United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
Reeaareh and Development
Method for
Determining
Potential  Odor
Contribution of
Selected  Kraft
Process Streams
EPA -600/2-79-117
June 1979'

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further deveJopment and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.  Environmental Health  Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

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

-------
                                     EPA-600/2-79-117
                                     June 1979
METHOD FOR DETERMINING POTENTIAL ODOR CONTRIBUTION
        OF SELECTED KRAFT PROCESS STREAMS
                       by
              Michael  E.  Franklin
                Andre  L.  Caron
    National Council of the  Paper  Industry
     for Air and Stream Improvement,  Inc.
          New York, New York   10016
              Grant No. R-804646
               Project Officers

       Donald L. Wilson, H. Kirk Willard
             and Victor J. DalIons
     Industrial Pollution Control Division
 Industrial Environmental Research  Laboratory
           Cincinnati, Ohio   45268
 INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
           CINCINNATI, OHIO    45268

-------
                                 DISCLAIMER


     This report has been reviewed by the Industrial Environmental Research
Laboratory - Cincinnati, U.S. Environmental Protection Agency, and approved
for publication.  Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial  products constitute endorse-
ment or recommendation for use.

-------
                           FOREWORD
     When energy and material resources are extracted, processed,
converted, and used, the related pollutional impacts on our en-
vironment and even on our health often require that new and in-
creasingly more efficient pollution control methods be used.
The Industrial Environmental Research Laboratory-Cincinnati (IERL-
Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and
economically.

     The subject of this report is to define the potential odor
contribution of selected kraft process streams that can be rou-
tinely sewered.  During the course of this study a method was
developed for determining the identifiable odor threshold of
liquid samples which appears to have utility in problem analysis
at the mill level.  The potential odor contribution of effluent
treatment systems was identified as a concern in the development
of total reduced sulfur emission standards for new sources in
the kraft industry.  Before decisions can be made regarding the
use of information generated in this study, some maximum accep-
table level for these odors should be established.  The benefits
of reducing the odor threshold level of individual process efflu-
ents should be determined.  If significant benefits are found,
then the search for alternate control technology for replacement
of the energy-intensive stripping procedures currently used
should be initiated.  For further information, please contact
the Food and Wood Products Branch of the Industrial Environmental
Research Laboratory, Cincinnati, Ohio.

                                David G. Stephan
                                    Director
                  Industrial Environmental Research Laboratory
                                   Cincinnati
                              111

-------
                           ABSTRACT
     The objective of this project was to define the potential
odor contribution of selected process streams and mixtures of
liquid process streams in the kraft industry that are routinely
sewered.  A procedure was suggested that can be used for this
purpose.

     Use of a dynamic olfactometer and odor panels to measure
odor thresholds determined by complete volatilization of the sam-
ple or stripping of the sample were unsuccessful.  No correlation
between odor threshold and reduced sulfur concentration in the
gas stream as measured by gas chromatographic techniques could
be obtained.

     Odor panels were employed using the head space analysis and
the forced-choice triangle technique.  It was shown that identi-
fiable odor threshold values were more reproducible and judged
more meaningful than simple odor threshold values.

     The group of process streams including white liquor, green
liquor, black liquor, and weak wash all had high pH and sulfide
concentrations.  They yielded the highest identifiable odor
thresholds of any other group of streams investigated with EDg0
(effective dosage at 50 percent level) values varying from
<2.6xlo6 to 6.3x10?.  This emphasized the need to keep these in
the process and to prevent their loss to the sewer.   Condensate
streams, including those from the multiple-effect evaporators,
digesters, and the turpentine decanter underflow also had high
     values, ranging from 3.1xl04 to >1.4xl08.
     The identifiable odor threshold of the condensate stream to
the steam strippers studied had a log average of 3-OxlO6 and the
stripper product samples had an identifiable odor threshold of
1.4xl04, a reduction in the odor level by a factor of about 200.
These were in the same general range as the odor thresholds mea-
sured on biologically treated kraft mill effluents.

     It was shown that independent of mixing techniques, odor
intensities of kraft mill process streams were additive.  This
was demonstrated for an acid sewer containing first chlorination
stage effluent, digester condensates, and multiple-effect evapo-
rator condensates, and for a total mill effluent with multiple-
                               IV

-------
 effect  evaporator  condensates,  decker water,  and odor-free dilu-
 tion  water  used  as make-up.

      This report is submitted  in fulfillment  of Grant No.  R-
 804646  by the  National  Council  of the Paper Industry for Air and
 Stream  Improvement,  Inc.  under  the sponsorship of the U.S.
-Environmental  Protection  Agency.  This work was performed  during
 the period  of  July 20,  1976  to  June 19,  1978.
                                 v

-------
                            CONTENTS
Foreword	iii
Abstract	iv
Figures	   ix
Tables 	    x
Abbreviations	xii

     1.  Introduction  	  1
     2.  Conclusions   	  3
     3.  Recommendations 	  5
     4.  Literature Review 	  6
          Odor Regulations 	  6
          General Precautions in Determining Odor
            Threshold	7
          Characteristics of Odor	8
          Physiological Aspects Affecting Response to Odor . 11
          Factors for Consideration in Panel Selection ... 12
          Odor Panel Management	14
          Reproducibility of Individual Panelist Response. . 14
          Combinations and Alternate Odor Analysis
            Techniques	17
          Kraft Mill Odor Identification and Threshold ... 18
     5.  Experimental Methods	23
          Sample Collection and Storage	23
          Human Odor Panel Selection	24
          Sample Preparation for the Total Volatilization
            Technique	24
          Sample Preparation for the Stripping Technique . . 24
          Sample Preparation for the Head Space Analysis
            Technique	26
          Preparation of the Calibration Gases 	 26
          Odor Threshold Determination by Dynamic
            Olfactometer 	 27
          Odor Threshold Determination by Head Space
            Analysis	34
          Chemical Analysis by Gas Chromatography	35
     6.  Discussion of Results	38
          Odor Thresholds of Individual Reduced Sulfur
            Gases by Olfactometer	38
          Odor Thresholds of Combined Reduced Sulfur Gases
            by Olfactometer	38
          Odor Threshold of Selected Streams by the Total
            Odor Volatilization Technique	39
                               VI1

-------
    6.   Discussion of Results (continued)
         Odor Threshold of Selected Streams by the
           Olfactometer and Stripping Technique  	  40
         Relationship of Measured Reduced  Sulfur
           Concentrations to the Odor Thresholds	41
         Preliminary Odor Threshold Measurement by  the
           Head Space Analysis Technique 	  42
         The Effects of Initial Stock Solution Concentra-
           tion and Odor Stability on the  Detectable Odor
           Level in the Head Space Analysis	42
         Reasons for Choosing Identifiable Odor Level
           Over Detectable Odor Level in the Head Space
           Analysis Technique	45
         Number of Dilution Stages Effect  on the
           Identifiable Odor Threshold 	  45
         Stock Dilution Level Effect on the Identifiable
           Odor Threshold	47
         Results of Duplicate Samples on Identifiable
           Odor Thresholds	47
         Storage Stability of Odor in Sample Bottle	49
         Summary of Head Space Analysis Procedure
           Evaluation	53
         Identifiable Odor Thresholds of Process Liquor
           Streams	53
         Identifiable Odor Threshold of Other Process
           Streams	55
         Summary of Process Stream Identifiable Odor
           Thresholds	75
         Additive Nature of Effluents	79
         Summary	82

References	83
                             Vlll

-------
                              FIGURES


Number                                                    Page

 1   Total odor volatilization technique schematic .... 25

 2   Strippable odor technique schematic 	 25

 3   Forced-choice triangle dynamic olfactometer 	 27

 4   Olfactometer flow data sheet	29

 5   Data collection and calculation sheet for
       forced-choice triangle test 	 30

 6   Plotting value vs. log (tolerance level)  	 32

 7   Chromatographic system schematic  	 35

 8   Operation of 8-port sample valve  	 36
                               IX

-------
                               TABLES
Number                                                      page

 1   Odor Threshold of Various Reduced Sulfur Gases as
       Reported in the Recent Literature 	 18

 2   Odour and Taste Thresholds* of Point Sources	20

 3   Relative Odour and Taste Contributions of
       Point Sources	21

 4   Odor Thresholds for Reduced Sulfur Gases Determined
       with Olfactometer	38

 5   Additive Nature of Pure Sulfide Gases on Olfactory
       Response	39

 6   Total Odor Technique:  Odor Unit Range	40

 7   Stripping Technique: Odor Unit Range	41

 8   Preliminary Head Space Technique: Odor Unit Range
       by Source *	'	42

 9   Effect of Stock Solution Dilution Level on Odor
       Threshold Values as Determined for MEE Surface
       Condenser Condensate from Mill C by Panels of Six .  .44

10   Effect of Dilution Stages on Identifiable Odor	46

11   Effect of Stock Solution Dilution Level on Identified
       Odor Threshold Values as Determined for Treated
       Total Mill Effluent from Mill C	48

12   Duplicate Sample Identifiable Odor Determinations
       on the Same Date	50

13   Storage Stability Studies on Mill E Samples 	 52

14   Identifiable Odor Threshold of In-Process Streams  ... 54

15   Identifiable Odor Threshold of Non-Condensible Line
       Condensate	54
                                x

-------
Number                                                      Page

16   Identifiable Odor Threshold of Turpentine Decanter
       Underflow	57

17   Identifiable Odor Threshold of Reboiler Condensate. .  . 58

18   Identifiable Odor Threshold of Decker Filtrate	59

19   Identifiable Odor Threshold of Forced Circulation
       Concentrator Condensates	61

20   Identifiable Odor Threshold of Combined Multiple-Effect
       Evaporator Condensates	62

21   Identifiable Odor Threshold of Multiple-Effect
       Evaporator Condensates from MillA	63

22   Identifiable Odor Threshold of Multiple-Effect
       Evaporator Condensates from Mill B	63

23   Identifiable Odor Threshold of Multiple-Effect
       Condensates from Mill D	64

24   Identifiable Odor Threshold of Multiple-Effect
       Evaporator Condensates from MillD	67

25   Identifiable Odor Threshold of Multiple-Effect
       Evaporator Condensates from Mi HE	68

26   Identifiable Odor Threshold of Multiple-Effect
       Evaporator Condensates from Mill 1	69

27   Identifiable Odor Threshold of Digester Condensate. .   . 71

28   Identifiable Odor Threshold of Chlorination Stage
       Filtrate	73

29   Identifiable Odor Threshold of Bleach Plant Acid
       Sewer Combined with Condensates from MEE and
       Digesters at Mill C	74

30   Identifiable Odor Threshold of Stripped Feed	75

31   Identifiable Odor Threshold of Air and Stream
       Stripped Material 	 77

32   Summary of Identifiable Odor Thresholds for Sources
       Evaluated	78

33   Additive Nature of Odor from Mill Process Streams  ... 80
                               XI

-------
                LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
ASB
b
BSW
°C
CFM
cm
comp.
cond.
dec.
dil.

ED50
elim.
evap.
FPD
ft
gal
GC
GLC
ID
ident ,
1
Ibs
liq.
max.
MEE
min.
ml
mm
NCASI

no.
NSSC
OT
panel .
ppm
ppt
-aerated  stabilization basin
-regression coefficient (y-intercept)
-brown stock washer
-degrees  centigrade
-cubic feet per minute
-centimeter
-composite
-condensate, condensible
-decanter
-dilution
-effective dosage at 50 percent level

-eliminator
-evaporator
-flame photometric detector
-feet
-gallon
-gas  chromatography
-gas-liquid chromatograph
-inside diameter
-identification
-liter
-pounds
-liquor
-regression coefficient (slope)
-cubic meters
-maximum
-multiple-effect evaporator
-minute or minimum
-milliliter
-millimeter
-National Council of the Paper Industry for Air and
 Stream Improvement, Inc.
•number
-neutral  sulfite semi-chemical
-odor  threshold
-panelists
-parts per billion
-parts per million
-parts per trillion
                               XII

-------
QD        —dilution air flow

QS        —sample gas flow
 2
r         —correlation coefficient
RD        —dilution ratio

SC        —surface condenser
S         —standard deviation of  "x" values
 jfX
S         —standard deviation of  "y" values
T         —ton
THE       —total mill effluent
US        —United States

SYMBOLS

CH_CH2SH  —ethyl mercaptan

CH3SCH3   —dimethyl sulfide

 [CH-SCH.,]—concentration of dimethyl sulfide

CH^SH     —methyl mercaptan

 [CH_SH]   —concentration of methyl mercaptan

CH3SSCH3  —dimethyl disulfide

 [CH-.SSCH.,]—concentration of dimethyl disulfide

C12       —chlorination

COS       —carbonyl sulfide
H2S       —hydrogen sulfide
 [H,jS]     —concentration of hydrogen sulfide

m         —millimicron
          —odor threshold of dimethyl  sulfide
         -
n         —odor threshold of methyl mercaptan
01CH3SH
          —odor threshold of dimethyl  disulfide
          o

  .J  .,     —odor threshold of hydrogen  sulfide
  n —fa

S02       —sulfur dioxide
S         —summation
 l        —microliter
                              Xlil

-------
                           SECTION 1

                         INTRODUCTION
     The potential odor contribution of effluent treatment sys-
tems was identified as a concern in the development of total re-
duced sulfur emission standards for new sources in the kraft in-
dustry.  Observations that have been made regarding the compara-
tive odor level of feed and product of selected process streams
that have been stripped for reduction of BOD contributing mater-
ials as an alternate means of reducing effluent load, let to
consideration of requiring steam stripping of condensates as an
integral part of these emission standards.

     At that time several information gaps existed regarding the
use of energy intensive measures such as stripping.  Just how
much odor could be tolerated from a biological treatment system
had not yet been established.  Perhaps the levels generated with
BPCT treatment are adequate for most current situations.   Cer-
tainly this was not the case where anerobic conditions existed
in earlier forms of treatment.  Neither was there any demonstra-
ted evidence that any change in odor level in the vicinity of
treatment plants occurs as the result of the use of strippers on
process streams such as condensates from, digestion and liquor
concentration.  This information gap still exists.

