to red
Violet to blue
Red to blue
Blue to green
Blue
Blue
Blue
Limits of Identification
in Microeratns
Aldehydes
~0.05-1
~1-30
0.2-0.4
0.2-200
0.01-3.0
0.1-80
0.01-11
Formaldehyde
0.2
1
0.25
0.2
0.01
0.1
0.03
Aero le in


0.2
0.3




-------
                                 -  40  -
           The  advantages of  gas chromatography for the detection of
 odorants  In  the  atmosphere was recognized in 1958 (49), and subsequent
 experiments  by Williams identified  fifty-eight atmospheric pollutants
 by  this procedure  (50).  His list of  compounds identified includes most
 of  those  found in  AWR emissions.  Williams used a short stainless steel
 column containing  Chromosorb P partially deactivated with di-ni-butyl-
 phthalate cooled to -80°C to collect  the samples.  The gas was then
 desorbed  from  the  column, dried, and  injected into a Perkin-Elmer Model
 154-D gas chromatograph fitted with flame ionization and electron capture
 detectors.   Williams  also tested a specially treated molecular sieve
 column heated  to 150°C, or one containing sulfuric acid on an inert sub-
 strate, as a means of  further separating air pollutants before analysis
 on  the GC columns.  Using various combinations of drying agents, sulfuric
 acid, and molecular sieves,  he was able to separate and identify hydro-
 carbons,  aldehydes, alcohols, organic acids, sulfides, disulfides, organic
 nitro compounds, nitriles, thiols, ketones, and halogenated compounds in
 the atmospheric  samples tested.

          Low molecular weight aldehydes have been studied in industrial
 effluents by Levaggi and Feldstein (51) and in automotive exhaust by
 Prater (52).  Levaggi  employed a Varian 1200 gas chromatography with a
 hydrogen  flame detector and  a stainless steel column packed with 15%
 Carbowax  20M on  Chromosorb,  followed by a 5" x 1/8" stainless steel column
 containing UCON-dinonylphthalate on fire brick.  This equipment enabled
 them to detect C^ and higher aldehydes.  Prater suggested the use of GC
 equipment with a dual  flame  ionization detector and an OD-17 column for
 analysis  of  the  aldehydes collected in 2,4-DNPH.   Using this equipment,
he was able  to detect  aldehydes from formaldehyde up to paratolualdehyde
 in concentrations ranging from .03 to 5 ppm.

          Several workers have used dual column GC techniques on air pollu-
 tants and found  they were not sensitive enough to identify all the odorant
 materials which  could be detected by  human  sensors.   A study of air pollu-
 tion from animal wastes by  Burnett  (53)  u»ed a  Varian Aerograph Model
 1520B equipped with dual flame  ionization detectors.   He  was  able to
 identify  8 or  9  compounds with good certainty but  some odorous  materials
 detected  by  human  evaluators did not  show up as peaks on  the  GC plots.
Analysis  of  odorant materials in the  atmosphere by Okita  (54) identified
a number  of  amines, mercaptans and  organic  sulfides,  but  even with  the
use of the GC with flame ionization detectors, he found the  equipment  was
not sensitive  enough  and he  felt there was  a need  for electron  capture and
electroconductivity detectors.

          Dravnieks (55)  has shown  that routine GC techniques with to* i
 stage detectors  are sensitive enough  to detect nanogram (10~9 gm) amounts
 of individual  substances.  This amounts to 6 x 101  molecules for a typical
 compound  having  a  molecular  weight of 100.  Assuming that the limit of
 detection of odorous materials is in  the range of 10  molecules per cubic
 centimeter of  air, it  is only necessary to concentrate the odorant materials
 in  air by a  factor of  6,000.  Such a  concentration factor is readily obtain-
 able using organics to absorb vapors, when the temperature is changed from

-------
                                  - 41  -
ambient to about 200-250°C.  Dravnieks has identified discriminators
which can be applied in gas chromatography for polar and non-polar
separations in animal-source materials.  These include optically
active dipeptides for the dispersion of optical isoners, cobalt
stearate for the strong dispersion of primary and secondary amines by
complex formation, sorbitol or diglycerol for the hydrogen bonding of
polar groups, and Emulphor 0 for the dispersion of polar functional
groups or double bonds with respect to paraffinic hydrocarbons.

          The complexity of odorant emissions from rendering plants and
other sources has led researchers to methods of further discriminating
the peaks obtained using GC columns and various detection techniques.
One cf the most widely used research techniques for the comparative
indexing of unknowns is the Kovats Index Sys tern (56) .  This technique
combined with a split column GC provided with a so-called "sniff port"
has been used by Dravnieks and Trief f  (57, 5_8 respectively).  Using these
combined techniques it is possible to detect specific compounds as well
as odorant types and relate the peaks noted from the GC analysis to odor
panel response.

           The most  complete qualitative  analysis which has been conducted
 to date  on rendering plant emissions  combines GC techniques with mass
 spectrometry  (59).   This  involves either single or  double  column GC
 equipment  withThe  effluent then being put  through  a mass  spectrometer.
 Using these techniques,  Dravnieks,  Burgwald (59,  60)  and co-workers have
 been able  to  identify  a  number of compounds by  chemical type.   In  this
 work, it was  necessary to identify  knowns by similar techniques and relate
 them to  the unknown emissions  from  the rendering  plant.   An indication  or
 positive identification of 30  different  compounds was obtained.  Such
 techniques, involving  the mass spectrometer, are  excellent for detailed
 identification of components  but  are not practical  for control purposes.

           Use of simpler GC columns and the detection equipment discussed
 above shows promise for identifying specific odorant compounds or types from
 rendering plant emissions and other odorants in the atmosphere.  Such a
 type analysis may be satisfactory for monitoring purposes, where the odorant
 originates from a fairly uniform source.  It must be recognized, however,
 that in the general case a type measurement for total aldehydes, for
 example, or total organic sulfur compounds, does not necessarily relate
 to the concentrations of the individual compounds present as principal
 odorants.

-------
                                 - 42 -
              4.  SPECIFIC INSTRUMENTATION FOR S ODORANTS

          This section contains brief summaries of the characteristics
of specific types of instruments which are considered as applicable to
the field measurement of sulfur compounds (and other odorants), and
literature sources for additional information.  It is divided into three
sections;  1) preferred commercial instruments, which can measure several
(or all) types of sulfur odorants; 2) alternate instruments, commercially
available, which can measure at least H2? , the principal S odorant; and
3) potential instrumental approaches, which need to be developed further
before they can be properly evaluated.  The capabilities of the instruments
described can be summarized briefly as follows:
          4.1  Preferred Commerical     H2S + RSH ^and others)
          4.2  Commercial Alternate     I^S (and others)
          4.3  Potential                R/D required

          Information regarding the individual techniques discussed
in each of these sections is organized into a brief review of the operating
principle, commercially available equipment for this specific use, biblio-
graphic references (for operating principle, applications, data on uses or
comparisons, and specific instrument descriptions), stated advantages and
disadvantages, and conclusions or recommendations.  The statements of advantage
and disadvantages listed are taken from the literature, with some editing, for
the purpose of this interim report.  It must be recognized that such statements
are subject to change on evaluation, and not always mutually consistent.

          Pertinent extracts of the manufacturer's operating, maintenance
or application manuals are appended for the commercially available instruments
in Section 4.1, to include statements of operating performance.  The originals
of these manuals should be consulted for details.

          This section of the interim report is a guide to source material
to be used in Contract 68-02-0219, for the "Evaluation of Measurement Methods
and Instrumentation for Adorous Compounds in Stationary Sources."
It is intended as a survey and not as a definitive review.  The reader who is
interested in critical reviews of various portions of the background informatlo
available will find numerous publications of interest.  Good analytical reviews
which cover the literature in detail on specific topics in the measurement of
odorous (sulfur) compounds have been published in 1943 by Thomas (36) , in
1952 by Bialkowsky (146), in 1963 by Kenline (26), in 1967 by Walther (91) >
in 1968 by Blosser (7), Thoen (8), Adams (9), and Leithe (147), in 1970 by
Hendrickson (6), Tokiwa  (35), and Stevens  (73), and in 1971 by Dravnieks (59)
and Austin (62) .

-------
                                  - 43 -
noULOMETRIC TITRATION  (S02> H2S, Mercaptans, Halogens and Nitrogen Compounds)

Principle of Operation

          Four  different types of coulometric  (amperometric) units are
available namely bromine, iodine, silver and hydrogen cells.  While the
principles  are  similar in each case, the chemistry and methods of analysis
are  somewhat different for each type, as discussed below.  The hydrogen  cell
is not  widely used  in  the industries of concern,  and is not  discussed here
in detail.

Bromine Cells  (For  SC^, H^S, Mercaptans)

          A constant level of  free bromine  is  generated by electrolysis
from a  solution of  KBr.  Reaction with  SCv  decreases the  free halogen
present and is  detected by bromine sensors  which  signal the  electrolysis
current to  increase.   The additional current  required  to  maintain the
original bromine  level is proportional  to the  S02 concentration  according
 to  the  following  reactions:
           2Br" - 2e
 Bromine units respond to SO^, H«S, mercaptans (35) and all oxidizable
 organic sulfides.   If desired, EJS and other reduced sulfides can be
 removed by passing the inlet gas over a heated (120°C) silver gauge, leaving
 S02 to be measured alone, or S02 can be scrubbed out leaving only the sulfides.

 Iodine Cells (S02, HjS and Mercaptans)

           A constant level of free iodine is generated by electrolysis
 from a solution of KI or Nal.  Reaction with oxidizable sulfur compounds
 reduces the amount of free iodine present.  This reduction is detected by
 polarizing electrodes.  The titrant I-j" consumed is proportional to the
 sulfur content of the sample as shown by the following reactions.  All
 sulfur compounds in the sample are first oxidized in a combustion tube, and
 swept into the tit rat ion cell.
                                                        2H+
                                           2e
            Sodium  azide  (NaN3)  is generally added  to the  iodide  electrolyte
  solution  to minimize possible  interference from chlorine and  nitrogen con-
  taining compounds in the  sample stream.

-------
                               - 44 -
 Silver Cell  (H S,  Mercaptans  and Chlorine)

           The  concentration of  silver ions  is  kept  constant by electrolysis
 from a solution of silver  iodide.   Sulfide  or  chloride present in the
 sample reacts  with the  silver ions  which  are coulometrically replaced.  The
 total current  required  to  replace the silver ions is  a measure of the
 sulfur or chloride present in the injected  sample.  Since this cell reacts
 only with reduced  sulfur and/or chloride  the unit can be used to measure
 reduced sulfur alone.

           If oxidized compounds are present, total  sulfur can be determined
 by a preliminary pyrolysis in a reducing  atmosphere to convert all sulfur
 compounds to H~S.   In this mode the hydrogen used in  the reduction reaction
 is humidified,  to  reduce coke formation and the  formation of small amounts
 of cyanide from any nitrogen  compounds in the  sample.  To stabilize the
 cell reaction,  the electrolyte  is saturated with a  gas mixture of 10% NH-
 in nitrogen.

           This  cell can also  be operated  to measure reduced sulfur and
 chloride  independently.  To accomplish this a  titration endpoint or "bias"
 potential is chosen for S  determination such that the chlorides will not
 consume silver ions.  For  determination of  chloride a higher "bias" potential
 is used.   Under these conditions neither  material interferes with the other  in
 desired analysis.