     If, however, odor from liquid process streams is to be han-
dled within the regulatory framework the energy use of practices
such as stripping is intensive, and its general application to
streams in a random manner without regard to their odor thresh-
old is not practical.  More extensive information regarding the
odor contribution potential of individual process streams is re-
quired before sound engineering judgment can be made regarding
(a) selective reuse of process streams within the operation as an
odor control measure, and (b) the extent of stripping or other
control technology applications can be defined which optimize
energy use and effective odor control.

     It was the objective of this project to define the odor
threshold of selected process streams and mixtures of process
streams that are routinely sewered.  In carrying out the project,
selection of suitable sample preparation and procedures for pre-
sentation of liquid samples to human panels for odor threshold
measurements represented a significant effort and greater than
initially envisioned.  The method ultimately chosen relied on

-------
determination of an identifiable odor above a previously agitated
sample in preference to definition of the detectable odor level
which was not found suitable for an array of sample preparation
methods.  Chromatographic analysis for sulfur compounds in the
liquid streams was also made to determine if this measurement
could be correlated with the odor level determined by a panel of
human observers.  Information on odor threshold of selected pro-
cess streams, components and mixtures of process streams was then
generated.

-------
                          SECTION 2

                         CONCLUSIONS
     Three methods of sample preparation and two methods of
sample presentation to panelists were investigated during the
course of this study.

     One sample preparation method involved suspension of a
liquid sample in an air stream.  Human observers were then used
to determine the detectable odor level of the gas stream.  There
was not significant difference in the measured odor threshold of
an array of effluent streams of divergent odor quality when using
this procedure.  These findings suggested that compounds were
present in these streams which altered the detectable level of
the presence of the liquid stream but which are not normally
volatilized with the degree of turbulence encountered in sewers
and treatment plants.  The use of this sample preparation method
was discontinued.

     A second sample preparation method involved stripping of the
liquid sample and measurement of the detectable odor level of the
off-gas.  Like the total volatilization method, there was no
significant difference in detectable odor level of samples of
divergent odor quality.  The use of this sample preparation and
presentation method was dropped for the same reasons that use of
the total volatilization method was discontinued.

     The method used for the bulk of the data generation was head
space analysis of a sample agitated just prior to presentation
to a panelist for identifiable odor threshold.  Forced triangle
procedures were used in making the panelists choice.  This method
differed in that the identifiable odor detection principle was
used in contrast to use of the identification of a detectable
change used in earlier portions of the investigation.  Odor thresh-
old values using this procedure were found to be reproducible and
differed widely between streams of varying odor quality.

     During the course of the investigation, information was
generated on the odor threshold of the reduced sulfur compounds
hydrogen sulfide, methyl mercaptan, dimethyl sulfide and dimethyl
disulfide.  The odor threshold for these compounds was found to
be generally consistent with those determined in recent years and

-------
reported in the literature.  Variation in observed  detectable
level between individuals and from day to day for the  same
individual was evident, however.

     No relationship between the level of reduced sulfur com-
pounds in the liquid effluent samples and the detectable odor
level using the total volatilization method could be determined.

     The identifiable odor threshold of mixtures of kraft mill
process streams was found to be calculable if the odor threshold
of the component streams was known.  The exceptions to this  were
mixtures where compounds of a mixture reacted, such as residual
chlorine in first stage bleaching effluent with odorous compounds
in other streams, resulting in a reduction in odor  threshold.

     The identifiable odor threshold of an array of process
streams was generated.  Those with the highest identifiable  odor
threshold were in-process streams not normally entering the
sewers, such as white and green liquor.

     Those condensate streams generated from digestion,  multiple-
effect evaporation of black liquor, and turpentine  recovery  were
found to have high identifiable odor threshold levels.   The
actual level reflected the amount of dilution provided in the
condensation step.  Important in this phase of the  study was the
observation that the bulk of the odor in multiple-effect evapo-
rator condensates is found in the condensates from  evaporators
containing the bulk of the BOD.  These are the first liquor
evaporation stages.

     During the course of this study a method was developed  for
determining the identifiable odor threshold of liquid  samples
which appears to have utility in problem analysis at the mill
level.  It is likely that the odor threshold measurements using
human panelists would have to be carried out away from the mill
site.  This is possible since odor threshold sample stability
of stored samples was demonstrated.  The findings of this study
demonstrate that generic descriptions of process liquids are not
generally adequate for problem definition for individual mill
situations and their odor threshold level should be determined
on a case by case basis.

-------
                         SECTION 3

                       RECOMMENDATIONS
     This study was concerned with only one element of at least a
three-element information generation need.  It dealt specifically
with measurement of the odor threshold of process streams and
effluents.

     Before decisions can be made regarding the use of informa-
tion generated in this study or that generated at individual mill
sites, the benefits, if any, of reducing the odor threshold level
of individual process effluents on odor generation at state-of-
the-art effluent treatment plants should be determined.

     If significant benefits are found, then the search for alter-
native control technology for replacement of the energy-intensive
stripping procedures currently used should be initiated.

-------
                          SECTION 4

                      LITERATURE REVIEW
ODOR REGULATIONS

     Regulations that address odors and their control are general
and elements of suitable regulatory approaches are frequently
being reviewed.  Historically, odor has been considered as a
nuisance.  Black's Law Dictionary defines a nuisance as:

     A class of wrongs that arise from the unreasonable,
     unwarrantable, or unlawful use by a person of his own
     property, either real or personal, or from his own
     improper, indecent, or unlawful personal conduct, work-
     ing an obstruction or an injury to the right of another
     or of the public, and producing such material annoyance,
     inconvenience, discomfort, or hurt that the law will
     presume resulting damage. (1).

     Cheremisinoff and Young  (1) stated,

     Nuisance actions boil down to balancing the equities
     between property owners, each asserting his own rights
     to use his land.  Factors that may be relevant in the
     case, however, include the availability of pollution
     control devices (2).  Failure to keep pace with the
     technological advances in pollution controls (as was
     done by a competitor) resulted in a court ordering adop-
     tion of such controls.  Of course, in any balancing of
     the equities the good faith efforts of the polluter,
     while not absolving him in any way, if they exist, would
     certainly be a factor if they were absent.

     The nuisance theory approach is, in essence, that taken by
the Texas Air Pollution Control Board for the regulation of emis-
sions from odorous sources (3).  Under this regulation, the mere
presence in the atmosphere of an odorant is not considered as
sufficient to prove a violation.  The odor must be of such a
concentration and of such duration so as to interfere with what
is considered to be normal use and enjoyment of the quality of
the property affected.  Every effort is made to obtain a resolu-
tion on a voluntary basis.  If this cannot be achieved, court
action is taken.

-------
     The Bay Area Air Pollution Control District has adopted  a
finite procedure to establish whether or not an odor exists  (4).
The regulation incorporates several novel features.  Prior to an
action being taken, ten complaints must be received from ten
individual complainants over a 90 day period.  Samples are then
collected of the emission itself or in the ambient air beyond the
property line.  The samples are presented to a panel of three
judges using a dynamic olfactometer.  Confirmation by two of  the
three panelists that an odor can still be perceived after appro-
priate dilution is an indication of violation of the regulation.
Dilution of samples for analysis collected from an emission point
is based on a predetermined factor based on stack height.  Dilu-
tion of samples collected from the ambient beyond the property
line is by a standard 4 to 1 ratio.

GENERAL PRECAUTIONS IN DETERMINING ODOR THRESHOLD

     As will be noted in subsequent sections, there are certain
general precautions that must be taken when utilizing either
human odor panels to determine odor thresholds or instrumentation
to estimate odor thresholds.

     Two publications (5,6) discussed methods of measuring odors.
The first article (5) discussed instrumentation, primarily gas
chromatography, and the use of humans as detectors of odorous
compounds emerging from a chromatographic column.  The objective
was to associate the odor with the chemical compounds responsible
for causing the odor.  This procedure was referred to as develop-
ment of an odorgram.

     The second article (6) discussed primarily human receptors
in some depth.  Odorgrams were mentioned, again utilizing human
panelists to identify odors as they are eluted from a chromato-
graphic column.  In addition, the selection of odor panels was
discussed.  It was suggested that persons of low, high, and aver-
age sensitivity to odors all be represented on the panel.  The
concept of ED_Q (effective dosage at 50 percent level) was
mentioned.  Tnis value represents that concentration at which
50 percent of the panelists could perceive the odor in the manner
presented and 50 percent could not.

     Duffee (7) stated:

     Not all of the problems of threshold determination are
     restricted to analytical inadequacies.  Sensitivity to
     odorants varies widely, both within an individual from
     time to time and among different subjects.  Exposure to
     suprathreshold odorant concentrations tends to lower
     olfactory sensitivity for that odorant.  Also, odorants
     are usually mixtures of several compounds whose odors
     the nose tends to synthesize into a single response with
     an intensity that cannot be predicted from a simple

-------
     addition of threshold values of the compounds.   On the
     other hand, analytical instrumentation such as  the GLC
     or infrared spectrophotometer tend to fractionate the
     mixture into its components.  For these reasons it is
     not surprising that reported odor threshold values vary
     by as much as six orders of magnitude for the same
     compound,  e.g.,  5.1xlO~  ppm to 5.1xlO~  ppm for methyl
     mercaptan (8).  Thus, we must rely on organoleptic
     techniques for most odor measurements."

     One investigator indicated (9):

     The most logical approach to the determination  of odor
     detection thresholds or 'odor units'  associated with
     sources emitting complex mixtures of odorous gases is
     to evaluate the dose-response relationship of the source
     gas itself using human subjects, rather than to rely
     upon threshold data obtained from pure compounds or
     synthetic mixtures since neither of these two latter
     procedures will reproduce accurately the emitted odor
     quality of the several sources within the kraft process.

     In the manual edited by Stern (10), it was reported to be
difficult, if not impossible, to translate detection limits for
single compounds prepared in a clean background to predicted odor
detection thresholds in real world situations involving mixtures
of both known and unidentified odorous compounds in  a variety of
background gases.  The author questioned the use of  synthetic
mixtures of several gases because of the possible odor contri-
bution from reported exotic sulfur containing cyclic or hetero-
cyclic compounds in certain kraft process emissions.

CHARACTERISTICS OF ODOR

Definitions of Odor Characteristics

     Huey (11) quoted the American Association of Heating,
Refrigerating, and Air Conditioning Engineers Handbook of
Fundamentals when he defined the sense of smell and odor as:

     The sense of smell is one of the five senses more or
     less involved in man's survival.  The stimulus sensed
     (odor) has been used and is based incidentally or
     intentionally in the search for and enjoyment of food,
     to stimulate emotions, to judge the health or safety
     qualities of places, food, air and water.  At appro-
     priate concentrations, odors may be good or bad, that
     is to have a good or bad effect; at excessive con-
     centrations, they are uniformly bad.

     Odorousness is defined as that property of a substance
     which excites the sense of smell.  To be odorous, a

-------
     substance is usually in a gaseous or vapor state, or
     possesses a vapor pressure.  Some odorants are pleasant,
     others unpleasant, depending on their psychological and
     sociological associations.

     Odorants in themselves are not the cause of organic
     disease.  The discomfort and disagreeableness that may
     be brought about by obnoxious odorants, however, may
     cause some temporary ill effects.  The effects that
     fringe upon ill health include lowered appetite,
     lowered water consumption, impaired respiration, nausea,
     vomiting, and insomnia.

     Odor was defined by Dravnieks (12) as four-dimensional,
containing the characteristics of intensity, detectability,
acceptability, and quality.  Odor intensity was described by the
mathematical equation:

                I = kSn
          Where I = perceived intensity
                S = intensity of the stimulus
          k and n = numerical coefficients specific to the order,

     The value of n tends to vary with the compound and ranges
between 0.2 and 0.8 for sulfur gases.  As an example, n-butanol
has an n = 0.63, while the dimethyl sulfide n value is 0.48.
There are indications that combinations of compounds exhibit
smaller values of n than do single compounds.  It further states:

          The human chemical sense cannot distinguish, by
     odor intensity, two concentrations of the same odorant
     if these concentrations differ by less than 15 to 30%
     (Weber-Fechner's law).  For n-butylmercaptan and
     t-butylmercaptan, the differential threshold of 30% has
     been reported (13).  Hence, an odorous effluent control
     which has reduced the odorous emission by only 15 to 30%
     will be barely observable even if the equally diluted
     'before1 and 'after1 samples were available for a direct
     comparison, as long as the odor is still perceivable.

          The power law, applied, for example, to methyl
     sulfide with n = 0.48, indicated that a reduction by
     75% (concentration to become 4 times lower) will reduce
     the odor intensity by only a factor of 2.  If the
     intensity is still perceivable, such reduction is not
     evident to most observers by memory.  Much more efficient
     reduction in the odorant emission is needed to produce a
     well-noticeable odor reduction effect, especially if
     odorants such as thiophenol and ethylsulfide, with even
     smaller n values, occur in the effluent.

-------
     Cheremisinoff and Young  (1) characterized odors dealing with
the phenomenon of human perception.  They identified adaptation
occurring when small changes  in the environment affect behavior,
indicating the changes can be pleasant or unpleasant, and that
behavior modifications occur  unconsciously to accommodate these
changes when they are of minor nature.  If an odor produces a
certain stimuli of the sense  of smell and the sensitivity of the
sense of smell is reduced during the next stimulus of the same
odor, adaptation has occurred, usually when smelling a single
odor.  If a second odor is introduced, adaptation will not be as
severe as when smelling the single odor.  Other terminology men-
tioned included anosmia, which is the loss of olfactory  sensitiv-
ity, hyposmia, which is a partial loss of olfactory sensitivity
and parosmia, which is a distorted olfactory perception,
carsosmia if the distortions  are unpleasant.