 Commercial Manufacturers

           The Barton Titrator is a  continuous-flow bromine microcoulometer
 which can accept samples for  direct measurement  of lUS and TRS at emission
 level concentrations.   It  is  widely used  in the  pulp  and paper industry.
 The  Philip S0_  Monitor  can be used  for continuous measurements of H^S and
 RSH  (but  not other sulfides)  at  ambient concentrations.  Commercial Dohrmann
 microcoulometers are available  as either  bromine, iodine, or silver ion
 cells.  They are primarily research laboratory tools, for the measurement
 of small  discrete  samples  such  as the fractions  produced by gas chromatography.

           Suppliers of  such instruments are as follows:

 Bromine Cells

 ITT  Barton                                      Models 286 and 400
 Dohrmann                                         Models C-200 + 250B
 Philips Electronic                              Model PW 9700

 Iodine  Cells

Atlas Electric  Devices                          Model L 1202
 Beckman Inst. Co.                                Model 906
 Dohrmann                                         Cell T-300-P

 Silver  Cell

 Dohrmann                                       Cell T-400-S
 Hartmann + Braun A.G., Frankfurt-am-Main, GFR.   Picoflux

-------
                                  -  45  -
References

Barton Model 286

          1.  Principle of  Operation:  61, 62, 63, £, 64,  65,  35,  66,  6J7

          2.  Applications:  61, 62, 63, 8_, 64, 65, 68,  66,  67,  j>9_

          3.  Data:   6.L,  62_,  63., IB, 64_, 65_, 66_, £7, 69.

          4.  Specific Instrument Description:  61, 62,  63,  8^, 65, 66, 61_

Bromine Cells

          1.  Principle  of  Operation:  70_, 7^, 72_,  73>,  35_, 22_, 7^, 75, 66_,
                                        76, ZZ» 78^

          2.  Applications:  70, 71., 72_,  7^,  22_,  T±, J5, 66_, 76, 77^, J8, 79.

          3.  Data:   70,  71, 72_, 73_, 22_,  74.,  75_,  66,  76, 77., 7^

          4.  Specific Instrument Description:   70^,  71,  72_, 22_, J5_, 66^

Iodine  Cells

          1.  Principle of Operation:   71, 35. 75,  66,  1J.

          2.  Applications:  71. 75, 66,  Ii8

          3.   Specific Instrument Description:   71, 75, 66. _18

Silver  Cells

          1.  Principle of Operation:   J30

          2.   Applications:  80

           3.   Data:  80

           4.   Specific Instrxanent  Description:   80

-------
                                 - 46 -
Discussion

Advantages

Barton Model 286
          1.  Reliable with a minimum of maintenance (weekly).

          2.  Gives rapid, continuous, direct measurement for SQ^.

          3.  Sensitivity to SO- in the range of 0.05 to 1.0 ppm (64)
              ranges from + 0.05 to + 0.24 ppm respectively.

          4.  Sensitivity to ILS and mercaptans is 0.02 ppm (64_).

          5.  Well suited for long sample lines with monitor at a distance
              from the source.

Others (Dohrmann, Philips)

          1.  Good sensitivity to H2? (10 ppb) and SO, (50 ppb): (Dohrmann, 22)
              (2 ppfa) and(4 ppb): (Philips)

          2.  Rapid response time.

          3.  Relatively trouble free:  Philips has internal (automatic)
              7ero and calibration.

          4.  Accurate.

          5.  Interference from other gases in the ambient air generally minor.

          6.  Applicable to both liquid and gases (Dohrmann).

          7.  Suitable for determination of oxidized and reduced sulfur
              compounds and compounds containing chlorine and nitrogen at
              levels ranging from 0.1 ppm to several percent.

Disadvantages

Barton Model 286^

          1.  Minimum sensitivity 30 ppb (H S) (66).

          2.  When H_S cone,  in gas is above 100 ppm over extended periods,
              some elemental  sulfur is deposited and may cause a sluggish
              response (62).

          3.  Chlorine compounds  Interfere.
Others

          1.  Dohrmann requires daily calibration by skilled technicians,
              sensitive to temperature changes.

          2.  lodometric cells do not respond to organic sulfides or disulfides,

-------
          3.   Reducing compounds  such  as  0«,  unsaturated compounds  and NO
              interfere in the  iodometric cells,  however,  these  may be
              removed with appropriate absorbers  (35).

          4.   Philips has limited range above ambient  levels.

Conclusions

          The Barton titrator is  well  suited for  automatic monitoring of
S02, ^S, and total reduced sulfides over a fairly wide range  of concentra-
tions.  Interfering compounds can generally be eliminated.  Maintenance
problems which can be troublesome in the field are due largely to the dif-
ficulties involved in providing a suitable sample stream.

          Other coulometric titrators  generally are just as sensitive as the
Barton and the technique can be adapted to monitoring a variety of compounds
including sulfides, chloride, phosphorous and nitrogen compounds.  One of
the major drawbacks with the microcoulometer is the frequent  (daily)  calibra-
tion  required, and the high degree of technical skills needed to get optimum
results.

Recomroendat ions

          Evaluate the Barton  instrument at plant  sites for monitoring  of
refinery as well as kraft mill emissions.

          Evaluate the microcoulometer further as  a possible  referee  method.

-------
                                               - 48 -



                                            BARTON

                                       Model 400 Titrator

                  Flow Diagram
                   Titration Cell
                                      General Specifications
Sensitivity
  Hydrogen Sulfide	0.02 ppm
  Mercaptans	0.02 ppm
  Organic Sulfides	0.05 ppm
  Sulfur Dioxide	0.05 ppm
Range	0-1000 H2S ppm (7 ranges)
Model 400 Outputs ... 0-100 ua, 0-10 mv or 0-100 mv
Function Controls	On, Off, Total Sulfur, Total
               Minus SO2, and Automatic Cycling
Power	115 v z-c, 60 Hz, less than 45 watts
Power Connection	6-feet, 3-conductor cable,
                                  3-prong plug
Sample Flow Rate	200-600 ml per minu
                 (250 ml nominal recommende
Inlet Connections	V4-inch tubing fittir
Outlet Connections	V4- inch tubing fitth
  Titration Module
  Height	10 inch
  Width	9 inch
  Depth	22 inch
  Recording Controller Module
  Height	,	141/2 inch
  w'dth	87/e inct-
  Depth	13y8 inct.

-------
            DOHRMANN  MICROCOULOMETER
                                  JT   ^
          Figure 3.   Comparison of titration celts
        Oxidativc cell
(T-300-P) Iodine cell
Electrolyte:
  0.05% potassium iodide
  0.50% acetic odd
  0.06% sodium azide
Sensor:  Pt°
Reference:
  Ptc—Triiodide Saturated
Anode:  Pt°
Cathode: Pt°
Bias Voltage: 160 mV
Titration:
  21° + SO, + H,O -*
  SO, + 21- + 2H+
Generation: 2I~ -* 21° -f 2e~
       Rcductive cell
(T-400-S) Silver cell
Electrolyte:
  0.3M ammonium hydroxide
  Q.1M sodium acetate

Sensor:  Ag°
Reference:
  H£°/HgO Saturated
Anode:  Ag°
Cathode: Pt°
Bias Voltage: 110 mV
Titration:
  2Ag* + S«- -* Ag,S

Generation:
  Ag° -^ Ag^ + e~
Au^illary gas flow:
  10% NH, in Ni at 40 cc/min

-------
 5.3  Standardization
                                                              -  50  -
                                             Oxidaiive cell                              Dohrmann
                                              Iodine cell                                Application Note  MC301
      5.3.1   Not all of the  sulfur  in the samples comes through the furnace as titratable S02. In the  strongly
      oxidative conditions of the pyrolysis tube, some of the sulfur is also converted to S03 which does not react
      with the  titrant.  Accordingly, system recovery should be verified  every four hours  (more often in critical
      applications) by using  the standards supplied with the  MCTS-30 System or in the CEM-30 Chemical Kit
      System recovery is typically 85% ± 10%.

   7  RANGE AND PRECISION
      T'   following specifications are based upon an injected  sample volume of 5 - 8 n\ at concentrations greater than
      2   sm and 30 - 40 (J.\ at concentrations of 2 ppm or less. The amount of analyte per injection should not exceed
      10,0(  ng and the sample injection rate should not exceed 1 /ul/sec.

      The following precision data is typical:

                C ppm                    PRECISION  (± ppm)         RELATIVE PRECISION  (± %)

                  0.1                           0.05                             50.0
                  0.5                          0.07                             14.0
                  1.0                          0.10                             10.0
                  10.0                          0.50                              5.0
                100.0                          3.20                              3.2
                500 and greater                  -                                3.0
                                              Reductive cell                           Application Note MC401
5.3 Standardization                             Silver cell

     5.3.1  System recovery should be verified every four hours (more often in critical applications) by using the
     standards supplied  with the MCTS-40 System or in the CEM-40 Chemical Kit. System recovery is typically
     100% ± 5%.
  7  RANGE AND PRECISION
    The following specifications are based upon an injected sample volume of 5 • 8 jul. The amount of analyte per in-
    jection should not exceed 10,000 ng and the sample injection rate should not exceed 0.2/^l/sec. The following pre-
    cision data is typical:

              C ppm                   PRECISION (± ppm)         RELATIVE PRECISION (±%)

                  1                             0.5                             50.0
                 10                             0.8                               8.0
                100                             3.5                               3.5
                500  and greater                  -                               3.0

    Above 10,000 ng of sulfur (eg. 1 JLI! of 10,000 ppm S) titration time becomes excessive which in turn degrades pre-
    cision. Concentrations greater than 10,000 ppm can be measured by diluting the sample with a suitable solvent
    C 'utions of up to 1000/1 will not significantly  degrade the precision specifications cited above. Dilutions to the
    100 ppm level provides a good compromise between analysis speed and precision as well as ease of sample handling

 8 ACCURACY
    System recovery is 100% ± 5% when the analyte concentration is calculated from instrumental data using Faraday's
    Law. System  accuracy  will be equal to the sum of system precision and recovery. Accuracy can be significantly
    improved when sample values are compared to those obtained with analyte standards. Accuracy  cannot be better
    than precision, however.

-------
                                           -  51  -
                               PHILIPS SOg MONITOR, ~PW 9700
DESCRIPTION OF  BLOCK DIAGRAM


        The air is introduced through the sampler unit comprising two filters. The air is then lead
        through a selectivity filter to the measuring cell. The measuring cell contains an aqueous solution
        of KBr; Br? and H2SO.. The SO2 in the air reacts with free bromine in this measuring solution.
        The bromine concentration is converted into a redox potential which is compared with a constant
        voltage.
        The difference between both voltages controls a current through two generating electrodes via
        an amplifier.  When the bromine concentration is decreased from a certain value,  this system
        regenerates bromine. This coulometric titration controls the bromine concentration in the
        measuring solution at a constant level.
                          SAMPLER UNIT
                          l>w J71G
                            6LOCKDIAGRAM  CHEMICAL UNIT
          Fig. 2.1  Block diagram SO2 monitor (chemical part)
                                                                                          \
          The current required to maintain this constant balance is directly proportional to the quantity
          of bromine used and consequently to the quantity of SO2 lead through the measuring cell.
          As a result of evaporation,  some bromine will leave the cell, thus causing a small zero current.
          Parts  have been built into the monitor for measuring the size of this zero current automatically.
          A resistance is fitted in series with the generating electrodes. This resistor converts the
          generating current into the input voltage of the output amplifier.

-------
  2. Mgasuring ranee,
            3                      3         -  52  -
    3 mg/m SO  (1.15 ppm SO2/nT air).
                                             3             3
    Other measuring ranges between 1.5 mg/m  and 10 mg/m  can be realised.
                                           q
    In measuring ranges below 0-1,5 mg/m  the calibration signal falls out of scale.
    0-20 mA into a resistance of 0-300 Q.

  3. Response time
    Approx. ij minutes (63 % of final value)
    Approx. 3 • minutes (95 % of final value).

  4. Accuracy (overall,  absolute)
    Better than 15 % of the measuring signal,  Better than 26 pig/m3 SO- (=0, 01 ppm) if the SO0
                                           3
    concentrations are  smaller than 183 /tg/m  (= 0,07 ppm).
    Provided calibration is performed once a day, for instance.

  5. Reproduclbility
    Better than  1 % of the measuring signal.

  6. Detection limit
    Smaller than 10 (Jg/m SO- (4 p.p.b.)
                            ,tt
  7. Size of zero current
                                                            3
    Corresponding to a measuring signal  smaller than 100 ngfm SO^

 8. Drift of the zero signal
    Smaller than 26 ^g/m3 SO0  (=0, 01 ppm) a day, non cumulative.
                            i
 9. Influence of  10 %m.iins voltage  variation
    Negligible.

10. Climatologies! inflviRnce
    Negligible.
11.  Selectivity
    The interference (I ), due to a substance X,  is expressed by the relation I  ^
                                                                               x 100 %, in
   which S  = measuring signal due to 0,05 ppm of substance X and S = measuring signnl due to
   0, 05 ppm SO. .
   X                                   Selectivity
   N'02                                    <   5 %
   O.,                                      <   1%
   H2S                                     <   ] %
   Methyl-mere aptan                        SB 180 %
   Klhylene                                <   2 %
   C12                                     <   2 %
   NO                                      <   1 %
   Aldehydes                               <   ] %
   Henzene, chloroform                     <  j %

-------
                             - 53 -
GAS CHROMATOGRAFHY + VARIOUS DETECTORS (SC>2, l^S, RSH, RjS, RSSR)

          Gas samples may be separated in a chromatographic column and
the eluted components identified by a variety of detection methods.  The
volatile sample is transported through the chromatographic column by an
inert gas (N2> He, A, etc.) and separated by preferential adsorption on
a  finely dispersed column packing.  The adsorbed materials are then removed
with an inert gas and the various components, which desorb at different
rates, may be separately identified and measured as they issue in the
carrier gas.

          Many different types of column packing materials have been
evaluated (95., 81,  T^, 94_,  86., 10., 74, 75_, 96,, 87, 97., 98, 89_, 90) but two
appear most promising for the separation of sulfur compounds:  Triton X-305
coated on Chromosorb  (99) and (100) polyphenylether coated on finely
devided Teflon  (101).  For  the separation  of amines,  aldehydes, ketones,
alcohols and mixtures of these components,  two GC columns  in series contain-
ing Chromosorb 102  coated with Emulphor 0  in the first column and, either
sorbitol or diglycerol is the second  column, were found  to be efficient
by Dravnieks  (69).

          Following separation of the test sample into its individual
components  or simpler mixtures by GC, these materials may then be
 identified  by a  variety  of  detector techniques,  such  as  those  listed  below.

                Colorimetric
               (Micro) Coulometric (M)CD
                Electron  Capture  (EC)
                Flame lonization  (FID)
                Flame Photometric (FPD)
                Infrared Spectroscopy
                Thermal Conductivity  (TCD)
                Ultraviolet Spectroscopy

           Most of the above techniques are discussed elsewhere in this
 report and only the FPD, FID, TC and EC detectors will be covered here.
 FPD is a preferred method, highly sensitive and specific for sulfur in any
 volatile form.  FID responds to any organic sulfide, but not to S02 or l^S.
 Thermal conductivity and electron capture are older methods, less sensitive
 to sulfur and of limited use at present.

 rtommercial Equipment

            Gas chromatographic columns, packings, coating agents for  various
 packings and detectors  are available from a wide variety of suppliers  and a
 complete  Listing of  all is beyond the scope of  this  report.  Commercial
 units  combining GC/FPD  which should  be  suitable for  use in kraft pulp
 mills  or  petroleum refineries are offered by  Bendix, Tracor,  and Varian
 Aerograph.  While these components  are  also  available  for monitoring
  specific  odorants in emissions  from animal waste rendering or other  sources,
  no combinations have been offered  or recommended as  a  package unit for this
  purpose.

-------
                               - 54 -
References

GC + FPD
           1.   Principle of Operation:   81^,  82. J73,  8_3,  74_, 66_
           2.   Applications:  81, 82_, 73» il'  74,  84,  85_
           3.   Data:  jft, 82, 13, J33, 74_, 85.
           4.   Specific Instrument Descriptions:   JU,  82_,  73. 83, 74, 84_
GC + (M)CD
GC + TCD
GC + FID
GC + EC
           1.   Principle of Operation:  61.,  71,  17,  86,  22_, 75^, 74_, 84_
           2.   Applications:  jtt, 71, 17,  86,  64_,  22_,  75,  74, 84, 90
           3.   Data:  6^L, 71, 17_, 86, 74_,  75_,  90
           4.   Specific Instrument Descriptions:  61,  71,  17. 86, 74. _75»
                                                   84,  90
           1.   Principle of Operation:  17,  74,  91
           2.   Applications:  17, 64, 74, jtt
           3.   Data:   _17, 74_, 9^
           4.   Specific Instrument Descriptions:  17, _74.
          1.   Principle of Operation:  17, 86, 22,  74
          2.   Applications:  .T7, 86_, 64_, 22^, 92_, _7_i, 75,  93
          3.   Data:   1.7, &6, 64, 22^, lk_
          4.   Specific Instrument Descriptions:  17, 86,  _74_
          1.  Principle  of Operation:  74
          2.  Applications:  74, 93.
          3.  Data:   74
          4.  Specific Instrument Descriptions:   74

-------
                                 - 55 -
Discussion

Advantage

GC + FPD

          1.  Capable of detecting S compounds at 2 ppb level (81) .

          2.  Can detect a wide variety of S compounds such as SC-2,
              G1~C3SH> and higher molecular weight sulfides + disulfides (82)

          3.  Shows  potential  for use  in monitoring  (73,  84).

          4.  Interference  from non-sulfur compounds  is small  (83), except
              for carbon oxides or hydrocarbons  at concentrations  greatly in
              excess of total  sulfur.

 GC + (M)CD

           1.  Best  combination found for kraft pulp  mill  emissions, but not
              completely satisfactory (22).

           2.  Minimum detection  level of 0.05  ppm P02 and 0.02 ppra H^S  (75).

           3.  Equal of similar in  sensitivity  to GC  + EPD (61).


 GC + TCP

           1.  Sensitive to S02, H2S and mercaptans.

 GC + FID

            1.  Method for carbon number analysis in  any homologous series.

           2.  Might be useful at high concentrations for S compounds (74).

            3.  Combined with a S detector it can be  very  useful  and has high
               sensitivity  for hydrocarbons.
  GC + EC
            1.   Good sensitivity to  compounds containing one or more
                functional groups.

            2.   Can serve as a complementary system with GC + FID.

-------
                                  - 56 -
 Disadvantages

 GC + FPD

           1.  Massive amounts of CO, C02 or hydrocarbon interfere.
 GC + MCD
 GC + TCD
           1.   Dohrmann type iodine cell  not  suitable  for  routine process
               analysis (22).

           2.   Requires standardization of the coulometric cell  for each
               type of compound eluted (22).
           1,   Sensitivity for the  detection  of H2S  and CH-SH about 500 ppm
               which is not suitable  for  direct analyses or process gases
               (74,  91).
 GC + FID
           1.  Not  suitable  for  sulfur compound  in the presence of hydrocarbons
              since they  are obscured by  S  free compounds, and the response
              does not  follow the "carbon number" rule as with aliphatic
              hydrocarbons  (94, 86, 22).

PC + EC

           1.  Not suitable  for sulfur compounds.

           2.  Small sample  capacity limits usefulness (36) because of
              interference  caused by ©2 (74).

           3.  Not sensitive to hydrocarbons.

Conclusions

          GC instruments combined with an appropriate detector are extremely
sensitive and discriminating.  Such combinations may be used for all known
gas mixtures, to Identify most individual compounds or to identify com-
pounds by type analysis.  The most versatile combinations for the kraft and
petroleum Industries are:

          1.  GC plus flame photometric detectors (FPD).
          2.  GC plus microcoulometric detectors (M)CD.

The other techniques have some serious drawbacks.

-------
                                - 57 -
Re commendations

          Continue evaluation of GC + FPD and GC + MOD systems for monitoring
emissions from kraft and petroleum industries.   Need to evaluate present
initial production instruments for commercial use.

-------
I
^MiMja
            /o
                  I' ~. oot
                                  TiTA-i-r
                                c
                                                   OdV
^fe
                                                                                      en
                                                                                      fl>
                                                                                      3
                                                                                      o
                                                                                      o
                                                                                      O
                                                                                      h1-
                                                                                      03
                                                                                      OQ
                                                                                      K
                                                                                      03
                                                                                      3
                                                                                          w
                                                                                          fD
                                                                                          D
                                                                                          Cu
                                                   w

                                                   <


                                                   o


                                                   3
                                                   0)
                                                   3
                                                   rt-
                                                   EU
                                                   I—*

                                                   O


                                                   o

                                                   3

                                                   rf
                                                   O
                                                   oq

-------
                               - 59 -
   The Bendix Environmental Chromatograph will meet or surpass the

following specifications if properly maintained and the specified operating

parameters are adhered to.

• Ranges:  0. 02 to 0-20 ppm on continuously adjustable attenuator

• Precision: *4% of full scale

• Minimum Detectable Limit:  less than . 005 ppm H£S and 0. 01  ppm SC>2

• Noise Level:  0. 5% of full scale

• Zero Drift:  Less than tl% per day and 2% per three days with automatic
              zero before each component.

• Span Drift:  Less than tl% per day and 2% per three  days

• Interference Equivalent: Less than 0.01 ppm

• Electrometer: Linearized electrometer corrects the signal to  provide output
                directly proportional to sulfur  concentration

• Linearity:  Less than 2% of full scale

e Oven Temperature:  Controlled to 0. 5°C

o Cycle Time:  Five minutes

« Operational  Period:  More than three days unattended

e Output Signal:  Trend outputs of 0-10 volts or 1-5,  4-20 or  10-50 made for
                each component

• Readout: Chromatogram, bargraph and trend

• Power:  115V 60 hz

-------
                                  - 60 -
 TAPE  SAMPLERS  (H2S^

 Principle  of Operation

           The  reaction of water  soluble  compounds of heavy metals such as
 Pb(Ac)2, HgCl2 or  Cd(OH)2 etc. with H2S  results in an insoluble sulfide
 precipitate:

           Pb(Ac)2  + H2S - >  PbS   +  2HAc

           Cd(OH)   + HS - >  CdS   +  2H0

 Filter paper on  other  appropriate material may be impregnated with Pb(Ac)
 or  Cd(OH>2  and air  drawn  through the paper will cause a response to sulfide
 compounds.  Mercaptans may or may not react.  Such papers may be supplied
 as  a tape for continuous monitoring of sulfide in air.