Relationship of Odor Characteristics to Perceived Odor

     NCASI Atmospheric Quality Improvement Technical Bulletin
No. 54 (12) was prepared by Dravnieks of the Illinois Institute
of Technology Research Institute to provide a basis for  the
interpretation of odor panel  findings derived from comparison
of pre- and post-control process emission odor levels.   The infor-
mation presented on the relation between chemical composition and
concentration, and odor level for a number of odorous sulfur com-
pounds is particularly valuable in this respect.  The extensive
bibliography included in this report should prove useful to those
endeavoring to improve their  understanding of the fundamentals of
this aspect of atmospheric pollution.

     Cheremisinoff and Young  (I) edited a book on odor which also
contains an extensive bibliography on the subject.  They reported
on the work of Berglund, Berglund, and Lindvall (14) in  which the
principal of odor interaction was explored.  The results indicated
that the odor intensity of a  mixture of odorous compounds was
directly proportional to the  arithmetic sum of individual inten-
sities of the components.  The concept presented suggested that
individual odors are independent but that odor intensities are
additive similar to light or  sound intensities.

     Cheremisinoff and Young  concluded the following from
additional work reported by Myddleton (1,15):

     "1.   Intensity of a mixture is about the same as the average
          of the component odors.
     2.   Pleasantness of a mixture is about the same as the
          average of the component odors.
     3.   Mixture nomenclature or response is the same as for
          single odorants.
     4.   No new qualities develop as a result of the mixture;
          that is the qualities of the mixture are determined
          by the qualities of the component odors.  Note that

                               10

-------
          this is a contradiction of Hartley  (16), where he
          finds that new characteristics may  develop.
     5.   Berglund, Berglund, and Lindvall's  work  (14) confirmed
          by Myddleton, since he also found the characteristics
          of individual odors to be additive  and to develop the
          new characteristics for the mixture.
     6.   Mixtures that involve three or four component odors
          do not become more complex olfactory experiences.  We
          respond to mixtures in the same manner that we respond
          to single odors."

Summa.ry

     In summary, the characteristics of odor  are as follows:

     1.   "That property of a substance which excites the sense
          of smell. To be odorous, a substance is usually in a
          gaseous or vapor state, or possesses a vapor pressure."
          (5)
     2.   Four-dimensional, with the characteristics of
          intensity, detectability, acceptability, and
          quality (6).
     3.   A reduction in odor concentration of 15 to 30 percent
          is barely observable as long as the odor is still
          perceivable (7).
     4.   It was reported (8) that the odor intensity of a
          mixture was the arithmetic sura of the components of
          the mixture.
     5.   The importance of adaptation was emphasized (1)
          suggesting that if an odor is not perceivable, it will
          not become perceivable by smelling  for a longer period
          of time.

PHYSIOLOGICAL ASPECTS AFFECTING RESPONSE TO ODOR

     Variables, other than dosage may effect  response (17).
Examples of such variables include the subject's age, sex, profes-
sion, attitudes toward air pollution in general or the source in
question, and differences in earlier experience of related environ-
mental events.  It was also suggested that for practical and
statistical reasons, the triangle test be utilized even though a
simple yes-no scheme would allow more samples to be run over a set
time period.

     It has been reported (17) that the threshold of dimethyl
sulfide and dimethyl disulfide when mixed together and determined
empirically was about 70 percent of the calculated odor intensity.
The conclusion arrived at was that an inhibitive or antagonistic
effect existed.  It was postulated that this  may have been the
result of some process at the perceptual level since the observa-
tion was particularly noticeable at lower concentrations.


                               11

-------
     Several articles and authors deal with the pleasantness and
unpleasantness of the odor or odor hedonics (9,17,18).  The lat-
ter article (18) presented detail on testing techniques and types
of responses (pleasant vs. unpleasant) to be expected and how the
data can be analyzed.

FACTORS FOR CONSIDERATION IN PANEL SELECTION

General Considerations

     Duffee (7) discussed some of the pitfalls associated with
the utilization of human odor panels.  He stated that,

     Since people's reactions are so unreliable and variable,
     worst of all so expensive to secure, it is considered
     impracticable to depend solely on psychometric measure-
     ments.  Instead, instruments must be developed that respond
     in a fashion similar to the human olfactory sense, but in
     olfactometry we are far from achieving the technical
     sophistication that is attained in audiometry or photometry.

     However, he further stated,

     Although the human olfactory system does not compare in
     sensitivity to the chemical senses of many animals or
     insects, for most odorants, it does far surpass the detec-
     tion limits of our analytical devices.

     The literature does not delve into the matter of panel size
other  than generally stating, "the more the better," or if a
small  sample is used, then it should be selected to best simulate
the norm.  However, some investigators commented as follows (19):

     As previously indicated, all panelists are screened by
     the 'triangle1 technique.  Presently, neither the most
     sensative  nor the most insensitive of those screened are
     being used, but those exhibiting average olfactory
     perception are selected.

          A sufficient number of screened panelists should be
     available  so that the tests can always be conducted with
     a minimum  of 6 panelists.

     The same author  (19) also made a list of items affecting
panelist reliability judged by reproducibility of their results:

      •Eating, drinking or smoking within 1 hour of panel.
      •Insufficient sleep the previous night resulting in
       over-tired conditions.
      •Personal  behavior, i.e. a lack of interest or objectivity;
       easily influenced by others: inability to follow instruc-
       tions, indecisiveness, etc.

                               12

-------
     • Use of personal odorants, such as colognes, perfumes,
      shaving lotions, etc. on day of panel determination.
     • Impairment due to colds, hay fever, sinusitis, etc.
     • Lack of training or experience in sniffing.

     In the chapter of Young and Cheremisinoff written by T. M.
Hellman (1) it was indicated that the response of an observer was
dependent not only on the nature and strength of the stimulus but
also on the degree of personal adaptation, motivation, attitude,
expectancy, previous experience, and variation in background.
Differences between observers due to age, health, sex, and smok-
ing habits have been reported to influence olfactory sensitivity.
However, investigations of these variables have generally pro-
duced contradictory and inconclusive results.

     In determining the degree of detectability of an ambient
odor or subjective annoyance, it was suggested that a balanced
panel rather than a trained highly competent panel be used to
best represent the community (12,17). This differs from the
approach which would be taken in product quality control and
research for the food and cosmetics industry (17).

Panel Screening Considerations

     Cheremisinoff and Young (1) suggested that presecreening of
individuals for the selection of panelists should be conducted
with odorants similar to those to be perceived.  It has been
found that an individual's sensitivity to vanillin - nethyl-
salicylate may not correlate with his sensitivity to sulfides
or fatty acids. They also mention that background odors often
hinder a determination and it may be necessary and appropriate
to train the panelists to screen background odor from the
determination.

     They (1) also indicated that other than in a few simple
cases, the chemical composition of odor is so complex and varied
that the difficulties of a detailed analysis are nearly
insurmountable.

     Another discussion on the selection of panelists was pre-
sented by Duffee et al. (20).  In the selection of an odor panel,
care was stressed, as the members are chosen for their sensitiv-
ity, perceptual reproducibility, and experience.  It was said
that application of the triangle screening procedure using vanil-
lin and methyl salicylate solutions in benzyl benzoate for the
selection of odor panelists did not yield a consistent correla-
tion between an individual's triangle test score and his relative
sensitivity to industrial odors.  This paper dealt with odor
associated with the rendering, pulp and paper, and paint coating
industries.  They subsequently substituted methyl disulfide and
butyric acid in a benzyl benzoate solution for the triangle


                               13

-------
tests   These latter results were found  to  have significant
meaning with regard to the ranking of  individual panelists.

ODOR PANEL MANAGEMENT

     One investigator (9) indicated that screening of panelists
and management of the program, including methodology, training,
and motivation of the panelists, are important in developing
reproducible results.

     It was further reported  (20) that panelists performed best
in the forenoon hours.  Just before and  for at least two hours
after  lunch, their responses were quite  erratic and on the low
side by a factor of as much as three from their optimum.  It was
also noted  that three hours panel duty seemed to be the effective
span for obtaining panel results.

REPRODUCIBILITY OF INDIVIDUAL PANELIST RESPONSE

     Another investigator  (21) determined that the standard error
of a particularly qualified observer could  easily be ±40 percent
on the same sample, whereas with a panel of four, the standard
error  could be reduced to  ±20 percent  on the same sample.

     Still  another article  (9) indicated that a qualified panel-
ist  can have a deviation of 50 percent to the low side to 100
percent on  the high side of a geometric  mean concentration
threshold.   A further study (10) indicated  that even though two
observers may have nearly  the same odor  threshold for one com-
pound  (benzophenone), the  same two individuals can differ by a
factor of as much as  five  for another  compound.

PRINCIPAL HUMAN ODOR  MEASUREMENT TECHNIQUES

     Duffee (17) identified four main  categories of dilution-to-
threshold measurement techniques which included:

     "1.   In-situ dynamic  dilution  (ISDD)—direct diversion of
          part of the odorous emissions  or  ambient air to the
          odor panelists on a continuous basis.
      2.   In-situ static dilution  (ISSD)—direct diversion of
          discrete  samples of odorous  emissions or ambient air,
          to odor panelists housed in  a  static mixing chamber.
      3.   Off-site dynamic dilution  (OSDD)—collection of dis-
          crete examples of odorous  samples in suitable con-
          tainers,  transport  to  the  odor panel, presentation
          to the odor panelists  on a continuous basis.
      4.   Off-site  static  dilution  (OSSD)—Same as 3 for collec-
          tion and  transport, but presentation to panelists in
          discrete parcels (e.g., syringe dilution)."
                                14

-------
     It is obvious that the ISDD technique has the  least number
of variables affecting the value derived, and thus  should give
the most reliable results since it eliminates the need for stor-
age and/or transport of discrete samples of the odorous emissions.

Syringe

     Much of the earlier odor panel work and even some of the
more recent information have been developed utilizing the syringe
dilution technique (19,22).  This is a technique in which the
panelists each receive a gas sample from the master syringe at a
known dilution level introduced into their own syringe.  The
panelists then provide the last stage of dilution by drawing
odor-free air into their syringe.  A "yes-no" answer is obtained
to determine the threshold level.

Blenders

     One study (23) reported 30 panelists selected  to represent a
desired populace were used simultaneously.  Odor meters (dynamic
gas blenders) were used with the scale of concentrations being
based on the fifth root of ten  (10 * ), i.e., a geometric pro-
gression having increments of about 59 percent.

Odor Rooms

     Others (24) used different methods of measuring odor or odor
thresholds.  The panelists in this survey were in a static air
system utilizing a "low odor" or virtually odor-free background
air from an activated charcoal filter as the dilution medium.
The odor threshold was defined, for this study, as  the lowest
level at which all the panelists could detect an odor. This was
said to be consistent with the definition for minimum identifi-
able odor.

Olfactometer

     Many types of instrumentation for the determination of odor
thresholds are presented in the literature.  However, the dynamic
forced-choice, triangle olfactometer as developed and used by
Dravnieks (5-29) seems to have maximum applicability when con-
sidering reproducibility, accuracy, and portability.  One refer-
ence (26) indicated that one of its advantages, namely a large
dilution range, may be a weakness since there appears to be a
systematic error of 20 to 50 percent to the high side when an
auxiliary 27:1 splitter is introduced to the system.  The olfac-
tometer is preferably used with 7 to 10 panelists.  Each is
forced to select one of the three sample ports which is different
than the others, even if a guess is required.  The  results are
analyzed statistically to derive the odor threshold.
                                15

-------
Instrumental Analysis for Selected Odorous Compounds

     The chemical analysis for odorous constituents is not a pro-
cedure that is widely practiced.  The exception is the measure-
ment of reduced sulfur compounds.  Gas chromatography using a
flame photometric detector is the most widespread application of
this principle.  The comments in this are therefore confined to
this measurement procedure.

     In two articles, Stevens (30,31) elaborated on various
detection methods for sulfur gases, discussing the use of a
34 foot, 0.085 inch I.D. FEP Teflon  tubing packed with poly-
phenyl ether five-ringed polymer containing phosphoric acid on
40/60 mesh Teflon  column prior to a flame photometric detector.

     Other articles (10,30-34) elaborated on the operation
principles of the FPD and other methods of measuring sulfur gases
such as coulometric detection, conductimetric detection, methy-
lene blue with STRaction 10  (arabin ogalactin) wet chemical
technique and the West-Gaeke colorimetric wet chemical technique.
The use of a 394 m  filter with the FPD was noted, yielding a
specificity to sulfur over hydrocarbons of at least 10,000 to 1.

     Several authors (35-39) relate instrumental response to
varying concentrations of sulfur present in different compounds.
However, some are related in peak height without actually giving
an attenuation base, making it virtually impossible to determine
a lower detection limit.  One article (38) indicated that for the
gases of prime interest  (H2S, CH-SH, CH3SCH3), a minimum con-
centration of 11.5 to 24 ppb could be detected.  Another author
(37) indicated that by taking extreme measures, the detection
limits for H2S could be stretched to 7 ppb and to 15 ppb for
CEUSH.  One additional author (40), using a wall-coated open tu-
bular column made of borosilicate glass (30-38 m x 0.25 mm I.D.)
and coated with OV-101 or SP-2100 was able to detect sulfur com-
pounds at ppt levels when using a cyrogenic enrichment sampler.
Good reproducibility was reported at the 5 ppb level without the
enrichment sampler.

     Two articles (41,42) discussed the use of permeation tubes
for the calibration of flame photometric detectors.  One author
(42) elaborated on the use of an exponential dilution flask to
obtain even lower levels of the sulfur gas for calibration
through the entire range.

     Columns for separating sulfur gases other than the previous-
ly mentioned polyphenyl ether column (Stevens column) (30), have
also been mentioned in the literature.  Brnner et al. (43) had
success with a 0.4 cm I.D. by 80 cm Teflon  column packed with
40/60 mesh Graphitized Carbon Black.  They reported detection
limits of 5 ppb for H2S or SO- and 15 ppb for CH,SH.  However,
for heavier sulfides the retention time becomes a factor on the

                               16

-------
limits of detectability.  deSousa and Bhatia  (44) reported on  an
acetone washed Porapak QS column that separated H2S, COS, SO,,,
CH3SH, CH3S£H and CH^SCH.,.  The column was a 18.5 cm by 3 mm
O.D. Teflon  tube that had been acetone washed prior to filling
with the acetone.

     Another researcher  (45) indicated that when measuring the
reduced sulfur gases of  interest at 0.2 to 0.8 ppm, columns con-
structed of 0.3 mm O.D.  stainless steel tubing cleaned with
benzene, chloroform, acetone, and distilled water followed by  a
rinse with 3 percent Siliclad solution and heated at 110°C for
one hour were adequate.  Several columns of 3 mm O.D. stainless
steel tubing were then utilized following this procedure, includ-
ing a 3 m 6 AW DMCS 80/100;  1.8 m column packed with specially
treated silica gel  (maybe deactagel); a 1.8 m column packed with
5% silicone QF 1-6500 on 80/100 mesh Porapak QS; and a 7.3 m
column packed with  10% polyphenyl ether (6-ring) and 0.4% phos-
phoric acid on chromasorb G AW DMCS 80/100.

     Two additional papers outlined the uses of various columns
for sulfur gas analysis.  The first (46) indicated the separation
of H2S, S02, and CH-SH on Carbopak B-HT-100, the separation of
COS, H_S, CS2, and SO- on Chromosil 310, the separation of H_S,
COS, SO-, CH SH, CH CH SH and CH3SCH3 on Polyphenyl ether/Chromo-
sorb T, and the separation of H2S, COS, S02, CH.,SH, CH3SCH3 and
CH3SSCH3 on Supelpak-S.  In  the second article T47) the use of
a 50/60 mesh Chromasorb  T column coated with 12% polyphenyl ether
and 0.5% H3PO,, a Carbopak B-HT-100 column, a Chromasil 310
column, ana a 120/140 mesh Deactagel column packed in 1 foot of
0.085 inch I.D. FEP Teflon  were compared.

COMBINATIONS AND ALTERNATE ODOR ANALYSIS TECHNIQUES

     Weurman  (48) analyzed samples by more than one technique.
He found that odor  sensory dilution techniques utilizing humans
and gas chromatography should be considered complementary to each
other.  A second investigator (49) reported that the ASTM syringe
dilution technique  did not correlate at all with either an odor
room or dynamic dilution technique.  Another approach reported
was that of human odor identification of individual compounds  at
the exit of a gas chromatographic column.  This is sometimes
called an odorgram  (50,51).  It has proven to be beneficial for
determining odor intensities of specific compounds as well as
identifying them.

     Cheremisinoff  and Young (1) further stated.

     It is our opinion that  the odor-in-water test is more
     reliable than  the odor-in-air test.  This may stem from
     the almost unlimited amount of sample to be sniffed from
     each dilution.  The biggest difficulty with odor-in-water
     testing is obtaining odor-free dilution water.  We have

                               17

-------
     on many occasions obtained positive responses from
     distilled water alone.

KRAFT MILL ODOR IDENTIFICATION AND THRESHOLD

     Sarkanen, Hrutfiord, Johnson, and Gardner (52) indicated the
primary constituents of kraft mill odor were the sulfide gases,
hydrogen sulfide, methyl mercaptan, dimethyl sulfide, and methyl
mercaptan's oxidation product, dimethyl disulfide.  This article
stated that kraft mill odor will be very difficult to totally
eliminiate or control because of two factors, (a) the very low
odor threshold of the previously listed gases, and (b) the large
number of sources from which odors emanate around a kraft mill.
This article also contained a very complete bibliography of
articles pertaining to kraft mill odor.

     Several authors published information relating to the odor
threshold of the reduced sulfur gases of interest in this pro-
ject.  This information is presented in Table 1.

     TABLE 1.   ODOR THRESHOLD OF VARIOUS REDUCED SULFUR GASES
                       REPORTED IN THE RECENT LITERATURE

Investigator
H2S
Odor threshold (ppb)
^"^ TT C? TT ^^TT £* X^TT ^^TT (2 CJ ^^ l^
<^rf jn A o o v^-Ti & o v^fi *K \_^in A o o v^n. «
Method
Leonardos (24)  4.7*    2.1
                0.47**

Nishida (53)   39.5

               60
                1
                               26
                                4.4
                                0.992
Polgar  (54)

Wilby (23)
               20

                4.1
0.80
2.1
5.6
Minimum identi-
fiable odor

Dynamic olfacto-
meter,
Syringe
Triangle bag
test

Dynamic dilution

Medians from
dynamic mixing
             s.d.=2.9 s.d.=0.71 s.d.=2.2 s.d.=6.4
*
**
     from Na2S
     gas
     Several investigators have reported in the literature
published prior to 1968 odor threshold values for these same
compounds.  However, much of this work was performed prior to the
development of what is considered current sample measurement and
handling technology capability.  The construction materials used
in some of these early dilution systems suggest system losses may
have influenced the results.  Thus, these threshold values tend
                               18

-------
to be significantly higher than those currently reported. One of
the better summaries of this work was compiled by Droege  (55).

     Limited work  (56) indicated dilutions of 200,000 for flash
and blowheat condensate, 35,000 for field treated condensate, and
2,000 for laboratory treated condensate were required to reduce
the odor of these  streams to threshold levels.  The same report
indicated that at  another kraft mill, the accumulator condensate
had an odor threshold at 240 dilutions whereas the steam stripped
condensate had an  odor threshold at 5 dilutions.

     In a progress report on a CPAR project by Domtor (57) the
following conclusions regarding kraft mill process streams were
presented:

    "(1)  Although all the area sewers contained some odour,
          those from the kraft mill were, by far the most
          obnoxious
     (2)  (From Tables V and VI)* it may be seen that the kraft
          mill evaporator condensates and the kraft mill digester
          foul water condensate have the highest odour inten-
          sities and contributions and in total, constitute,
          practically speaking the whole odour contribution.
          These condensates fortunately represent only 7.6% of
          the total mill effluent volume.
     (3)  Minor contributions to the odour problem comes from the
          recovery furnace flue gas condensate, no. 9 seal tank
          effluent (Kraft Bleach Plant after the Hypochlorite
          Stage) and the Sulphite Mill Waste Liquor.
     (4)  Bench scale activated sludge and aerated lagoon experi-
          ments on various combinations of the indicated problem
          sources  showed up to 99% reduction in relative odour
          contribution from a combination of the evaporator
          condensates and the digester foul water condensate, and
          only a slightly lower efficiency when including the
          recovery furnace flue gas condensate.
     (5)  Bench scale chemical oxidation/stripping by aeration
          gave between 95% and 98% reduction in the odour con-
          tribution of the same effluent combinations listed
          under (4).
     (6)  T.C.A. (Turbulent Contact Absorber), pilot plant
          oxidation/stripping by aeration gave 96% reduction in
          relative odour contribution of the same effluent com-
          binations as above."

     The potential for the evolution of odor from treatment
systems and some methodology of removing the odor in a laboratory
situation were discussed.  Two articles (58,59) addressed the
potential for evolution or stripping of organic compounds from


"*    Reproduced as Tables 2 and 3 on the following pages.


                               19

-------
                    ODOUR AND TASTE THRESHOLDS* OF POINT SOURCES
Source
Evaporator condensate (seal
tank)
Digester foul water
condensate
Evaporator condensate -
(barometric well)
Sulphite mill waste liquor
Area
Kraft mill -
Kraft mill -
Kraft mill -
Sulfite mill

old recovery
sewer
main sewer
old recovery
sewer
- main sewer
A.P.H.A.
odour
threshold
no. (1)
2,000,000
633,000
41,600
25,000
A.P.H.A
taste
threshold
no. (2)
200,000
500,000
500
142,800
#9 Seal tank (hypochlorite
              stage)

Recovery furnace flue gas
 condensate

Vanillin plant barometric well

Kraft mill west sewer

#11 Seal tank (unbleached
         pulp thickener)

Causticizing effluent


Vanillin plant cooling water

Kraft mill middle sewer

#8 Seal tank (chlorination
              stage)

Sherbrooke thickener

#12 Seal tank (CIO- stage)


#1 Machine white water


White water chest

Cowans rejects

#10 Seal tank (C102 stage)

Bleached tertiary rejects
Kraft mill - bleach plant
             sewer

Kraft mill - new recovery
             sewer

Vanillin plant - main sewer

Kraft mill - main sewer

Kraft mill - bleach plant
             sewer

Kraft mill - old recovery
             sewer

Vanillin- plant - main sewer

Kraft mill - main sewer

Kraft mill - bleach plant
             sewer

Sulphite mill - main sewer

Kraft mill - bleach plant
             sewer

Kraft mill - bleach plant
             sewer

Sulphite mill - main sewer

Sulphite mill - main sewer

Kraft bleach plant sewer

Sulphite mill - main sewer
10,000


 2,500

 1,430

 1,430


   910


   830

   770

   670


   590

   290


   220


    40
 2,500


 1,250

    10
  250


  500

  200




  170

1,430


  170




6,670

  500

  170

   30
*    Thresholds are at 75% level.

1.   Five day composites using detailed testing method.
2.   One day composite using screening method.
                                           20

-------
          TABLE 3.  RELATIVE ODOUR AND TASTE CONTRIBUTIONS OF POINT SOURCES
Source
Evaporator condensate
(seal tank).
Digester foul water
condensate
Evaporator condensate
(barometric well)
Recovery furnace flue
gas condensate
Sulphite mill waste
liquor
19 Seal tank
(hypochlorite stage)
Kraft mill west sewer

Vanillin plant
(barometric well)
Causticizing effluent

til Seal tank
(unbl. pulp thickener)
Sherhrooke thickener

Kraft mill middle
sewer
112 Seal tank
(C102 stage)
#1 Machine white water

Vanillin plant cooling
water
#8 Seal tank
(chlorination stage)
Cowans rejects

110 Seal tank
(ClOj stage)
Bleached tertiary
Total mill effluent
Area
Kraft mill
old recovery sewer
Kraft mill
main sewer
Kraft mill
old recovery sewer
Kraft mill
new recovery sewer
Sulphite mill
main sewer
Kraft mill
bleach plant sewer
Kraft mill
main sewer
Vanillin plant
main sewer
Kraft mill
old recovery sewer
Kraft mill
bleach plant
Sulphite mill
main sewer
Kraft mill - main
sewer
Kraft mill - bleach
plant sewer
Kraft mill - bleach
plant sewer
Vanillin plant - main
sewer
Kraft mill - bleach
plant sewer
Sulphite mill - main
sewer
Kraft mill - bleach
plant
Sulphite mill - main

Relative *
odour
contribu-
tion

28,000

6,330

790

180

175

120

17

13

12

11

9

3

3

3

1

1

—

—
-
(1,890)
Relative **
objection-
ableness of
odour

10

9

6

4

7

5

4

3

2

6

6
o

4

4

3

—

4

~

7
3

Relative *
taste
contribu-
tion

2,800

5,500

10

90

1,000

30

-

1

7

3

44

-

2

—

0.2

2

3

2
0.1
(5,880)
 *   Plow of sewer as percent of total mill flow divided by threshold concentration.

**   Rated on a 1-10 scale using 6 and 7 paper machine sewer as "1" and seal tank
     new evaporators as "10".
                                         21

-------
biological treatment facilities.  In the first article  (58) the
author discussed the calculated desorption of volatile  gases and
liquids from aerated stabilization basins using a mathematical
model and certain fixed parameters.  These parameters were con-
trolled to determine their effect on the desorption rate.  It was
concluded that the most important parameters that affect strip-
ping in aerated stabilization basins were temperature, wind
velocity, liquid droplet diameter, and aerator interfacial area.
The second publication (59) discussed a method of physically
removing the volatiles from solution by simulating an aerated
stabilization basin, and measuring the volatile organics removed.

It is recognized there are many additional references on
odor and its measurement.  The review selectively presents
information in the literature of the last decade, several arti-
cles of which contain comprehensive bibliographies and in general
represent the advances in odor measurement technology during that
period.
                               22

-------
                          SECTION 5

                    EXPERIMENTAL METHODS


     During the course of this investigation three  separate
procedures were used in preparing samples  for  their presentation
to human odor panelists.  Two were discontinued since they were
not found to be applicable in reaching  the objectives of this
study.  The first involved total volatilization of  a known volume
of liquid sample into a known gas volume.  This procedure was
found not to be applicable since it was incapable of distinguish-
ing between the total odor level of the sample and  that portion
which was readily volatilized with agitation at 40°C.  Use of
this procedure also produced odor thresholds which  differed only
slightly, if at all, between sources.

     To overcome problems associated with  the  total volatiliza-
tion procedure a stripping procedure was investigated.  It was
found to suffer some of the problems of the total volatilization
method, namely inability to distinguish odor level  between widely
varying sources which were evident, even to the casual observer.

     A head space analysis was ultimately  used which incorporated
the forced choice triangle test in conjunction with the proce-
dures outlined in Standard Methods (60).

     Each of these sample preparation procedures is described in
the order they were used during the course of  the study.  The
bulk of the data were developed using the  head space analysis
procedure, however.

SAMPLE COLLECTION AND STORAGE

     The procedure for sample collection and storage was consis-
tent during the course of this study.   Samples to be analyzed
were collected in 300 ml glass stoppered bottles and delivered
to the NCASI West Coast Center where they  were refrigerated over-
night prior to odor analysis.

     Samples collected from West Coast  mills generally arrived at
the West Coast Center on the same day they were collected, where-
as samples from the Southeastern U.S. were in  transit four to six
days after collection.  Storage of this duration was found to
have an insignificant impact on the "identifiable"  odor level.