          A number  of  schemes have been devised to automate tape samplers
 and various means are  available for determination of the concentration
 level.  One of the  simplest methods is to  read  the color intensity or
 amount  of precipitate  developed with some sort of spot evaluator (102).
 Other  methods involve  dissolving the precipitate formed on the tape with
 reagents such as NItyOH and measuring the intensity of the brown color
 developed colorimetrically (102, 103) .  The sulfide may also be dissovled
 and measured colorimetrically by the methylene blue method (104) .   In each
 of  these cases the  sulfide concentration observed is correlated with a
 standard ILS curve.

          Modifications have been made by various researchers and  equip-
ment suppliers to overcome the inherent problems such as  light sensitivity
humidity control, and variations in tape response.   The limits of  detection
 range from  0.5 to 100 ppb.  Lead acetate detectors have a lower limit of
about 50-60ppb, Cd(OH)2 about 1-3 ppb and the HgCl2 type unit is sensitive
to 0.5 ppb.

Commercial Equipment

          No instrument is available which gives a direct  output or linear
readout proportional to sulfide concentration.   Lead acetate tape  samplers
according to several different principles are available.

          Houston Atlas 825:   continuous readout of exposed trace  on moving
                              tape.

          Houston Atlas 855:   catalytically hydrogenates all sulfides to
                              H2S,  fed to tape  sampler.

          RAG 5000:   cumulative spot analysis on 5  minute cycle (or
                     multiples),  each spot  pre-corrected for tape  zero.

          Bendix  Monocolor (Maihak) :   continuous readout on moving tape,
                                      zeroed against light absorbance of
                                      unexposed tape.

-------
                                 - 61 -
          Simplified samplers are also available for use in field stations
with provision for separate tape readout in a control laboratory.
Calibration curves are generally provided with each instrument, but these
have been found to be inaccurate by a number of investigators.

References

          1.  Principle of Operation:  ji8, £6_, 105, 103, 104, 106. 101

          2.  Applications:  88, 66, 105

          3.  Data:   88, 66, 105

          4.  Specific Instrument Description:  88, 66, 105

Discussion,

 Advantages

           1.   Low cost.

           2.   Simple operation and maintenance.

           3.   Cd(OH)2 and HgCl2 detectors sensitive to very low
               concent rat ions.

           4.   No interference  with CH..SH, at lUS  to CH-SH ratios
               greater than 1.0.

 Disadvantages

           1.   Lead sulfide impregnated tape sensitive to  light and fades
               with time, other materials such as  Cd(OH>2  and Hgd-2 more
               stable.

           2.   Lead sulfide not accurate at concentrations below 50 ppb.

           3.   Need to control humidity of gas stream.

           4.   Ozone and S02 cause fading of lead acetate tapes.

 C!onclus ions

           Tape samplers good for detection of I^S  in air above  50 ppb
 provided adequate precautions  are taken to control humidity  of  gas and
 to minimize  fading problems.

 Recommendations

           Further development  needed to  fully automate the  system and
 improve accuracy.

-------
                                 -  62 -
                               TAPE SAMPLERS
                MonocolotMClOOl      Maihak AC, Hamburg.  (147)
                                        sample cell;
                                        reference cell;
                                        lump;
                                        photocell 1;
                                        photocell 2;
                                        paper tape impregnated with rciigrfnt;
                                        box for papur lapc;
                                        spool head for paper t'tj^;
                                        indicating device.
  The instrument  operates according to the following principle: A paper tape
impregnated with the reagent (e.g., lead acetate for H2S) is successively Jr;i wn thrc;:^!;
two cells. One of these, the rr-fcrencc ccl!, is fiiicd with pure air, while tlie air sample
passing over the  paper  is blown into  the sample cell.  Bolh  tells,  which  are
symmetrically arranged are illuminated by a common lamp. The light reflected from
both lape surfaces is respectively measured by two photocells. The current diflcroiKo
corresponds to the blackening of the paper caused by the foreign matter in the air
sample (e.g., H2S). The tape feed rate is 1 mm/tnin; since the diameter of the sample
cell is 12 mm, the paper is  illuminated 12min and a mean value over this period is
measured. The throughput rate of the  sample gas (c.g., lOOml/min) can be adjusted
by means of control valves and measured. The lowest measurement range for H2S
is 0-14 ppm.
  A similar instrument, the AISI Automatic Smoke Sampler (Research Appliance
Co., Pittsburgh, USA),    can also be used for determining H3S when a  paper
impregnated with lead acetate is inserted.
                              Air MM
                                          Revolving spool  Motor for spool
    Stated  accuracy  + 15% PS  at  0,4  ppm,  +  30% at 0.05 ppm,
      higher concentrations measured by dynamic  dilution.

-------
                                     -  63 -
                            HOUSTON  ATLAS,  INC.



 The Model 855 Analyzer is comprised of two components:
 The PYROLYZER*  for sample preparation and conversion of ad liquid hydrocarbons to yases  and ai
 sulfur compounds into hydrogen sulfide. The Model 825 H2S Analyzer for readout of sulfur content.



                                GENERAL SPECIFICATIONS
 RESPONSE  TIME:



  RANGES:




 LINEARITY:


 SUPPLIES:



 CONSTRUCTION:
93% in 30 minutes at 2 ppm full scale plus flow tubing resident time. More rapid
response at high sulfur levels.


As specified.  0-2  parts per  million to  0-20% by weight sulfur.  Normally

continuous reading, peak reading below 50 ppm.



± 3%  of scale  (+  3%  at  50% FS).


Hydrogen gas  (sulfur free) at  1  SCFH  required.  Lecd acetate sensing tape. 5%
acetic acid solution.


Flow system stainless steel, 1/16  inch tubing. Available in explosion proof model.
                              RAC  Model  5000
OPERATION
  Filter paper tape is fed through the sampling nozzle where
air is drawn through tape  by an oil-less vacuum pump at
approximately 0,25 cubic feet per minute. Pump rate  (1.1
CFH) is controlled and set  at 0.25 CFM (15CFH) by a  flow
adjust valve.

  Filtered exhaust air places slight positive pressure in sam-
pler's front  area  to prevent any  contamination of  tape
around sampling spot.

  Tape,  punched on two inch centers for positive indexing,
is moved automatically  by timer which can  be  pre-set for
any sampling period from 5 minutes to 3 hours 20 minutes
in 5 minute increments. When  sampling period  ends,  tape
is automatically indexed.
  Samplers  are easily adapted  for sampling  gases such as
hydrogen sulfide. A prefilter and humidity jar is attached to
air inlet and exhaust gases are  passed through a soda and
lime tube to scrub out any hydrogen sulfide  that may  have
passed through the tape.

Direct reading range:   0.001  to 0.5 ppm

      in 1  hour sample  (x  12 iri  5  minutes)
      higher  levels by  flow dilution.
Stated accuracy:

      + 15% of  scale  at 0.4 ppm

      + 30% of  scale  at 0.05 ppm
                                                                         ; ,-JT*»rt\ ** I*
                                                          2^.   ~3



                                                    y  '*'."•%
"Irii; •"
o if^
. !
,
5;
'-,-

>}

-
u
                                                                    re
                                                        12     10

-------
CONDUCTIMETRY
Principle of Operation
          The air sample is absorbed in a dilute solution of sulfuric
acid (H2SO^) containing hydrogen peroxide (H202) .   Any S02 in the sample
is converted to SO 3 and H2S04 according to the reaction below, thus
increasing the conductivity which is recorded as S02 concentration,
Most conductivity systems are equipped to measure conductance in solutions
between two platinum electrodes.   Distilled or deionized water should be
used.  Absorption of C02 is prevented and its interference reduced by using
enough I^SO^ to give an equivalent of 2X1Q-5N and enough H202 (3%) to bring
the absorbent to a strength of 3X10~% with respect to peroxide.  Instru-
ments are calibrated against primary standards.

Commercial Equipment

          Instruments from a number of suppliers are available as listed
below :

          Beckman Instruments, Fullerton, Calif.; Models K78 and K1006
          Inst. Development Corp; Reston Va. ; Models 903-1, 902-3
          Kimoto Electric, Osaka, Japan; Model S-350N
          Leeds and Northrup, North Wales, Pa.;  Model 7860
          Research Appliances
          Hartmann-Braun = Intertech, Princeton, N.J.
          Wo'sthoff (Mikrogas) - Calibrated Instruments, Ards ley,  N.Y. ;
             Model U3S
          Davis Instruments, Charlottesville,  Va.;  Model 11-7000
          Scott Aviation; Model 7010 RPA
          Scientific Industries,  Hempstead, N.Y.; Model 67

References

          1.  Principle of Operation:  63, 73, 35

          2,  Applications:  6_3,  107, ^73

          3.  Data:  £3_, 107, _7_1

          4.  Specific Instrument Description:  63

-------
                                - 65 -
Discussion

Advantages

          1.  Very good performance in the range of 0.2 - 4 ppm with a
              standard deviation of 0.01 - 0.9 ppm.

          2.  Specific when SCL is the principal acid gas oresent.

          3.  Interference due to NOX at concentrations generally found
              in ambient atmosphere is not significant.

          4,  Used widely by many air pollution agencies.

Disadvantages

          1.  Errors can occur if interfering substances exceed 10% of S00
              level.                                                     L

          2.  Interfering substances are not easy to monitor.

Conclusions

          Most instruments are sensitive enough  for  general  use with  a
minimum of  interference problems.  The various instruments generally
correlate fairly well.  Improvements  are still being made  for more
selective monitors, which will be  reliable under varying conditions.

Re commen dat ions

          Most instruments perform well but need frequent  checking  against
standards.

-------
                              - 66 -
ULTRAVIOLET SPECTROPHOTOMETRY  (H2S, S02>

Principle  of Operation

           Direct measurement of  S02 is possible  in  the  ultraviolet  (UV)
range  since this gas  absorbs strongly at  about 285  nm.  H2S may be  con-
verted to  S02 by oxidation, or converted  to other derivatives which
 ibsorb in  this  range.  The effects  of UV  interferences  can be  minimized
by dual beam measurements  at two wavelengths.  Methods  and equipment have
also  been  developed for the detection of  H2S  and SC>2 in the presence of
one another  (108,  109).

           In  the method for detection  of  H2S  and S02 simultaneously (109),
 the sample gas  is  mixed with air, filtered and split into two  streams.
 One stream passes  through an oxidation  furnace to convert the  H2S to S02
 and the other  stream, which is used as a reference, is passed  through a
 dummy furnace  and  exits unchanged.   The two streams are analyzed for S02
 and H2S content determined by  difference.  Systems  can be arranged to
 use the unchanged  stream in a  reference cell  and compare the differences
 in UV absorption  electronically to give a direct reading proportional to
 the HLS content of the sample  gas.   Such a system can measure  S02 and/or
 H2S in the range  of 10-2500 ppm.  In another  system, H2S is absorbed on
 a filter medium impregnated with Pb(Ac)2 and  the PbS precipitate extracted
 with  organic solvent (CH3COOCH3, Ci^OH and HOAc) (no).  The resulting brown
 suspension is  measured with a spectrophotometer at 350 nm.  The amount of
 H2S is determined by comparison to a standard curve.  Reproducibility is
 reportedly +15% (110) with very little  humidity effect.