                               23

-------
HUMAN ODOR PANEL SELECTION

     Panelists were initially selected on their ability to detect
(a) the standard kraft pulping process reduced sulfur gases such
as hydrogen sulfide, methyl mercaptan, dimethyl sulfide and di-
methyl disulfide and (b) n-butanol at or close to the literature
reported values using a dynamic olfactometer.  Panelists were
screened on a regular basis and accepted or rejected on their
continuing ability to detect the standard sulfur gases at or
close to the detectable concentrations found in this study of
1.9 ppb hydrogen sulfide, 0.43 ppb methyl mercaptan, 3.3 ppb di-
methyl sulfide and 0.80 ppb dimethyl disulfide.  The purpose of
the latter was to eliminate those panelists that had periodic
extreme olfactory responses, either high or low.  This was nec-
essary to maintain an "average" sampling of olfactory responses.
To qualify, the panelist had to be no more than a factor of three
higher or lower than the threshold level.  This variation was
selected since the olfactometer had a factor of three difference
between dilution stages.  Thus, missing the correct concentration
by one dilution stage would result in an error by an approximate
factor of three.

SAMPLE PREPARATION FOR THE TOTAL VOLATILIZATION TECHNIQUE

     A total odor or volatilization method involved the volatili-
zation of about 0.1 ml of sample into 17 liters of clean, dry
air.  This was accomplished by placing a midget impinger with two
side ports in a 100°C bath, and purging with charcoal filtered,
desiccated air as shown in Figure 1.  The^air volume was measured
by a wet test meter.  A Teflon  or Tedlar  sample bag was con-
nected to the outlet of the impinger and 0.1  1 of sample was
injected into the impinger.

SAMPLE PREPARATION FOR THE STRIPPING TECHNIQUE

     The stripping technique for sample preparation was simpler
in design.  The principle was stripping at room temperature to
remove the volatile compounds.  This utilized a glass stripper
column, 2.5 cm I.C., filled with glass beads and helixes to a
height of about 30 to 35 cm, that received liquid sample at a
flow of about three to five ml/min and utilized nitrogen as the
stripping gas at a rate of about two 1/min as shown in Figure 2.
The flow of liquid sample to the column was controlled by the use
of an orifice and constant head device.  The gas handling se-
quence included a nitrogen cylinder, regulator, micrometering
valve, wet test meter, desiccant, activated charcoal filter,
stripping column, and a 17 liter Teflon  or Tedlar  bag.  During
the stripping process, temperature was measured by using a ther-
mometer that had been placed in the column packing.
                               24

-------
MICROMETERING
   VALVE
                                      ACTIVATED
                                      CHARCOAL
                                       FILTER
      PUMP
                                                             100* C
                                                             WATER BATH
                     17 LITER TEFLON®

                     OR TEDLAR®BA6
      Figure 1.    Total  odor  volatilization technique.
                         ACTIVATED
                         CHARCOAL
                           FILTER
                                                            17 LITER INERT BAG
                                             STRIPPED
                                             SAMPLE
        Figure  2.    Strippable odor technique.
                                  25

-------
SAMPLE PREPARATION FOR THE HEAD SPACE ANALYSIS TECHNIQUE

Odor Free Water Preparation

     In the early stages of the project when the panelists were
being trained, dilution water was prepared by filtering distilled
water through a charcoal column at a rate of about 100 ml/min  as
described in Standard Methods (60).  A five centimeter in diame-
ter glass column filled to a depth of about one meter with acti-
vated charcoal was used.  This technique was never perfected to
the point that the water no longer exhibited a "burned" and/or
"musty" odor and consequently was not used.

     Odor free water used during this evaluation was prepared  by
purging nitrogen regulated from 200 to 500 ml/min through dis-
tilled water heated to approximately 50°C for a minimum of six
hours.  This allowed for stripping of the residual materials that
might be detected at the elevated temperatures at which the tests
were conducted.

Dilution

     The samples were sequentially diluted in stages of not more
than a factor of 2,000 to 1 and no less than 2.5 to 1 by use of
volumetric pipettes and flasks.  A stock solution of a given
dilution factor was prepared.  To select the dilution submitted
to the panelists, the odor detectability was checked by placing
200 ml of the solution in a 500 ml"large-mouth erlenmeyer flask.
The flask was then placed in a 40°C water bath and allowed to
equilibrate.  After a sufficient period of time (five to ten
minutes) the flasks were smelled to determine if there was an
odor present.  If no odor was present or if the odor present was
judged too strong to identify an odor threshold, the appropriate
corrections were made and the new solution was evaluated in a
similar manner.

     The procedure employed during this study had six dilutions
or concentration levels with the stock solution representing the
strongest dilution and each subsequent dilution stage was weaker
than the previous stage by a factor of three.

PREPARATION OF THE CALIBRATION GASES

     TeflonR or TedlarR nine liter bags that had been filled with
known quantities and dilutions of one or more of the gases,
hydrogen sulfide, methyl mercaptan, dimethyl sulfide, and di-
methyl disulfide were prepared from permeation tubes.  Each day
the chromatograph calibration curves were prepared at four or
more different concentrations for each gas.  Bags filled in a
similar manner with one of the sulfur gases were used to screen
the panelists on a regular basis.
                               26

-------
ODOR THRESHOLD DETERMINATION BY DYNAMIC OLFACTOMETER

Principle

     The dynamic olfactometer was designed for use with the
forced-choice triangle test and utilized six dilution stages.
The sample was diluted with charcoal filtered, odor-free room
air.  The unit, as shown in Figure 3, was designed by Dravnieks
of the Illinois Institute of Technology (27).  The samples were
presented in the order of the most dilute through least dilute.
The panelists were forced to identify which port in a cluster of
three contained odors.
     Figure-3. Forced-choice triangle dynamic olfactometer.
     Data accumulated during this portion of the project included
only the odor threshold or that concentration level at which a
difference could be determined between the blank and sample odor.

Equipment Description

     The olfactometer was used to determine the odor threshold of
various process streams and to screen the panelists on a regular
basis.  The olfactometer received a gas sample from an inert bag
at a rate of about 100 ml/min.  Dilution air flow rate was about
nine 1/min.
                               27

-------
     Each dilution stage of the olfactometer had three ports and
there were a total of six dilution stations.  There were 18 indi-
vidual ports for a panelist to smell in each test run.  One of
the ports at each dilution station received a known rate of sam-
ple from the sample bag plus a dilution air flow of approximately
500 ml/min.  The other two ports received only dilution air at a
rate of about 500 ml/min.  The flow of sample from the bag to
individual ports varied from about 3.2 to 32 ml/min.

     The olfactometer flows were controlled by the use of sec-
tions of capillary tubing, with the lowest flows having the lon-
gest section of tubing.  For a more complete descripition of the
olfactometer, it is suggested that Source Emission Odor Measure-
ment Via Dynamic Forced-Choice Triangle Olfactometer by Dravnieks
of the Illinois Institute of Technology Research Institute and
Prokop of the National Renderers Association (30) be studied.

Procedures

     The flows to each port receiving a sample were measured and
recorded daily.  These values were recorded in the table shown in
Figure 4.  Flow at attenuators A and B, theoretically 27:1 were
also measured daily.

     Experimental data gathered while utilizing the forced-choice
triangle test technique were collected on the data sheets shown
in Figure 5.  Included on the data sheets were sample identifica-
tion, date of analysis, panelist response, and the information
needed to determine the ED__, that concentration at which 50 per-
cent of the panelists detected a differentiation in odor.

     Each panelist was identified by name and order of partici-
pation in the specific test.  The responses were noted in the
appropriate box.  Only the left-hand side of the sheet was used
for data collection.  The right-hand side of the sheet was the
table of "plotting values" utilized in the calculations describ-
ed later.

Calculations

     Dilution factors at each stage of the olfactometer were
calculated.  The flows of the sample gas  (Q ) and dilution air
(Qd) were both measured and recorded as shown in Figure 4.
The dilution ratio  (R,) was calculated by the formula:
the dilution ratio R^ was then multiplied by another  constant if
it was applicable, such as when a portion of the  sample  went
through a conditioning step.
                                28

-------
                                              IITRI January 1974
              FLOW CALIBRATIONS FOR DYNAMIC TRIANGLE
              OLFACTOMETER.   DATE 	

                      (ALL FLOWS IN ML/MIN)
PRE-ATTENUATOR (IF ANY)
THROUGH C
5.0209
BYPASS
94,7297
PRE-ATTENUATION
Factor A
19.8671
IF NO PRE-ATTENUATION, A=l
DILUTION LEVELS:
No.
6
5
4
Odor
29.1971
11.0092
3. 4582
Dilution Air
366.9725
489.7959
444.4444
ATTENUATOR
THROUGH C
8.3916
BYPASS
80.6939
Dilution
Factor D
13.5688
45.4897
129.5190
Attenuation
Factor B
10.6160
D x A
269.5724
90S. 7471
2573.1637
A x B
210.9095
DILUTION LEVELS:
No.
3
2
1
Odor
31.0078
9.8361
3.6474
Dilution Air
500 . 0000
485.8300
472.4409
Dilution
Factor D
17.1250
50.3925
130.5281
D x A x B
3611.8203
10628.2674
27529.6257
Port
No.
6
5
4
3
2
1
Log Total
Dilution
Factor
2 4307
2.9560
3.4105
3.5577
4.0265
4,4398
Figure 4.   Olfactometer  flow  data  sheet.
                             29

-------
                      Evaluation Form for Forced-Triangle Head Space Analysis
U)
O
Sample:
Evaluati
elist
1
2
3
4
5
6
7
8
9
10
	 .Stock .OUutipn. =_
on r)a 1" P *

RES!

Name
A
B


C.
D
E
F
G
11


Frequency Tal 1
Average Rank
= Plotting Value
y
/
A
1
, 	 f\ — .-_
A
B
A
C
B
C
B



)
Lxlt)5 : 1



ILT: Log CDs0=5.5736 E0so=31746xl05 r2=0.9598
B
2
A
R
C
C
B
A
A



I
C

B
	 B
C
B
B
A
A
(c)

I
y
22-0.
1 A
4
B
C
C
B
(£}
v_x
U_A
B
A

1
Y
77-0.

5
C
C
A
0,
B
A
C
B


1
y
43 fO.
L_£__
6
©
J(eX
B
	 ii 	
C
(^
fc)
C


4
5.5,
2f +1.
"7"



0




—



1
f
22
V
For Rank
Count
1
2
3
4
5
6
7
8
9
10
Plotting Values
Average Number of Panelists
Rank 6 7 8 9 10
1.0 -1-07 -1.15 CTT.2C)-1.28 -1.33
1.5 -0.79 -0.89 -0.97 -1.04 -1.10
2.0 -0.57 -0.67 dO-70-0.84 -0.91
2.5 -0.37 -0-.49 -0.59 -0.67 -0.75
3.0 -0.18 -0.32 Co. 4Q -0.52 -0.60
3.5 0 -0.16 -0.28 -0.39 -0.47
4.0 +0.18 0 -0.14 -0.25 -0.35
4.5 +0.37 +0.16 0 -0.13 -0.23
5.0 +0.57 +0.32 +0.14 0 -0.11
5.5 +0.79 +0.49 C+oH?D+0.13 0
6-0 +1.07 +0.67 +0.43 +0.25 +0.11
6.5 +0.89 +0.59 +0.39 +0.23
7.0 +1.15 +0.77 +0.52 +0.35
7.5 +0.97 +0.67 +0.47
8.0 C+1.2£>+0.84 +0.60
8.5 +1.04 +0.75
9.0 +1.28 +0.91
9.5 +1.10
10.0 +1.33
            X


            Y = Log (tolerance level)]


            Log (dilution factor)


            Dilution no.

<1
7.39
1
6.91
2
6.43
3
5.95
4
5.48
5
5.00
6

? • 7 ; t
             Figure  5.   Data collection and calculation sheet for forced-choice
                                              triangle test.

-------
     The 50 percent effective dose  (ED  Q)  levels  were  calculated
in the following manner.  The log of  tne dilution factors  for
each dilution stage were entered.   It was  assumed that when the
individual was able to differentiate  an odor  at a particular
dilution level and respond accordingly, that  their true odor
threshold was at some dilution between  that level and  the  preced-
ing more dilute level.  The panelist  was given "credit" for this
inequity by computing the log average for  these two concentration
stages, and indicating that the odor  was detected at this
calculated level.  These values are shown  as  "Y = log  (tolerance
level)" in Figure 5.  The log (tolerance level) values for either
extreme were obtained by assuming they  were as far from the
nearest log (dilution factor) value as  the previous log tolerance
level .

     The "X = plotting value" numbers were obtained by tabulating
the results of the testing.  First, the "correct" answers  were
identified and circled as shown in  Figure  5.  This "correct"
response correlated to the most dilute  sample at  which the panel-
ists were able to detect odor and continue with correct responses
after that point.  The number of circles per  dilution  stage were
then tabulated with the totals entered  as  shown in the "Frequency
Tally" row.  The "Average Rank" was then computed utilizing the
column entitled "For Rank Count."   Starting with  the highest
dilution on the left side of the page,  the numbers in  the  "For
Rank Count" column were marked with the quantity  between marks
the same as the numbers in the "Frequency  Tally"  row.   Thus,  as
in example 1,1,1,4 and 1, numbers were  checked off sequentially
as they appeared in the "Frequency  Tally"  row.  Once this  had
been done, the values between the marks were  averaged  and  noted
in the column entitled "Average Rank" such that in the preceding
example the values were 1,2,3,5.5 and 8.