 Commercial Equipment

           The duPont Model 461 UV Spectrophotometer  is available to
 automatically  determine S02 (and H2S).   The major components for the oxida-
 tion  method including  the sample handling system, have also been made by
 Analytic Systems Co. to be used with their Series 600 UV analyzer, but not
 on a  commercial basis.

 References
        1. Principles of operation; 110,  111, 109

        2. Applications: 110,  109

        3. Data:  112,  110, 109

        4. Specific instrument description:  110, 109

-------
                              - 67 -
Discussion

Advantages

       1. Specific for S02> at high concentrations.

       2. Reliable measurement over a wide concentration range.

       3. The Pb(Ac)2 extraction method gives a collection efficiency of
          greater than 90% and compares favorably with the
          •methylene blue method (J12) .

Pis advantages

       1. Dual measurement is more reliable for differences or
          ratio measurements of l^S and SC^ than for absolute values.

       2. Other reduced sulfides also oxidized to  S02«

       3. 862 and ozone interfere in  the Pb(Ac)2 method but
          can be removed by scrubbing for I^S only determinations.

       4. The Pb(Ac>2 tape fades with time.

Conclusions

          Satisfactory for field sampling, but other methods such  as  the
coulometric  titrators, tape samplers and combined  GC techniques generally
better where absolute values are required.

Recommendations

          Not recommended  for  complex mixtures, useful where H S and S09
are the principal constituents to be determined.              2

-------
                                - 68 -
COLORIMETRY (S02, H2S, Mercaptans)

Principle of Operation

          A sample of gas is bubbled through a solution which selectively
absorbs the component desired.  The absorbed compound is than reacted with
specific reagents to form a characteristic color which is measured
spectroscopica]ly.

          Separate methods have been developed for the determination of
S(>2 and sulfides  (^S and mercaptans).  Both types of gases may be detected
in the presence of the other by use of scrubbers to absorb the undesired
materials.

          For the determination of SC>2 a dilute aqueous solution of sodium
tetrachlororaercurate is used for  absorption to form nonvolatile techlorosulfito-
mercurate ion.  This is reacted with formaldehyde and bleached pararosaniline
to form a red-purple color and measured at 560 nm.   This is the modified
West-Gaeke method and has been tentatively adopted as a standard by the
NAPCA (114).

          "Total  reduced sulfides" including H2S  and volatile mercaptans
may be determined colorimetrically by the  methylene blue method which
involves absorption of the gas in an alkaline suspension of Cd(OH)2 to
form CdS.  The precipitate is then reacted with a strongly acid solution
of N, N dimethyl-p-phenylenediamine and FeCl3 to give methylene blue, which
is measured spectrophotometrically at 500 nm.  This procedure has been
adopted as a standard method for H0S (108).
                                  £,     ' '

Commercial Equipment

1.  SO:? Analysis

          Instrument manufacturers (Atlas, IDC, Monitor Labs, PSD, Technicon
and WACO) supply components to collect and analyze gases for S02.  Selective
prefilters may be added depending upon the nature of the gas sample to be
analyzed.  Automated analytical trains for this procedure are supplied by
Technicon and others in this country and by  several  manufacturers  in
Germany, England, and Japan.

2.  H?S Analysis

          Instruments such as those noted above for S02 may be adapted  for
H2S analysis.

References^

          1.   Principle of Operation:
              For SO?:   6!3,  64.,  _73, 115.  35_,  116, 6_7, 114,  117
              For H?S and Mercaptans:   64_, _73,  115, 92_,  118, 119, 120, 121
                                       122,  116,  67

-------
                     - 69 -
2.  Applications:
    For S02:  63, 64, 73
    For H2S and Mercaptans;  64, 115, 92, ^7

3.  Data:
    For S02:  63, _73, _35_
    For H2S and Mercaptans;  73, 115, 92_, 123. 124, 108, 67,

4.  Specific Instrument Description:
    For 502:  63., 125. 99., 100
    For H2S and Mercaptans:  115, 108

-------
                                    -  70  -
Discussion

For SO?

Advant ages

          1.  Specific to S02.

          2.  Levels from 0.002 to 5 ppm can be detected,

          3.  Sensitivity in the range of 0.02 - 0.04 ppm or better
              (in the 1 ppm range).

          4.  Correlates well with most other test methods.

          5.  Response time to changes in S02 concentration fairly rapid
              ( 15 min.)

          6.  Has been chosen as a tentative standard by EPA.

Disadvantages

          1.  Ozone, C12, HC1, NH3 and N02 interfere if more than 10% of
              S02 concentration, by destroying some of the dye.

          2.  Requires careful reagent quality and temperature control.

For H2S

Advantages
                                               3
          1.  Sensitive to 1-100 ppb at 10 mg/m .

          2.  Specific for H-S and reduced sulfides.

          3.  Standard deviation + 3.5%, recovery 80% or better.

          4.  Being used as a tentative standard for ambient H2S measurement,

Pis ady ant age s

          1.  Strong reducing agents (e.g., S02) inhibit color development.

          2.  Total reduced sulfides respond.

          3.  N02 above 0.3 ppm and 0. above 57 ppb interfere.

-------
                               - 71 -
          4.   Takes discrete samples, not a continuous method.

          5.   Economics are favorable only when a large number of similar
              samples are to be measured in a short time.

Conclusions

          Colorimetric methods for both SOo and t^S (and mercaptans) are
very satisfactory, but have not been developed yet to give direct, continuous
measurements,

Re commen dat ions

          Further work is needed to develop fully automated, continuous,
rapid response techniques.

-------
                              - 72 -
ELECTROCHEMICAL

Principle of Operation

          The pollutant gas diffuses through a selective membrane and
dissolves in the underlying electrolyte film and becomes adsorbed at a
sensing electrode where a charge transfer occurs.  The flow of electrons
forms a current which is proportional to the partial pressure of the
pollutant.  Such sensors are reported to be liquid-state devices that con-
tain surfaces onto which the pollutants are adsorbed.  The adsorption is
followed by electron transfer either to or from the adsorbate.  The cur-
rent flow is determined mainly by the rate at which the gas molecules
reach the catalytically active surface and by the valence change which
the adsorbed species undergo at this surface.

          Presumably the catalytically active electrodes and the electrolyte
are enclosed in the sensor that is equipped on one surface with a semi-
pereable membrane.  The membrane is exposed to the atmosphere being
monitored.  The current generation by the sensor is permeation-controlled
by the transport of the pollutant gas through the membrane.  The rate of
permeation is proportional to the pollutant concentration in the atmosphere.
The selectivity of the sensor is attained through the appropriate combination
of catalytically active electrodes and electrolyte.

Commercial Equipment

          Prototype instruments are available by Dynasciences and by Enviro
Metrics, Inc. for SO  or H S, but not for other sulfur-containing gases.

Discussion

          Although the principle has been demonstrated, application of this
method is still under development.  Test results to date indicate that
sensors in the emission concentration range  (0-1000 ppm) are more reliable
than those offered for the ambient range (0.04-2.0 ppm).

References;  (35)

-------
BIOLUMINESCENCE (H S, RSH)

Principle of Operation

          This detection system involves the utilization of extremely
sensitive and highly reliable biosensors - living, light-emitting
(bioluminescent) organisms.  The biosensor cartridge is placed in an
electro-mechanical system which is equipped with a photocell to monitor
the light generated by the bacteria.  An air pump pulls the sample gas
across the surface of the biosensor whose response is monitored electronically
and displayed on a meter.  Specific sensors must be selected to monitor
specific pollutants.


Commercial Equijjmenjb

          Prototype units and sensors are available from RPC Corporation,
El Segundo, California.

Discussion

          The supplier's information indicates a diverse capability by use
of individual strains of bacteria which respond selectively to different
chemicals.  Sensitivity can be greatly improved by selective breeding of
bacterial strains and has been extended in some cases  to concentrations
below 10~12 mol fraction.  Certain biosensors are capable of responding
to classes of chemicals, but yet are capable of differentiating qualitatively
between members of that class.  Sensitivity of specific cultures now
available permits detection of H^S at 70 ppb, S09 at 0.2 ppb, or NH0 at
0.001 ppb.                                      L                  3


Reference:  126

-------
                                    - 74  -
 CHEMILUMINESCENCE


Principle of Operation


          NH3 Is converted to N02 by thermal conversion at above 300°C
and then reacted with ozone to produce nitric oxide

                            A
               ANH3 + S02 ——	>   6H2O + 4NO                       (1)

                N0+03	>   N02*+02                           (2)


          Chemiluminescence is produced by deactivation^of the N02*


               NOo*  	>  NOo + hV (     - 6300 A)                (3)
                 *•              ^        max                            r

           Chemiluminescence detectors, based  on  a prototype  unit developed
by Ford Research of Dearborn, which have been developed for the monitoring
of NOX in ambient air, may be used combined with a thermal preconverter to
 oxidize  the NH3.  This equipment has a fast linear response  and is  sensitive
to concentrations of O.OOSppm (127).  The equipment consists of a variable
temperature chamber containing catalytic surfaces in which the gas stream
and oxidant are reacted.  Conventional.,ozone .generators with appropriate,
scrubbers to provide clean air may be used to furnish the atomic oxygen
for the conversion of NO to N02*.

          N02, S02 and CO also react with atomic oxygen to produce
chemiluminescence.  S02 and C0~ activated oxygen  compounds emit at different
wavelengths (3500 A and 4350 A.) and may be distinguished from N0£* emissions.
To detect NH3 in the presence of NOjj the NH3 may be separated from the NO*
components and determined by difference using either a split gas stream or
separate analyses for total and total less NH3.

Commercial Equipment

          Commercial units are available from Aerochem, Bendix, Scott Research,
REM and Thermo Electron.
References
          Principle Of Operation:  127. 128, 129
          Applications:  127. 130. 131
          Data:  127, 129
          Specific Instrument Descriptions:  127, 132

-------
                                - 75 -
Discussion
Advantages
          1.  CO and SC>2 cause no interference.
          2.  NOX can be readily distinguished by separating out NH3,
          3*  Rapid response (1 sec.)*
          4.  High reliability anticipated.
          5.  Reasonable cost.
          6.  Range of 0.002 -1000 ppm.
Disadvantages
                         ,3:'-     ..      .        •
          1.  Continuous evacuation to 1-3 torr required for reactor flow.
          2.  Other reduced H compounds interfere i.e., amines.

Conclusions

          Not specific enough for NH3, other methods such as direct titra-
tion probably better suited, but combinations with NOX can be measured.

Recommendations        'r"

          Satisfactory for use in combinations containing NOX and NH3.  A
similar method might be possible for determination of S02*

-------
                               - 76 -
 FAST SCAN INFRARED SPECTROMETER (S02)

 Principle of Operation

           A single-beam spectrophotometer with automatic repetitive
 scanning in 5 or 12.5  seconds,  of wavelengths in three divisions 2.5
 to 4.5 microns;  4.4 to 8.0 microns and 7.9 to 14.5 microns  with
 interference filters is used, coupled with a high speed hot-stylus
 oscillographic recorder.   Bands in the spectrum from C02 and moisture
 in the air are corrected by means of a background trace before and after
 analysis.  Spectra are obtained by simultaneously passing the pure gas
 and dilution gas (He or N2)  through the gas cell in a flow-through manner.
 The small spectrum makes it essential  that measurement of the band
 absorbance  is accurate.   To assure this,  part of the standard IR absor-
 bance scale is redrawn,  photographed,  and reduced to the size of the
 instrument spectra.  A transparent plastic overlay is then  placed  over
 the spectra and  the absorbance  of the  individual band is measured.  With
 the aid of an illuminated magnifier, readings can be obtained with  an
 accuracy of +J0.05  to +0.1 and good reproducibility.

 Commercial Equipment

           Several  fast-scan  IR  spectrophotometers are available such as  the
 Warner-Swasey (OCLI) and Beckman IR-102.   A completely automated analyzer
 system has  not been  produced.

 References
       1. Principle of operation:  133
       2. Applications:  133
       3. Data:  133
       4. Specific instrument descriptions:  133
Discussion
Advantages
       1. Gives a complete spectrum in 12.5 seconds or less.
       2. The instruments can be made portable and may be used up
          to temperatures of 300°C.
Disadvantages

       1. Cannot be used in actual stack gas analysis.
       2. Suited primarily as a research tool.

Conclusions

          A good method for research studies with clean, dry gases.

Recommendations

          Not practical for control studies.

-------
                                - 77  -
FLUORESCENCE (H2S)

Principle of Operation

          Fluorescent liquids can be made which react with I^S and other
sulfur compounds resulting in a reduction of fluorescence.  The change in
fluorescence may be related to the concentration of S compound absorbed.
The original procedure developed (73) utilized tetra-acetoxymercarifluorescein
(IMF) and a method has been suggested for its use in the determination of
H£S in the atmosphere (48).  Recent improvements (47) include use of
fluorescein mercuric acetate (FMA) instead of (TMF) since the former is
commercially available.

          Photofluorometers are available for measurement of the solution
fluorscence in conjunction with a paried optical filters  (Corning Glass
No. 5113 and 2424).  The method involves passing a known volume of air
through a specific volume of FMA solution in 0.01N NaOH.  Fluorscence is
measured and compared with a standard calibration curve.  A linear response
is obtained at H2S concentrations ranging from about 4-6 yg of I^S in
25 ml. of solution.

Commercial Equipment

          None available, but photofluorometers are available such as the
Model 12C from Coleman Instruments, which could be built  into a prototype
unit for monitoring ambient air or emissions.
References
          1.  Principle of Operation:  4J7,  4£,  73,  35_
          2.  Applications:   47. 48.  35_
          3.  Data:  47. 4j8
          4.  Specific Instrument  Description:   47

-------
                                - 78 -
Discussion
Advantages
          1.  FMA reagent stable up to 16 days.
          2.  Method reliable and accurate to ± 1% in the range of 2-8 yg
              of H£S.
Pisjtdvant age s

          1.  Mercaptans and disulfides also cause quenching of fluorescence
          2.  Effect is non-linear in multicomponent blends.
          3.  Solutions stable for only 16 days.
          A.  Acidic materials in air such as C02, S02 could affect reagent
              if over SOOppm by volume in air.

Conclusions

          Good for H2S in presence of C02 + SC^ at low air concentrations.
Mercaptans + disulfides may interfere.

Re commen dat ions

          Consider for a prototype instrument.

-------
                              - 79 -
INFRARED SPECTROSCOPY (H2S, CI^SH, RSH, RSSR)

          Infrared spectroscopy measures the energy absorption pattern
in the vibrational and rotational modes of organic molecules.  This
characteristic can be utilized to measure organic compounds such as
mercaptans directly with the IR, or derivatives of H2S which are
prepared subsequent to collection.  IR is not a good method for I^S itself
          For direct measurement of gases, the material is collected in
a suitable container, and water generally removed prior to measurement
of I.R. absorption.   Direct measurement has limited use due to the
complexity of most odorous emissions and a preliminary separation is
usually preferred.  Combinations of IR with gas chromato graphic (GC)
separation techniques are discussed in the section on GC.

Commercial Equipment

          IR spectrophotometers suitable for the detection of sulfides and
sulfur derivatives are available from Bausch and Lomb, Beckman, Perkin-Elraer,
and many other manufacturers.

References
       1, Principles of operation:  137

       2. Application:  137

       3. Data:  13"7

       4.  Specific instrument description:  137

-------
                                 -  80  -


 Discussion

 Advantages

        1.  Permits positive identification of components.

        2.  Interference effects are slight in most simple  mixtures.

 Pis advan t_ages

        1.  Separation techniques generally required prior  to analysis
           due  to the complexity of ambient or emission gas samples.

        2.  Complete analysis is a research procedure, not  for routine
           use.
        3.  Not a good  method  for H-S ,  which has  a. weak spectrum.

Conclusions

          IR methods have limited application for  direct measurements,

Re commendat ions

          IR is a powerful tool when combined with an appropriate
separation technique and should be considered for  identification tech-
niques in such systems.

-------
                                 -  81  -
IR LASER RADIATION

Principle of Operation

          Carbon dioxide or Iodine infrared lasers are attenuated when
passed through a  sample gas and this change is related to the concentra-
tion of pollutants in the gas sample.  The narrow spectral width of the
laser emission permits sensitive detection of pollutants.   Three modes of
operating the laser probe are possible (1) using a folded  beam laser,
(2) reflecting the laser beam from a retro-reflector, and  (3) reflecting
the laser beam from natural targets.

Commercial Equipment

          No commercial equipment is available although the laser components
are produced.

Discussion

          Currently efforts are being made to build prototype units for
field testing.  This technique shows promise because it permits remote
sensing, is not affected by atmospheric moisture, and minimizes interferences
among pollutants.  it is anticipated S02 can be detected at 1.5 ppra using
a C02 laser.

Reference:   (138)

-------
                               - 82  -
METALLIC SILVER FILTER MEMBRANE  (H2S)

 Principle of Operation

          Metallic silver reacts with H2S and/or mercaptans to produce
 silver sulfide by the following reaction:

                                    (HO)
                H2S + 2Ag + 1/202  —=	>-  Ag2S + H20


                                    (H70)
               2RSH + 2Ag + 1/202  —-	>•  AgSR + H20
The black silver sulfide formed changes the reflectance of the original
silver surface.  This can be measured by a reflectometer and related to
sulfide content of the ambient gas.  The reflectance of the initial surface
may be compared to a standard magnesium oxide surface as defined by the U.S.
Bureau of Standards.  A membrane of metallic silver reacts quantitatively
with H2S or mercaptans causing a decrease in the reflectivity of the surface.
Each membrane is calibrated before use to the nearest 1/2 reflectance unit
using a reflectance meter equipped with a green tristimulus filter.

Commercial Equipment

          No equipment is available for automatic determination of gaseous
sulfides by this method.  Components are available for reflectance measure-
ments and air sampling, but a complete system has not been devised.

References

       1. Principle of operation:  139, 140, 141. 142, 143

       2. Applications:  139

       3. Data:  139

       4. Specific instrument descriptions:  139

-------
                              - 83 -
Pis cussion

Advantages

       1. Sensitive to very low concentrations of lUS and RSR.

       2. NOX, 03 or ultraviolet light do not interfere, as  they
          do with Pb(Ac>2, HgCl2 and other similar detectors.

       3. Simple, low cost and rapid.

Disadvantages

       1. Mercaptans interfere.

Conclusions

          Needs further testing, since the method  appears quite promising.

Re commendations^

          For the detection of very  low  concentrations  of  mainly H.2S,
the metallic silver membrane  seems appropriate  for  further development.
To detect high concentrations other  methods  probably better suited to
automation.

-------
                              - 84 -


PLASMA CHROMATOGRAPH   (S02, RSH, RSSR)

Principle of Operation

          Gases are detected by conversion to ion-molecules and separated
by drifting through an electical field toward a detector.  The detector
output of ion current versus time presents a plasmagram of the reactant
ion and the different ion  molecules formed. The plasmagram resembles a
chromatogram,with a millisecond time scale.

Commercial Equipment

          None available.

Discussion

          This technique enables measurement of the mass of an organic
molecule with an instrument that operates at atmospheric pressure.  Thus,
it may be possible to use as a simpler instrument than the mass spectrometer
for detailed analyses.  This instrument, with further development may
become a valuable laboratory tool.

References:  144» 145

-------
                               -  85  -
                              BIBLIOGRAPHY


 (1)   Duffee, R.  A.,  JAPCA 18:472-4,  July, 1968.

 (2)   Katz,  S. H. and Talbert, E.  J., U.S. Bur. Mines Tech.  Paper 480 (1930).

 (3)   Ryazanov, V. A., Arch.  Environ. Health ,5:480-94, 1962.

 (4)   Manufacturing Chemists  Assoc.  Research on Chemical Odors.   Research report
      from A. D.  Little, Inc., October 1968.

 (5)   Dravnieks,  A.,  "Theories of  Olfaction", in  Chemistry and Physiology of
      Flavors, ed. H. W. Schultz,  Avi Publishing  Co., Westport,  Conn. (1966).

 (6)   Hendrickson, E. R., Control  of  Atmospheric  Emissions in the Wood Pulping
      Industry, Vol.  3, Final Report  of EPA Contract CPA 22-69-18, 1970.
      NTIS No. PB 190,353.

 (7)   Blosser, R. 0.  and Cooper, H.B.H., Jr., "Compendium of Methods for Measuring
      Ambient Air Quality and Process Emissions", National Council for Air and
      Stream Improvement, Technical  Bulletin No.  38, New York, Dec. 1968.

 (8)   Thoen, G.,  DeHaas, G. G. and Austin, R. R., TAPPI j>l:246, 1968.

 (9)   Adams, D. F., Presentation at  the Ninth Methods Conference in Air Pollution
      and Industrial Hygiene Studies, Pasadena, California.   "Analysis of
      Malodorus Sulfur Containing  Compounds" February 7-9,  (1968).

(10)   Stevens, R. K., O'Keeffe, A. E. and Ortman, G. C., Environ. Sci. & Tech. _3»
      652 (1969).

(11)   Turk,  A., Heating, Piping, Air Cond. _39 207-10, 1968.

(12)   Larsen, R.  I.,  APCA Meeting New York, June  1969, Paper 69-210.

(13)   Mills, J. L., Walsh, R. T.,  Luedtke, K. 0.  and Smith, L. K., JAPCA JL3_:467-75,
      1963.

(14)   Benforado,  D. M., Rotella, W.  J. and Horton, D. L., JAPCA 19_: 105-5, 1969.

(15)   Banks, 0. M., "Problems of Odor Measurement and Control", American
      Petroleum Institute, New York, Oct. 1969.

(16)   Landsberg,  H. and Escher, E. E.,  Ind.  Eng.  Chem.  46_:1422-8,  1954.

(17)   Walther, J.  E.  and  Amberg,  H.  R.,  TAPPI  51:126A-29A (1968).