     The "Average Rank" values were converted to  the probability
related "Plotting Values" using the table  shown on the right  side
of the data sheet.  These values are  equal to probits-5 (61).
The values obtained were from the column for  eight panelists  as
circled.  The Y values, log (tolerance  level), were then plotted
against the X values, plotting value, the  ED5_ level was the
point where the log tolerance level crossed tne plotting value
equal to zero as shown in Figure 6.   This  process was  simplified
using the method of least squares and the  formulas to  determine
the equation of the "best fit" line through the points as
follows:

     the general formula for the straight  line,

     Y = mx+b where m is determined by  the formula

                       x  -  x
               m =
                    n x2 -  ( x)2
                               31

-------
     and b is determined by the formula

                  2
          b  =
                    y -  x xy
                 n  x2  - (  x)2
 UJ
 UJ
 o
 CC
 UJ
     where n  is  the  number of paired data points  and

     x and y  are the individual data points
   7.O
   65
   6.0
 o 5.5
 o
   5.0
                                                 X = 0
                                                 Y = 5.57
                                                       3.7xlO
     -1.6    -1.2     -0.8     -0.4     0.0     0.4

                               PLOTTING VALUE
                                 0.8
1.2
     Figure  6.   Plotting value vs. log (tolerance  level).
     The ED5Q value  was determined by substituting  0  for x in" the
general line equation and solving for y which was the log of the
ED - level.  The  correlation coefficient r  was also  determined
for each trial and was calculated by determining the  standard
deviation of the  x values (S )  and y values  (S  ) by the following
formulas:                                      v

                 -2    ( x)2
                x   -
          Sx '
                       n
n - 1
                                32

-------
                       y>2

and then

           2 _ n xy -  x y
             " n(n-l) SxSy


     The correlation coefficient was determined  for  each
individual run.

     An example of the calculations, using  the information as
shown in the Figure 5 would be as  follows:


     m = 5 (-8.0878) - (-0.9200)  (28.5784)   = _0>7723
               5 (3.8330) -  (-0.9200)2


     b = (3.8330) (28.5784) -  (-0. 9200 ) (-8. 0878 )
               5 (3.8330) -  (-0.9200)2           =  5-5736


and substituting back into the equation  for a line  (y = mx+b) and
setting x=0 yields y= (-0. 7723 ) (0) +(5. 5736 )=5. 5736 which is the
antilog pf the ED _ which can  then be calculated to  be equal to
3.746x10 .  Next, the correlation  coefficient is calculated by
first determining the standard deviation of the  x and y values
as follows:
           {3.8330) -  -

     Sx =V - 5-1   -  =  °-957°


          /165.6214) -  <2V784)2
     Sy =J - 5-1    -  '  °'7544


and then  it follows that

      2 _  5  (-8.0878)  -  (-0.9200H28.5784) 2
        ~      (5) (5-1) (0.9570)(0. 7544)


     Another calculation  used extensively in this report was the
log average which was computed by  adding the log of all the
dilution  factors of concern, determining the average of these
log values, and then obtaining the antilog of that average,
hence the log average.
                                33

-------
ODOR THRESHOLD DETERMINATION BY HEAD SPACE ANALYSIS

Principle

In the head space analysis, an equilibrium is established  at  a
preset temperature in a confined and known volume between  com-
pounds in a known liquid volume and the air space above  it.   The
presence or absence of an identifiable odor in the head  space was
determined by trained panelists in this study.

Equipment Description

     The head space analysis was conducted using 500 ml  erlen-
meyer flasks with standard taper pennyhead stoppers containing
200 ml of solution and maintained at a temperature of 40°C.   Six
dilution stages were used with the triangle test employed  at  each
stage.

Procedure

     The samples were presented to the panelists in the  order of
most dilute through least dilute.  The panelists were forced  to
identify which flask in a series of three contained an odor and
at which dilution stage the odor could be identified with  a
sample of the same solution that he had been exposed to  earlier.
This is in contrast to the olfactometer exposure method  in which
difference in perceived odor was a determining criteria.   Before
smelling each sample, the panelists were instructed to shake  the
stoppered erlenmeyer flask.  It was hypothesized that this action
established an equilibrium between the liquid and gas phases
present in the closed system.

     Experimental data gathered while utilizing the forced-choice
triangle test technique was collected on the data sheets shown in
Figure 5 in the same manner described earlier while using  the
olfactometer.

Calculations

     The calculations for the head space analysis technique were
the same as for the dynamic olfactometer with one exception.  The
calculation steps for determining the dilution factors at  each
stage of the olfactometer were not required since this step was
simply recording the sample concentration in the flasks  at the
six dilution stages.  The assumptions made earlier concerning the
inequities in the computation of the log average odor threshold
values also apply to this method when determining "identifiable"
odor thresholds.
                               34

-------
Alternate Head Space Procedure Used

     Another method of  head  space analysis was used early in the
process and for a  limited  time only.   This method utilized seven
flasks containing  different  dilutions of the process stream of
interest arranged  in ascending concentrations at 40°C.  Inter-
spersed with these flasks  were three  flasks that contained only
the odor free dilution  water.   The panelists were asked to indi-
cate the concentration  level where they first perceived an odor
that was different from a  blank.

CHEMICAL ANALYSIS  BY GAS CHROMATOGRAPHY

Principle

     The gas chromatographic analytical procedure separates the
individual constituents of a gas  stream by the use of a colunn
that selectively adsorbs and desorbs  the compounds of interest at
varying rates.  The system utilized in this study was equipped
with a flame photometric detector (FPD) which is sulfur specific
at a wave length of 394 m  .

Equipment Description

     The chromatographic system is depicted in Figure 7.   As can
be seen from the figure, a hydrogen sulfide permeation tube was
placed in the carrier gas  stream, following the eight-port sam-
pling valve but prior to the chromatograph oven.  The tube deliv-
ered hydrogen sulfide to the detector at about 0.2 ppm.  The
column which contained  a 3.35  x 6 mm  (11 ft. x % in.) Teflon
column packed with 12%  polyphenyl ether + 0.5% H3P04 on a 40/60
mesh Chromasorb T  column.  This technique allowed a constant flow
of sulfide to continuously fill the active sites on the column.
In addition, it raised  the background level of sulfide to the
detector, since the FPD response  curve indicates linear response
from 5 ppb to 0.9  ppm at the detector.  An electrical heating
tape was used on the exhaust of the flame photometric detector to
prevent condensation and subsequent intermittent blockage of the
gas exit that resulted  in  "spikes" on the recorder readout.
                          CHROMATOGRAPH OVEN
                               R
        8-PORT
      SAMPLE VALVF
                T
                 H2S
              PERMEATION
                TUBE
POLYPHENOL ETHER
COLUMN ON CHROM T
                pEXAUST


                 	1 FLAME PHOTOMETRIC
                 	1 DETECTOR
          Figure  7.  Chromatographic system schematic,
                                35

-------
Procedure

     The sample  was purged through  the  eight-port sample valve
and through one  of the two 10 ml sample loops before injection  as
shown in Figure  7.   The injection proper was simply the reversal
of these two  sample loops, trapping  10  ml of sample and injecting
it to the  system.

     The chromatograph column oven was  temperature programmed
starting at 60°C for 4 minutes and increasing at the rate of
16°C per minute  to 150°C which was held as long as required to
elute all  the peaks, generally not more than 20 minutes.  The
carrier gas or nitrogen flow rate was maintained at 90 ml/min
and the oxygen was set at 31 ml/min.  The detector and injector
ovens were maintained at 105°C and the  FPD block temperature was
controlled at 95 to 100°C.
                            POSITION   I
           SAMPLE IN
           N2 OUT
           (TO G.C.)
10ml SAMPLE LOOP


            N2 IN
             SAMPLE QU-
              ITO PUMP)
                                     10ml SAMPLE LOOP
                           POSITION  2


                                    10ml SAMPLE LOOP
           SAMPLE IN
           N2 OUT
           (TO G.C.)
                                                No IN
             SAMPLE OUT
              (TO PUMP)
                                    10ml SAMPLE LOOP
          Figure  8.   Operation of 8-port  sample valve,
                                36

-------
Calculations

     Concentration  of  a  specific gas in one of the nine  liter
bags was calculated by the following formula:
{perm rate!/  min  perm  j/22414 ml^  [room temp (°K)]/  1 mole  \      fi
[ (g/min) J\tube in  line/I mole   /  L  273°K      J\mol wt gas/x  (10
           [total volume in filled bag (ml)]

     The  daily calibration curves for the chromatographic peaks
were established by comparison of peak heights for H?S and CH-.SH
and by comparison  of peak areas for CH,SCH., and CH..SSCH.,.
                                  37

-------
                          SECTION 6

                    DISCUSSION OF RESULTS
ODOR THRESHOLDS OF INDIVIDUAL REDUCED SULFUR GASES
BY OLFACTOMETER
                                                                j^
     By utilizing the olfactometer, permeation tubes, and Teflon
or Tedlar  bags, it was possible to generate standard concentra-
tions of the reduced sulfur gases, hydrogen sulfide, methyl mer-
captan, dimethyl sulfide, and dimethyl disulfide and to determin-
ing their respective odor thresholds.  Bags containing one or
more of these gases served as a panelist source gas on a regular
basis.  Odor thresholds for the reduced sulfur gases were devel-
oped and are shown in Table 4.  The log average of the values
obtained indicated that the gas's odor thresholds were 0.43 ppb
for methyl mercaptan, 0.80 ppb for dimethyl disulfide, 1.9 ppb
for hydrogen sulfide, and 3.3. ppb for dimethyl sulfide.  Litera-
ture values for H_S vary from 0.47 to 60 ppb, from 0.80 to 2.1 ppb
for CH.,SH, from 0.992 to 2.6 ppb for CH SCH., and one value report-
ed for CH-SSCH- was 5.6 ppb.  Thus, a good correlation for all
gases witn the exception of dimethyl disulfide, for which there
is limited literature information available was obtained.  The
column entitled "max." denotes the highest odor threshold level
obtained for that specific gas while "min." denotes the lowest
value obtained for that specific gas.

     TABLE 4.  ODOR THRESHOLDS FOR REDUCED SULFUR GASES
                    DETERMINED WITH OLFACTOMETER

Gas
Hydrogen sulfide
Methyl mercaptan
Dimethyl sulfide
Dimethyl disulfide
No. of
trials
16
13
13
16
(max)
ppb
4.68
0.79
9.82
3.37
Odor threshold
( 1 og ave . ppb )
1.89
0.43
3.31
0.80
(min)
ppb
0.79
0.12
0.55
0.18

ODOR THRESHOLDS OF COMBINED REDUCED SULFUR GASES BY OLFACTOMETER

     Cheremisinoff and Young (1) indicated that odors such as
sulfides are additive.  If this assumption is true, then the
following equation should be true:


                                38

-------
     OT =  [H2S1     [CH3SH]     [CH3SCH3]

             ~~
                                                        =

                                          +
                              OT           DT1
                       ^SH    U1CH3SCH3    U1CH3SSCH3

In the equation  [H2S], [CH3SH],  [CH3SCH3], and  [CH3SSCH3J, repre-

sent the concentration of that specific  gas  at  the  ED    level

calculated and OT     OT    R, OT         and OT          represent

the previously listed odor thresholds  for that  specific  gas.

     Data generated early in this  study  as shown  in Table 5,
indicated that the odor threshold  of sulfide gases  are not addi-
tive when measured using this technique.  When  considering that
the odor threshold range for a specific  gas  could vary as much as
an order of magnitude from day to  day, it was not expected that
the odor threshold of a complex mixture  could be  predicted by
calculation with any degree of accuracy.

     TABLE 5.   ADDITIVE NATURE OF  PURE  SULFIDE GASES
	ON OLFACTORY  RESPONSE	

     Gas combination                     Log  average OT (ppb)
TT O
H2S
H2S
Ho
9
CH,
CH3
CH3
+ CH
+ CH
+ CH
SH +
SH +
SCH3
3SH
3SCH3
3SSCH3
CH3SCH3
CH3SSCH3
+ CH3SSCH3
0.
0.
0.
0.
0.
0.
806
398
491
192
291
232

ODOR THRESHOLD OF SELECTED STREAMS BY THE TOTAL ODOR
VOLATILIZATION TECHNIQUE

     A portion of this study was spent attempting to develop a
technique by which the dynamic olfactometer described by Dravnieks
and Prokop (28) could be utilized.  In order to obtain a sample
that could be delivered to the olfactometer, a total volatiliza-
tion technique was developed.  In this procedure, a small amount
of liquid process stream effluent was volatilized in an appro-
priate amount of dry, odor-free air to maintain the relative
humidity in thepmixture below the dewpoint and was captured in
either a Teflon  or Tedlar  bag.  The technique used is presented
in the section on Experimental Methods.  The detectable odor
threshold of the sample in the bag was then determined by the
                                39

-------
use of dynamic olfactometer and human odor panels.  This method
was labeled the "total odor" technique since the whole sample and
therefore all compounds had been volatilized.

     The data as shown in Table 6, utilized a unit, the "log
average —odor threshold— /(1 air)n  This value was calculated
     3        ml sample
by dividing the odor threshold developed using the olfactometer
by the volume of liquid sample volatilized and the volume of
odor-free dilution air used.

     TABLE 6.  TOTAL ODOR TECHNIQUE: ODOR UNIT RANGE	

No. Of
Source mills
Steam stripper product
Digester condensate
Turpentine decanter
bottoms
Air stripped condensate
Stripper feed
MEE body condensate
MEE surface condenser
ASB influent
ASB effluent
Steam stripped cond.
2
1

3
2
6
2
2
4
2
4

Range
No- ?f Log ave. °
samples 3 m
3
6

7
8
26
15
5
16
9
10
14
1

1
0
0
0
0
0
0
0
(xlO~5)
dor
units ,
1 sample '

.7

.2
.91
.64
.64
.64
.37
.25
.15
- 350
- 20

- 27
4.
- 47
- 15
- 12
- 11
6.
- 14
1 air)




3




6


     From the data in Table 6, it can be seen that all of the
process streams investigated gave the same range of values.  This
included such widely variant steams as the treated liquid from a
.steam stripper, turpentine decanter underflows and aerated stabi-
lization basin effluents, etc.  Observation based on odor inten-
sity of the head spaces above stored samples by the investigators
indicated this to not be the case.  It was hypothesized that the
volatilization of the entire sample allowed compounds that would
not normally volatilize with the degree of turbulence encountered
in sewers and waste treatment systems to do so.  Following this
line of reasoning, compounds not normally volatilized were assumed
to be contributing significantly to the detectable odor threshold
which was measured using this procedure.  Alternate procedures,
judged to more closely simulate the turbulence encountered within
a mill environment, for generating the odor panels source gas
were then investigated.

ODOR THRESHOLD OF SELECTED STREAMS BY THE OLFACTOMETER AND
STRIPPING TECHNIQUE

     Since there did not appear to be a significant difference in
the odor threshold levels developed by the "total odor" technique,


                               40

-------
a stripping technique was used  in  an  attempt  to  more  closely
simulate the odor liberation process  from  the treatment  facility
and sewer lines.  This method included  the use of  an  olfactometer
for determinable odor threshold as previously discussed.   Samples
were prepared by nitrogen stripping and capturing  the off-gases.
The technique used is presented under the  appropriate section of
the Experimental Methods section of this report.

     On a trial run, this method yielded odor threshold values
that were similar to the "total odor" technique  results.   Varia-
tion was only about 100 fold between  the cleaner samples observed
to have a low odor level, such  as  the aerated stabilization basin
effluent, and the samples with  high odor intensity such as
digester condensates.  These results  are shown in  Table 7.  It
was interesting to note that in general, these values were 2 to
3 orders of magnitude smaller than those derived by the "total
odor" technique indicating  that, turbulence or a gas-liquid
interface phenomena plays a significant role  in  the amount of
evolved and measured detectable odor  of an effluent or process
stream.  Alternate means for simulating the turbulence encoun-
tered by effluent streams at mill  sites were  considered.  Being
largely emperical, they were not pursued and  more  conventional
methods for odor generating from a liquid  and its  measurement
were pursued.

	TABLE 7.  STRIPPING TECHNIQUE; ODOR UNIT RANGE	

                                               Range  (xlO  )
Source Log ave .
Digester condensate
Lab aerated digester condensate
Clarifier outlet
ASB effluent
Odor units mtrr -,ir)
	 	 1 J.J_^tJi <1XX|
ml sample
0.043 - 0.23
0.0032 - 0.0074
0.0015 - 0.0057
0.0015 - 0.0031

RELATIONSHIP OF MEASURED REDUCED  SULFUR CONCENTRATIONS TO THE
ODOR THRESHOLDS

     During the portion of this investigation in which the odor
threshold values were determined  with the panelists using the
olfactometer, the concentration of reduced sulfur gases in the
bag introduced to the olfactometer were determined with the gas
chromatograph equipped with a  flame photometric detector.

     The measured concentration of reduced sulfur compounds were
compared with the odor threshold  determined  for each sample
analyzed.  For each source~with more than three samples, the
correlation coefficient (r ) between the concentration of the
reduced sulfur compounds and the  odor threshold varied from 2
0.00046 to 0.171.  Similarly,  the correlation coefficient (r  )


                               41

-------
varied from 0.011 to 0.190 when comparing the odor threshold
determined for the sample bag to that calculated from the re-
duced sulfur gas analysis.  These calculated values for the
correlation coefficient were deemed low and indicated no corre-
lation between the measured reduced sulfur concentration in the
sample and its odor thresholds.

     It was further determined that the individual reduced sulfur
gases were not additive in their contribution to the perceived
odor threshold when using this methodology.

PRELIMINARY ODOR THRESHOLD MEASUREMENT BY THE HEAD SPACE
ANALYSIS TECHNIQUE

     With the lack of variation in odor threshold measured when
employing either the "total odor" or "stripping" techniques for
sample generation and using the olfactometer, use of the head
space method of odor analysis was investigated.  The procedure
used in this preliminary study utilized seven flasks of differ-
ent concentrations of the stream of concern arranged in ascend-
ing order with flasks containing odor-free water interspersed
among them.  The panelists were asked to identify the first
flask in which they could detect an odor.

     The results, as shown in Table 8, indicated that detectable
odor thresholds determined by this technique varied over a range
of five orders of magnitude.  These values were more in order
with the investigators observed odor intensity of the head space
of these streams in containers.  For this reason a head space
analysis procedure was selected for use in this study.

     TABLE 8.  PRELIMINARY HEAD SPACE TECHNIQUE: ODOR
                  UNIT RANGE BY SOURCE
                                              Range  (xlO   )
     Source
Log ave.
Odor units
ml sample
liter air
Digester condensates
Lab aerated digester condensate
Clarifier outlet
ASB effluent
           0.0045     - 0.049
           0.000028   - 0.00056
           0.00000036 - 0.00000081
           0.00000039 - 0.0000019
THE EFFECTS OF INITIAL STOCK SOLUTION CONCENTRATION AND  ODOR
STABILITY ON THE DETECTABLE ODOR LEVEL  IN THE  HEAD SPACE ANALYSIS

     As the early trial period progressed,  it  became  obvious to
the investigators that the response of  the  panelists  may have
been dependent upon the original sample stock  solution concentra-
tion.  Therefore, different detectable  odor thresholds could be
                                42

-------
U)
              TABLE 9.  EFFECT OF  STOCK  SOLUTION DILUTION LEVEL ON ODOR THRESHOLD
                        VALUES AS  DETERMINED  FOR MEE SURFACE CONDENSER CONDENSATE
                                       FROM MILL C BY PANELS OF SIX

Run
1
2
3
4
5
6
7
8
Number of
panelists
per flask
2x2x2
3x3
x6
x6
x6
2x2x2
3x3
x6
Stock
dilution
3.
3.
3.
3.
3.
2.
2.
2.
2 x
2 x
2 x
2 x
2 x
56x
56x
56x
10
19
10
10
10
10
10
10
10.
•
10
•
•
10.
•
10.
10.
*
17.
•
17.
17.
•
1
1
1
1
1
1
1
1
Dilutions of 1
odor threshold
(xlO D)
5,
3,
I,
4,
1,
13,000,000,
5,300,000,
230,000,000,
300,
300,
400,
000,
400,
000,
000,
000,
000
000
000
000
000
000
000
000
dumber of panelists
missed
high
2
0
0
1
1
0
2
1
low
2
0
1
0
2
1
0
0

-------
measured for the same source.  Leonardos, et al., (24) called
attention to this phenomena in their work.

     To learn if the above effect was biasing the results, the
initial stock solution concentration for the preparation of a
series of samples was changed.  A trial was run utilizing
multiple-effect evaporator surface condenser condensates.  The
panelists were givenna sample prepared from stock solution
dilutions of 3.2x10  :1 and 2.56x10  :1.  The head space
analysis was set up for the forced triangle test, that is one
flask in three at each concentration level to which the panelists
were subjected contained the samples.  This allowed a factor of
243 difference between the most concentrated and weakest dilution
in the series.  The results shown in Table 9 indicated that there
was a difference recognizable to the panelists between the blank
flasks and the flasks containing sample since the dilution level
at which a response to odor was detected could be calculated for
both dilution series.  The calculated ED_Q appeared to vary in
proportion with the dilution factor of tne stock solution.
     The results indicated that two,stock solutions, one at
3.2x10  :1 and the other at 2.56x10  :1 which were made from the
same sample and given to the panelists for evaluation yielded
distinctively different but seemingly valid odor thresholds.
This indicated that the calculation of an odor detection level
by determining which flask smelled "different" was not a valid
testing procedure for the purposes of this study.

Odor Stability

     Table 9 presents one additional piece of information, the
column entitled "Number of Panelists per Flask."  This is the
number of panelists that smelled a single sample from identically
prepared flasks and the number of sample flasks used at each
concentration level during that run.  For example 2x2x2 indicates
that six panelists participated in that run, two panelists each
per sample flask were used prior to exchanging that flask for an
identically prepared sample, and a total of three flasks were
used at each dilution level.  Additionally, 3x3 means 6 panelists
were used, two identically prepared sample flasks were utilized,
being exchanged when three panelists had used them.  Finally, x6
signifies that six panelists were used and all six made their
odor determinations using the same flask containing the sample at
each concentration level.  These data indicated that there was no
significant variation in the detectable odor threshold for the
effluent used in this study by six or less panelists using the
same set of flasks.  However, the head space analysis method for
determing odor thresholds is prone to the loss of volatile com-
pounds which could be responsible for the odor and the above is
not intended to be applied as a general rule.
                                44

-------
REASONS FOR CHOOSING IDENTIFIABLE  ODOR LEVEL  OVER DETECTABLE
ODOR LEVEL IN THE HEAD SPACE  ANALYSIS  TECHNIQUE

     For the reasons previously  explained and shown  in  Table  9,
namely the original sample  stock solution concentration could
bias the determined odor  threshold level,  alternate  approaches
were considered.  That selected  was the identifiable odor  thresh-
old.  In this procedure the panelists  were required  to  smell  the
odor of the diluted sample  prior to smelling  the  test flasks  and
were instructed to respond  to the  odor in the flask  that first
smelled similar to the sample odor.

     The head space analysis  data  was  generated after refinement
of the techniques for this  procedure and conditioning the  panel-
ists to the test and mode of  operation.  Since there was a factor
of three difference in the  sample  concentration between dilution
stages, any results within  a  factor of three  were considered  as
virtually the same value.

NUMBER OF DILUTION STAGES EFFECT ON THE IDENTIFIABLE ODOR
THRESHOLD

     Since this project addressed  odor, or volatile  constituents
of the mill process streams and  often  required dilutions of
several orders of magnitude prior  to testing,  there  was concern
for the ability to maintain the  integrity of  the  sample during
dilution.  The same sample, subjected  to different dilution
sequences might actually  yield a different value  for the ED^-
due to systematic errors.

     To determine what effects,  if any, might be  caused by the
number of dilution stages to  which the sample was subjected,  an
investigation was undertaken.

     Data presented in Table  10  shows  that the number of dilution
stages used in preparation  of the  sample did  not  have a signifi-
cant effect on the identifiable  odor threshold.   The only  excep-
tion, was the last sample of  vapor compression evaporator  conden-
sate from Mill I (runs 8a and 8b).   However,  the  bulk of the  data
indicated the number of dilution stages was of little or no con-
cern in the preparation of  samples. Thus,  the results  indicated
that there were no systematic errors associated with the number
of dilution stages used in  preparation of the sample.   Even when
diluted in three different  manners, the ED5Q  values  measured  were
similar.

     A brief explanation  of the  table  format  is in order.  Table
10 is similar to most presented  in the remainder  of  this report.
The column entitled "ED^ Dilution Factor"  is the column of
interest and signifies the  dilution at which  50 percent of the
panelists identified the  characteristic odor  of the  sample.
These values have all been  multiplied  by a factor of 10   prior

                               45

-------
                   TABLE 10.  EFFECT OF DILUTION STAGES ON IDENTIFIABLE ODOR
(Tt

Run Sample Mill
la S.C. cond. C
Ib
2a S.C. cond. C

3a S.C. cond. C
3b
3c
3d
4a Cone. cond. I
4b
5 a Cone. cond. I
5b
6a Cone. S.C. I
6b cond.
7a VCE cond. I
7b
7c
8a VCE cond. I
8b
No. of
dil.
stages
2
3
2
3
2
2
2
3
2
3
1
2
1
2
1
2
3
2
3
No. of
panel .
9
8
8
9
7
9
9
8
8
8
8
8
8
8
8
8
8
8
8
No. of
panel
ident.
odor
at all
dil.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Maximum EDso No* °^
test dil. panel.
dil. _ factor missed
(xlO ) (xlO~b) low
21
12
17
19
19
22
22
14
1.8
0.85
0.12
0.16
1.2
0.81
16
12
24
16
2
0
2
1
2
1
1
1
2
0
0
0
0
0
2
0
0
0
0
0
Min. dil
at level
missed
(xlO b)
_
2.
2.
2.
2.
2.
2.
2.
__
-
_
-
_
0.
_
—
-
_
—

3
3
3
3
3
3
3





12






-------
to entry.  The number of panelists  that  participated  in  the  test
is also indicated.  On a few  samples,  some  panelists  either  iden-
tified the odor at all six dilution levels  or  failed  to  identify
it at all dilution levels.  Criteria were established to deter-
mine the validity of a sample run as follows:  (a)  if  more than
one-half of the panelists failed to identify the odor in the
range of dilutions, the run was discarded,  (b)  if  two or more
panelists in a team of six, seven,  or  eight, or three or more
panelists in a team of nine or ten  identified  the  odor at all the
dilution stages, the ED5_ was noted as "greater than" (  ) the
calculated value using all panel participants,  and (c) if two
or more panelists in a team of six, seven,  or  eight,  or  three or
more panelists in a team of nine or ten  failed to  identify the
odor at all the dilution stages, the ED5Q was  noted as "less
than" ( ) the calculated value using all panel  participants
observations.

     An entry in the "Maximum Test  Dilution" column signifies the
lowest effluent concentration of that  run at which one or more of
the panelists detected the identifiable  sample odor at all test
dilutions.

     An entry in the "Min. Oil. at  Level Missed" column  signifies
the highest effluent concentration  of  that  run at  which  one or
more of the panelists failed  to detect the  identifiable  sample
odor at that concentration.

STOCK DILUTION LEVEL EFFECT ON THE  IDENTIFIABLE ODOR  THRESHOLD

     The panelists were required to smell the  odor of the diluted
sample prior to smelling the  test flasks and were  instructed to
respond to the odor in the flask that  first smelled similar to
the sample odor.  This proved to be a  more  reliable value since
it could be duplicated when starting with different stock dilu-
tions as shown in Table 11.   The only  discrepancy  noted  was in
run 1 and appeared to be associated with the number of panelists
per bottle rather than the initial  dilution of the stock solution.
The discrepancy, however, was not considered significant.  Runs
2 and 3 displayed essentially the same "identifiable" odor thresh-
old when the dilutions were prepared from different stock solu-
tions.  The results also indicated  that  as  many as 7  panelists
could be run per dilution series without affecting the results.

RESULTS OF DUPLICATE SAMPLES  ON IDENTIFIABLE ODOR  THRESHOLDS

     Throughout this study, duplicate  runs  were made  on  individu-
al samples.  Duplication, in  this context,  means the  individually
prepared dilution series of the sample was  presented  to  the
panelists two, three, and four times during the day.