(18)   Tarman, P.   B.,  Andreen,  B.  H.  and Kniebes, D.  V.,  Chem. Eng. Progress 61(17):l-8
      (1960).                                                               ~~

(19)   Washburn,  H. W.  and Austin, R.  R.  in  "Air  Pollution", ed.  McCabe,
      Chapter  72,  576 McGraw Hill, New York, 1952.

(2°)  Fredericks,  E.  M.  and  Harlow,  G. A.,  Anal. Chem.  36:263,  1969.

-------
                             - 86 -
 (21)  McWilliams, I. G.  and Dewar,  R. A., Nature  181:760,  1958.

 (22)  Adams, D. F., Jensen, G. A.,  Steadman,  J. P.,  Koppe, R.  K.  and
      Robertson, T. J.,  Anal. Chem. _36:1094  (1966).

 (23)  British Columbia Research  Council, "A Manual Describing  a Method  for
      the Collection and Analysis of Pulp Mill Effluents", Vancouver, 1959.

 (24)  Adams, D. F. and Koppe, R. K., TAPPI ^1:366, 1958.

 (25)  Blaedel, W. J. and Laessig, R. H., "Automation of  the Analytical  Process
      through Continuous Analysis", in Advances in Anal. Chem. and  Instrumentation
      2:69-168, J. Wiley, New York, 1966.

 (26)  Kenline, P. A. and Hales,  J.  M., "Air Pollution and  the  Kraft Pulping
      Industry11,, PHS Publication 999-AP-4, 1963.

 (27)  Walther, J. E. and Atnberg, H. R., JAPCA 20(1):9-18,  1970.

 (28)  Schwalbe, C. G., Paper Trade  J. 2i:52»  1922.

 (29)  Colombo, R., Corbetta, D., Pirotta, A.  and  Sartori,  A.,  TAPPI ^0_  490,  1957.

 (30)  Bergstrom, H. and  Trobeck, K. G., Svensk Papper  48_:49-54, 1945;
      see also U.S. Patent 2,406,581.

 (31)  Control Techniques for Sulfur Oxide Air Pollutants,  NAPCA Publication
      No. AP-52, Chapter 5.2.

 (32)  Air Pollution Control Assoc. Petroleum  Committee TI-3, JAPCA ^ (1), 1964.

 (33)  Termeulen, M. A.,  "Air Pollution Control by Oil  Refineries",  First
      International Clean Air Congress, London 1966, Paper IV-5.

 (34)  Palmer, H. F., Rodes, C. E. and Nelson, C. J., JAPCA JL9:778-86, 1969.

 (35)  Tokiwa, Y., "Sulfur Dioxide", paper at  llth Methods  Conference in Air
      Pollution, Berkeley, Cal.,  April 1970.

 (36)  Thomas, M. D. and  Cross, R. J., Ind. Eng. Chem. J20:645,  1928;
      ibid.  Anal. Ed 4_:253-6, 1932 and L5:287-90, 1943.

(37)  Danielson,  J.  A.,  "Air Pollution Engineering Manual", Air Pollution
      Control District,  County of Los Angeles, U.S.  Dept. of HEW  (1967).

(38)  Teller, A.  J.,  APCA 13 No.  4 Pg.  148 (April 1963).

(39)  Ronald, D., "Handbook on Offensive Trades",  Wm. Hodge, London (1935).

(40)  Wood,  T.  and Bender,  A.  E., "Analysis  of Tissue Constituents - Commercial
      Ox-Muscle Extract", 67 (1957).

-------
                              - 87  -
(41)   Burchfield,  H.  P.,  Johnson, D.  E., Rhoades, J. W., and Wheeler,
      R.  J.,  J.  Gas Chromatog. _3:28,  1965.

(42)   Hemeon, W.C.L.,  "Malodors  - A Basis  for  Regulations", JAPCA,  2.1  (12)  Pg.  770
      (Dec.  1971).

(43)   Sullivan,  R. J., "Air Pollution Aspects  of Odorous  Compounds", Technical
      Report by  Environmental Systems, Bethesda, Md.,  (For Nat'l Air Pollution
      Control Administration, USPHS).

(44)   Stahl, Q.  R., "Preliminary Air  Pollution Survey of  Aldehydes",  U.S.
      Dept.  of HEW, Publ. Health Service,  NAPCA, (Oct. 1969).

(45)   Dravnieks, A.,  "Current Status  of Odor Theories", Advances in Chemistry,
      Series #56, Pg. 29 (1966)  ACS.

(46)   Dravnieks, A. and Lafford, P.,  "Possibilities of Predicting Odor Intensity
      and Quality by Means of Chemical Analysis", Conf. on Odorous Pollutants,
      The Karolinska Institute,  Stockholm,  (June, 1970).

(47)   Miller, F., et  al., "Determination of Acetic Acid  in Air", Amer.  Ind.
      Hyg. Assoc., 1£> (1) Pg. 221  (Mar. 1955).

(48)  Holmes, R.  G.,  "Air Pollution  Source  Testing Manual", Los Angeles  Air
      Pollution Control  District,  Los Angeles County,  Calif.


 (49)  Stahl,  W. H.,  "Instrumental  Methods Used  in the Analysis  of  Odorants",
      Proceedings of  the 3rd Annual  Symposium on Problems in  Air Pollution,
      Franklin  Institute (Oct.  1958).

 (50)  Williams, I. H.,  "Gas  Chromatographic Techniques for the  Identification
      of Low Concentrations  of  Atmospheric  Pollutants",  Anal. Chem.  3]_ (13)
      Pg. 1723  (Dec.  1965).

 (51)  Levaggi,  D. A.  and Feldstein,  M.,  "The  Determination of Formaldehyde,
      Acrolein  and Low Molecular Weight Aldehydes  in Industrial Emissions"
      JAPCA 20.  (5) Pg.  312  (May 1970); 19_ (1) Pg. 43-5 (Jan.  1969).

 (52)  Prater, T.  J., "Exhaust Gas  Carbonyl Measurements  by the Gas Chromato-
       graphic 2,4-Dinitrophenylhydrazine Method",  Ford Motor Co.,  Dearborn,
      Mich.

 (53)   Burnett,  W. E., "Air Pollution from Animal Wastes", Environmental Science
       and Technology 3_  (8) Pg.  744  (Aug.  1969).

 (54)   Okita, T., "Filter Method for the Determination of Trace Quantities of
       Amines, Mercaptans and Organic Sulfldes  in the Atmosphere", Atmospheric
       Environment, Pergamon Press 1970 Vol. 4  Pgs. 93-102, Printed in Great
       Britain.

 (55)  Dravnieks, A.  and Krotoszynski, B. K., "Systematization  of  Analytical
       and Odor Data  on  Odorous Air", from  Symposium  Bulletin,  "Odors  and
       Odorants:  The Engineering  View", Amer.  Soc. of Heating, Refrigerating
       and Air  Conditioning  Engineers, Inc.  (ASHRAE), (1968).

-------
                              - 88 -
 (56)  "The Kovats Retention Index System", Anal. Chem. _36  (8) Pg. 31A  (1964).

 (57)  Dravnieks, A. and O'Donnell, Anne, "Principles and Some Techniques of
      High-Resolution Headspace Analysis", J. Agr. Food Chem. JL9  (6) Pg. 1049
      (1971).

 (58)  Trieff, N. M., et  al., "Analysis of Industrial Odor Emissions", 19th
      Pittsburg Conf. on Analytical Chemistry and Applied Spectroscopy, Inc.
      Cleveland, Ohio (Mar. 1968).

 (59)  Dravnieks, A., "Odors:  Their Nature, Measurement, and Control", IIT
      Research Institute, Chicago, 111.

 (60)  Burgwald, T. A., "Identification of Chemical Constituents in Rendering
      Industry Odor Emission", IIT Research Institute Report.


 (61)  Smith, D. 0. and Tauss, R. H., "Modification of a Coulometric Titrator
      for the Monitoring of Gaseous Pulp Mill Effluents", S. Pulp Paper Mfg
      33 #2, 32, 4-6 (1970).

 (62)  Austin, R. R., "Sampling and Analysis of Pulp Mill Gases for Sulfur
      Compounds", TAPPI, 54 #6, 977-80 (1971).

 (63)  Palmer, H. F., Rodes, C. E. and Nelson, C. J., "Performance Characteristics
      of Instrumental Methods for Monitoring Sulfur Dioxide II Field Evaluation",
      JAPCA, 19_, #10, T18-86 (1969).

 (64)  Kesler, R. B., "Modern Methods of Monitoring Gaseous Pulp Mill Effluents
      for Sulfur Compounds", Advan Automat Anal., Technicon Int. Congr.
      2, 111-14 (1969).

 (65)  Austin, R. R., et  al.,  "The Application of Couloraetric Titration to
      Atmospheric H^S and S02 Measurements and to Hydrogen Reforming Measurements",
      Ann u. ISA Conf.  Proc. 2Q_, #3, 2.1-1-65 (1965).

(66)  Bamesberger,  W.  L. and Adams, D. F., "Field Comparison of the Coulometric,
      Colorimetric, and Lead Acetate Tape Analysis Methods for Sulfur Containing
      Gases", TAPPI, 52, #7, 1302-1306 (1969).

(67)  Garber, W. F., et  al.,  Journal WPCF,  42.  #5, 209-220 (1970).

(68)  Thoen, G.  N., et   al., "Continuous Measurement of Sulfur Compounds and
      Their Relationship to Operating Kraft  Mill Black Liquor Furnaces",
      TAPPI, 52, #8, 1485-87 (1969).

(69)  Tretter,  V.  S.,  TAPPI, _52, #12, 2324-26 (1969).

(70)  Adams, D.  F., et  al., "Analysis of Sulfur Containing Gases in the Ambient
      Air Using Selective Pre-Filters and a Microcoulometric Detector",
      JAPCA 18 #3,  145-8 (1968).

 (71)  Andree, B. H., et  al.,  "Instrumental Methods of Analysis for Odorant
      Compounds in Natural Gas", Inst. of Gas Technology, Tech. Report No. 7  (1963).

-------
                             - 89 -
(72)   Nader,  J.  S.  and  Coffey, W.  L.,  "Automatic  Zero Drift  Correction
      of  the  Titrilog",  Am.  Ind Hyg Assoc.  J  2l4_ (8), 563-5  (1963).

(73)   Stevens,  R. K., "Review of Analytical Methods for  the  Measurement  of
      Sulfur  Compounds  in the Atmosphere",  llth Conf. on Methods in Air
      Pollution and Industrial Hygiene Studies, Calif.  State Dept.  of Public
      Health  (1970).

(74)   Adams,  D.  F., et   al., "Analysis of Kraft-Mill, Sulfur Containing  Gases
      with GLC lonization Detectors",  JAPCA !L5_, #1, 31-33  (1965).

(75)   Adams,  D.  F.  and  Koppe, R. K.,  "Direct  GLC Coulometric Analysis of
      Kraft Mill Gases", JAPCA,  1J,  #3, 161-65 (1967).

(76)   Drushel, H.  V., Preprints, Div.  Petr. Chem. ACS,  14,  #3, B232  (1969).

(77)   Drushel, H.  V., Analytical Letters, _3_,  353,  (1970).

(78)   Killer, F.C.A. and Underhill,  K. E., Analyst, £5_, 505  (1970).

(79)   Marsh,  W. W., Jr., Analytical Letters,  .3, 341 (1970),

(80)   Engelhardt,  H., Instr. Rev.  Pg 98-99,  (Feb.  1968).

(81)   Stevens, R. K., et  al.,  "Gas Chromatography of Reactive  Sulfur Gases
      in Air at PPB Level", EPA Publication.

(82)  Mulik,  J. D., et  al., "An Analytical  System Designed  to  Measure
      Malodorous Compounds  Related to  Kraft  Mill Activities" TAPPI  Water
      and Air Conf. Apr. 1971.

(83)  Mulik, J. D., "Continuous Determination  of Sulfur Dioxide with Flame
      Photometric  Detector,  ibid  (1970).

(84)  McGinnity, et al., "Mobile Source  Sampling  Laboratory for the Pulp
      and Paper Industry Presented at  a  Symposium  on Methods for Measuring
      and Evaluating Odorous Air Pollutants  at the Source and in the Ambient
      Air  6/1-5/70 Dept. of HEW, NAPCA,  Durham,  NC.

(85)  Harrison, G.A.F.  and  Coyne, C.  M.,  J.  Chromatogr., 41, #3,4,  45.3-5  (1969).

(86)  Brink, D. L., et  al., "Analysis of Sulfur Containing and Sulfur  Frea
      Organic Products Formed in  Kraft Black Liquor Pyrolysis", TAPPI 54,
      #5,  714-20  (1971).

 (87)  Adams, D. F. and Koppe, R.  K.,  TAPPI,  42,  #7, 6Q1-6Q5 (1959).

 (88)   Siu, W., Levaggi, D.  A.,  et al., "Modifications  to an H2S Tape Sampler for
      Increasing  Sensitivity and  Accuracy in H S Sampling", JAPCA, 21,
       #10, 636-8,  (Oct. 1971).                 z                    —

 (89)  Dravnieks,  A., "Olfactory Information Information Processing and
      Mechanisms", AGARD Symposium  on Bionics, Brussels,  Belgium,  Sept. 1968.

-------
                                - 90 -
  (90)  Applebury, T. E. and  Schaer, M. J.,  JAPCA,  20,  #2,  83-6  (1970).

  (91)  Walther, J. E. and Amberg, H. R.,  "Continuous Monitoring of  Kraft
       Mill Stack Gases with a Process Gas  Chromatograph", TAPPI, 50, #10,
       108A-110A (1967).

  (92)  Adams, D. F., "Analysis of Malodorous Sulfur-Containing Gases",  TAPPI
       51, //I, 53-58 (1969).

  (93)  Craig, J. C., et. al., "Analysis of  Organic Pollutants", Univ. of Calif.,
       Project Clean Air, Res. Rep. 6^ Vll-115, Sept. 1, 1970.

  (94)  Feldstein, M., et. al., "Gas Chromatographic Determination of Mercaptans",
       JAPCA 15, #5, Pg 215-17 (1965).

  (95)  Williams, I. H. and Murray, F. E., "Studies of  the  Gas Chromatographic
       Analysis of Kraft Mill Sulfides Part I Collection and Analysis of
       Gaseous Samples", Pulp Paper Mag Can 67_, #8, T347-52, (1966).

  (96)  Martin, R. L. , Anal.  Chem., _38_, 1209 (1966).

  (97)  Adams, D. F., et. al., TAPPI, 43_,  #6, 602-608 (1960).

  (98)  McKean, W. T., Jr., et. al., TAPPI,  48, #12, 699-704 (1965).

  (99)  "Environmental Measurements" (1965), PHS Publication No. 999-AP-15.

(100)  Morgan, G. B., et.  al., Automation in Analytical Chemistry,  Technicon
       Symposia 1966, Mediad, 1967, p 526-523.

(101)  Jacobs, M. B., "The Analytical Chemistry of Industrial Poisons,  Hazards
       and Solvents", Interscience, N.Y.   (1949).

(102)  Pare,  J.  P.,  JAPCA JL6, #6, 325-7 (1966).

(103)  Budd,  J.  and Bewick, H. A., Anal.  Chem 2_3, 1536 (1952).

(104)  Marbach,  E.  P.  and Doty,  D. M.,  J. Agr and Food Chem., 4_, 881 (1956).

(105)  Hochheiser,  S.  and Elfers, L. A.,  "Automatic Sequential Sampling of
       Atmospheric  H^S by Chemisorption on Mercuric Chloride-Treated Paper Tape",
       Environ.  Sci. Technol. _4,  #8, 672-6  (1970).

(106)  Sensenbaugh,  J.  D.  and Hemeon,  W.C.L.,  Air Repair 4_, 1 (1954).

(107)  Shikiya,  J.  M.  and MacPhee, R.  D., "Multi-Instrument Performance
       Evaluation of Conductivity-Type Sulfur  Dioxide Analyzers", JAPCA 19,
       #12,  943-45  (1969).

(108)  Intersociety Committee,  Health Laboratory Sci,  _7_, #4 (Oct. 1970).

(109)  Risk,  J.  B.  and Murray,  F. E.,  Can. Pulp and Paper Ind., 1_7_  #10, 31-34 (1964)

-------
                                    -  91  -
(110)   Oklta,  T.,  et.  al.,  "Filter Method for  the Measurement  of  Atmospheric
       Hydrogen Sulfide", Environ. Sci.  Technol 5_, #6,  532-4 (1971).

(Ill)   Lodge,  J. P.,  et.  al.,  Amer Ind Hyg.  Assoc. J,,  24,  380 (1963).

(112)   Sanderson,  H.  P.,  et.  al.,  "Limitations of the Lead  Acetate Impregnated
       Paper Tape  Method  for  Hydrogen Sulfide", JAPCA 1£, #6,  328-30 (1966).

(113)   High, M. D. and Horstman,  S.  W.,  "Field Experience  in Measuring Hydrogen
       Sulfide", Am Ind Hyg Assoc J. 2£, #4, 366-73 (1965).

(114)   Intersociety Committee, Health Laboratory Sci. £, #4, (Oct. 1969).

(115)   Adams, D. F., "Tentative Method of Analysis for Hydrogen Sulfide Content
       of the Atmosphere",  Health Lab Sci _7_, *3, 157-63 (1970).

(116)   Bostrom, C. E. and Brosset, C., "A Method for Simultaneous Determination
       of H2S and  S02 in Flue Gases", in "Atmospheric Environment" Pergamon Press
       Vol 2» p 407-16 (1969) Great Britain.

(117)   Gaeke, G. C. and West, P. W., Anal Chem 28, 1816-19  (1956).

(118)   Sliwinske,  R. A. and Doty, D. M., Agr. Food Chem _6,   41  (1958).

(119)   Jacobs, M.  B., et. al., Anal. Chem. 29, 1349  (1957).

(120)   Mecklenburg, W. and Rozenkranzer, R., A. Anorg Chem  86, 143  (1914).

(121)  Almy, L. H., J Am Chem Soc., £7,  1381  (1925).

(122)  Sheppard,  S. E. and Hudson,  J. H., Ind. Eng.  Chem.,  Anal.  Ed. 2_,  73 (1930).

(123)  Moore, H., Helwig, H. H. and Graul, R.  J.,  Ind. Hyg. J. 21,  466  (1960).

(124)  Bamesberger, W. L. and Adams, D.  F., Environ. Sci and  Tech. _3_, 258  (1969).

(125)  Stern, A.  C.  (Editor), "Air  Pollution", Vol II  (1968),  Academic  Press,  N.Y.,
       Chap. 17:  "Analysis of Inorganic Gaseous  Pollutants".

(126)  Bulletin No.  101, RPC  Corporation, El  Segundo,  Calif.

(127)  Hodgeson,  J. A., et. al., "Application of Chemiluminescence Detector  for
       the  Measurement of  Total Oxides  of Nitrogen and Ammonia in the Atmosphere",
       EPA  Tech.  Center.

(128)  Hodgeson,  J.  A., et. al., Anal.  Chem.  43, 1123 (1971).

(129)  Parts,  L.  P.,  et. al.,  "Instrumentation for the Determination of Nitrogen
       Oxides  Content of Stationary Source  Emissions", Interim Report Contract No.
       EHSD 71-30 for Air  Pollution Control Office of EPA, Durham, N.C.

 (130)  Fontijn, A. and Ronco,  R.  J.,  Aerochem Research Labs, Princeton, N.J.,
       EPA, APCO  Contract  No.  70-79 (1971).

-------
                                  - 92 -
 (131)  Ballard, L. F. and Decker, C., Research Triangle Institute,  Research
       Triangle Park, N.C., EPA Contract No.  70-101 (1971).

 (132)  Stedman, D. H., et. al., Publication Preprint,  Ford Research,  Dearborn,
       Michigan.

 (133)  Comberiati, J. R., "Application of  Rapid Spectrometry to Air Pollution
       Research", Anal Chem 43, #11, 1497-9 (Sept.  1971).

 (134)  Wronski, M., Z. Anal Chem. 180, 185 (1961).

 (135)  Andrew, T. R. and Nichols, P.N.R.,  "The Determination of Hydrogen Sulfide
       in the Atmosphere", Analyst 90 (1071)  367-70 (1965).

 (136)  Hardwick, B. A., et. al., "A Fluorometric Method for  Determining Low
       Concentrations of Hydrogen Sulfide  in  Air",  Atmos.  Environ 4^  #4, 379-85 (1970)

 (137)  Thoen, G. N. and Nicholson, D. C.,  "infrared Analysis of Kraft Pulping
       Process Gases", TAPPI _53_, #2, 224-6 (1970).

 (138)  tianst, P. L. and Morreal, J. A., "Detection  and Measurement  of Air Pollutants
       by Absorptions of Infrared Radiation",  JAPCA, 18_, #11,  754-9 (1968).

 (139!)  Falgout, D. A. and Harding, C. I.,  "Determination of  H2S Exposure by
       Dynamic Sampling with Metallic Silver  Filters",  JAPCA 18_, #1,  15-20 (1968).

 (140)  Drott, J., Arkiv for Kemi, 15, 14 (1960).

 (141)  Sullivan, J. L., "An Investigation  of  Air Pollution Problems in the
       South Auckland Area", R. E. Owens,  Govt.  Printer, Wellington,  New Zealand.

 (142)  Lilienfield, S. and White, C. E., JACS .52_, 886  (1930).

 (143)  Ricca, P. M., "The Effect of Temperature and Humidity on the Odor
       Threshold of Hydrogen Sulfide Gas",  Masters'  Thesis,  Univ. of  Florida.

(144)  Karasek, F. W., "A Drift-Mass Spectrometer",  Research Development
       Pg. 26 (Dec. 1970).

(145)  Karasek, F. W., Research Development 21_,  34  (1970).

(146)  Bialkowsky, H. W., DeHaas, G. G., "Catalytic Oxidation Procedure for
       Determining Sulfur Compounds in Kraft  Mill Gases", Pulp  and  Paper Mag.
       Canada 53 (11) 100-(1952).

(147)  Leithe, W., "The Analysis of Air Pollutants"  (Stuttgart  1968,  Transl.
       1970 by Ann Arbor-Humphrey Science Publishers,  Inc.)


                    ftd.S. Govranunt Printing Office: 1972—746-763/4109 Region Ho. 4

-------