     The results of these duplicate runs are shown in Table  12.
The last column of the table  was derived by dividing  the largest

                                47

-------
         TABLE 11.   EFFECT OF STOCK SOLUTION DILUTION LEVEL ON IDENTIFIED ODOR  THRESHOLD
                      VALUES AS DETERMINED FOR TREATED TOTAL MILL EFFLUENT FROM MILL  C
OO

Run
1



2


3



No. of
panel. Stock
per flask dilution
x6
3x3
x6
3x3
x7
3x4
x7
3x4
x7
3x4
x7
100:
100:
25:
25:
500:
500:
50:
100:
100:
25:
25:
1
1
1
1
1
1
1
1
1
1
1
No. of
panelists
ident. odor
at all dil.
0
1
0
3
0
1
0
0
0
0
1
Maximum EDso Min. dil.
test dilution No. of at level
dilution factor panelists missed
(xlO D) (xlO~D) "missed low" (x!0~D)
0.
0.42 0.
0.
0.11 0.
0.
2.1 0.
0.
0.
0.
0.
0.11 0.
0083
095
0059
043
021
053
026
013
020
022
019
0
0
0
0
1 0.0029
0
0
0
0
0
0

-------
ED5Q recorded on a particular sample by the smallest ED_.  record-
ed for that same sample.   The largest factor between any"0two runs
was 3.34 for run 23.  The  smallest difference between two  runs
was a factor of less than  1.1 for run 2.   The average factor be-
tween duplications on 23 runs was 1.8.

     Results of these tests  indicated that the panels were
capable of duplicating  their observations  on samples prepared in
a similar manner.

STORAGE STABILITY OF ODOR  IN SAMPLE BOTTLE

     There was concern  for the sample stability when shipped
across the country which required several  days transit time.
Tests were run to determine  the storage stability of samples
that were analyzed several days after collection.

     The results of this study are presented in Table 13.  The
study can be broken into two segments,  the first, runs 1 through
4 and the second, runs  5 through 7.  In the first four runs,  the
samples analyzed on the second day were from the same bottle.
The second set of three runs were analyzed on duplicate samples
such that the first time each bottle was opened was  the day  on
which it was analyzed.

     The first four runs gave inconclusive results.   The first
two samples indicated that the odor intensified in the sample
bottles after exposure  to  air several days prior to  analysis.
This was opposite of the anticipated results.   The potential  for
oxidation of sulfides and  loss of volatiles through  the air  space
above the sample would  be  expected to yield an identifiable  odor
threshold at a lower dilution.  Runs 3  and 4 indicated storage
stability of previously opened sample bottles was good.

     These mixed results prompted additional study of sample
storage stability, hence runs 5, 6, and 7.  These three samples
were collected in triplicate, refrigerated and tested after  one
or two, seven, and ten  days  of storage. Only two bits of  data
might appear irregular. The first, the turpentine decanter  under-
flow value obtained on  March 14, was somewhat high when compared
to the low "identifiable"  odor threshold of March 8.   However,
since both of these values were within a factor of three of  the
median threshold measured  on March 17,  the concern was minimal.
The second was also minor  and concerns the data generated  from
the multiple-effect  (MEE)  hot well sample  analyzed on March  17.
The value recorded is low, but was also noted as "greater  than"
the value presented.  Both the high and the low values from  this
set of data were within a  factor of three  from the mean.

     It was concluded from this exercise that samples could  be
stored in completely filled  glass containers under refrigerated
conditions for periods  of  up to ten days.   During this study a

                                49

-------
                          TABLE 12.  DUPLICATE  SAMPLE IDENTIFIABLE ODOR DETERMINATIONS  ON THE SAME DATE
o
Run
1

2

3


4

5

6

7

8

9


10

11

12

13


Sample Mill
Surface condenser C
condensate*
Surface condenser C
condensate*
Surface condenser C
condensate +


Surface condenser C
condensate
#3 MEE body C
condensate
Chlorination stage C
filtrate
Chlorination stage C
condensates
Concentrator I
condensate
VCE condensate I


Cone. S.C. I
condensate
Concentrator I
condensate
VCE condensate I

Stripper feed F


Date
analyzed
1/20/78

1/20/78

1/20/78


1/26/78

1/26/78

2/02/78

2/02/78

2/09/78

2/09/78


2/10/78

2/10/78

2/10/78

2/14/78


No. of
panel .
9
8
8
8
7
8
9
9
7
7
7
7
6
6
6
6
8
8
8
8
8
8
8
8
8
8
8
7
7

No. of
panelists
ident. odor
at all dil.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Maximum ED n Min. dil. Largest
test dilution No. of at level -IPgo 	
dilution factor panel. misse^ smallest
(xlO~ ) (xlO~ ) missed low (xlO ) EDcn
ou
21
<12
17
<19
19
22
22
1
2
0
0
<0
0
11
4
1
0
12
24
16
1
0
0
0
2
5
350
350








.9
.6
.021
.017
.00081
.0011

.7
.8
.85



.2
.81
.12
.16
.0
.3



0
2
1
2
1
2
1
1
0
0
0.
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
1
1

-
2.
2.
2.
2.
2.
2.
2.
_
-
_
-
_
-
_
-
_
-
_

-
_
0.
_
-
_.
-
120
120


3 1,8
3
3 <1,1
3
3
3
3 >1.6

1.4

1.2

1.4

2.3

2.1


2.0

12 >1.5

1.3

2.7

1.2
(continued)

-------
Ut
TABLE 12.
Run
14

15

16

17

18

19


20

21

22

23




Sample Mill
#2 concentrator D
condensate
f2 MEE body D
condensate
#2 MEE body D
condensate
#4 MEE body D
condensate
VCE condensate J

Turpentine
decanter J
underflow
#2 MEE body E
condensate
#2 MEE body B
condensate
Decker filtrate A
(black liquor)
Turpentine
decanter E
underflow


Date
analyzed
2/16/78

2/16/78

2/17/78

2/17/78

2/28/78


2/28/78

3/02/78

4/07/78

4/12/78


4/12/78



No. Of
panel .
8
8
8
8
7
7
7
7
6
6

6
6
8
8
7
7
6
6

6
6


(continued)
No. of Maximum EDSQ
panelists test dilution
ident. odor dilution factor
at all dil. (x!0~3) (xlO"3)
0
0
0
0
0
0
0
0
0
1

0
0
0
0
0
0
2
2

0
0


0.44
1.2
0.87
0.69
0.89
2.1
0.89
1.3
540
8400 940

63
58
2.2
2.7
<13
<16
2.1 >0.23
21 >0.76

< 36
120


Min. dil. Largest
No. of at level ED5Q
panel. missed smallest
missed low (xlO~ ) EDso
1
0
0
0
0
0
0
0
0
0

0
0
0
0
3
3
0
0

2
1


0.058
2.7
_
1.3
_
2.4
_
1.5
_
1.7

-
1.1
_
1.2
5.8
5.8 1.2
_
3.3

20
20 M..3
n = 23
x = 1.2
           l/'3  of  panelists/flask,  3 flasks/test
           1/2  pf  panelists/flask,  2 flasks/test
           All  panelists on same flask,  1 flask/test

-------
                                         TABLE 13.  STORAGE STABILITY STUDIES ON MILL E SAMPLES
01




Run
la
Ib
2a
2b
3a
3b
4a
4b
5a
5b
5c
6a
6b
6c
7a
7b
7c
*
+





#2 MEE

« 5 MEE

MEE hot





Sample
body cond . *

body cond . *

well*

Turp. decanter


Date

collected
3/2/78

3/2/78

3/2/78

3/2/78
underflow*
#3 MEE


body cond.+


Turp. decanter
3/7/78


3/7/78
underflow*

MEE hot


Samples
Samples

well*


not kept sealed
kept sealed pric

3/7/78


prior to
>r to use


sampled

analyzed
3/3/78
3/7/78
3/3/78
3/7/78
3/3/78
3/7/78
3/3/78
3/7/78
3/8/78
3/14/78
3/17/78
3/8/78
3/14/78
3/17/78
3/9/78
3/14/78
3/17/78
use on second
on second and


No. of
panelists
No. of ident. odor

panel
8
8
8
8
8
8
8
8
7
9
.6
7
9
6
7
9
6
date
third

at all dil.
0
2
1
3
1
1
2
2
0
0
0
0
0
0
0
0
2
listed
Maximum
test
dilution
_c
(X10 3)
-
420
4.2
4200
840
840
8400
8400
_
—
—
-
—
-
_
-
8.4

date -(individualized samp]
ED,-

dilution
factor

£
(xlO~J
1.
<16
1.
>470
55
79
>1200
>1100
5.
0.
2.
72
120
88
9.
15
>2.

les for
8

1





4
75
2



0

4

each

No. of
panelists

missed low
0
1
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0

source and
Min. dil.
at level
missed
_s
(xlO 3)
-
0.58
-
-
_
-
-
-
1.2
0.12
-
12
—
-
_
-
—

date)

-------
total of 181 different  samples  were processed and only 27  were
stored for a period exceeding eighteen hours.

SUMMARY OF HEAD SPACE ANALYSIS  PROCEDURE EVALUATION

     The procedure evaluation indicated that the  panelists should
be required to express  their response on the basis of  the  dilu-
tion stage at which an  identifiable odor similar  to that of the
sample was recognized.   The preliminary investigation  showed that
for these samples as many  as 6  panelists could be used on  a
single solution series  without  the solution losing its odor integ-
rity.  The storage stability study indicated that samples  collect-
ed in completely filled glass bottles could be stored  under re-
frigerated conditions for  periods  of up to ten days.

IDENTIFIABLE ODOR THRESHOLDS OF PROCESS LIQUOR STREAMS

White Liquor

     Samples of white liquor were  obtained for odor analysis.
This stream is the pulp cooking liquor and has an in-process  flow
rate of about 3100 liters/metric ton (750 gal/ton)  at  15 to 20
percent solids is an in-process stream and loss to the sewer  is
only intermittent through  spills or other process abnormalities.
Two samples were obtained  from  different locations.  The identi-
fiable odor threshold varied from  about 2.0x10 to 4.0x10   as
shown in Table 14.

Green Liquor

     Samples of green liquor were  also obtained for odor thresh-
old determination.  This in-process flow results  from  dissolution
of the inorganics in the heat recovery furnace smelt and is the
first step in cooking liquor manufacturing step.   It represents
a flow at about 3750 liters/metric ton (900 gal/ton) at about 15
to 20 percent solids and losses are only due to spills or  process
irregularities.  Three  samples  were obtained from two  different
mills.  The identifiable odor threshold values measured from  these
samples varied from less than 2.6x10  to as high  as 6.3x10   as
shown in Table 14.

Weak Wash

     The third in-process  stream analyzed was the weak wash or
that liquid stream recycled in  the causticizing system. The  in-
process flow of this stream is  about 4,200 liters/metric ton
(1,000 gal/ton) at 4-9% solids. Two samples were obtained from
different mills.  The results as shown in Table 14,  indicated
that the identifiable odor threshold had a range  of 9.0x10  to
4.0x10.
                                53

-------
TABLE 14.  IDENTIFIABLE ODOR THRESHOLD OF IN-PROCESS STREAMS



Mill
Run ident.
White liquor
1 C
2 E
Green liquor
3 C
4 C
5 E
Weak wash
6 C
7 F
Black liquor
8 C
9 C


No. of
panel.

6
6

6
8
6

8
7

7
7
No. of
panelists
ident. odor
at all dil.

0
0

0
2
0

0
0

0
0
Maximum
test
dilution
(xlO~5)

—
-

-
8400
—

-
-

-
—
ED50
dilution
factor
(xlO~5)

200
410

<26
>170
630

90
400

0.31
0.10

No. of
panelists
"missed low"

0
0

2
0
0

0
0

0
0
Min. dil.
at level
missed
(xlO~5)

-
-

5.0
-
—

-
-

-
—

TABLE 15.
IDENTIFIABLE ODOR THRESHOLD OF NON-COMBUSTIBLE
LINE CONDENSATE



Mill
Run ident.
1 E
2 E


No. of
panel .
6
7
No. of
panelists
ident. odor
at all dil.
0
0
Maximum
test
dilution
(xlO~5)
—
—
ED50
dilution
factor
(xlO~5)
1,100
11,000

No. Of
panelists
"missed low"
0
1
Min. dil.
at level
missed
(xlO~5)
_
1200

-------
Streams Containing Black  Liquor

     These were effluent  streams  that are associated  with  the
pulp washing or screening process.   Samples  were  collected from
two different locations at one  mill.   The first sample  was col-
lected after the final stage  of conventional pulp washers  and
the second was collected  after  the  decker.   At this location the
decker was used as a  last stage of  pulp washing.   The filtrate
from the decker was used  as wash  water on the last conventional
washer stage.  The soda content in  the pulp  from  this "decker"
averaged about five kilograms per metric ton (10  Ibs/ton) .  The
black liquor flow at  the  last conventional stage  brown  stock
washer was measured at about  10,000 liters/metric ton (2,300 gal/
ton) at less than 1 percent solids  from the  "decker"  filtrate.

           5Q identified  odor threshold values obtained were
            the washer sample and 1.0x10  for the sample fro
         ED
3.1x10  for the washer  sample  and  1.0x10  for the samle from the
"decker" as shown in Table  14.
Summary of Process Liquor Streams  Identifiable Odor Threshold
Measurements

     As a category, these in-process  liquor  streams generally had
the highest identifiable odor  thresholds of  any of the streams
measured during this  survey.   Identifiable-odor thresholds were
measured in the range of 2.6x10  to 6.3x10   with the exception of
the sample containing black  liquor with ED,-n values found of
1.0xlOJ and 3.1x10 .

     Distinctively, these process  streams did not smell particu-
larly strong prior to analysis.  However, after dilution it
became apparent that  the odor  intensities increased.  All four
streams are normally  basic and contain high  concentrations of
sulfides in solution.  Dilution  of these process streams resulted
in lowering of pH, the liberation  of  sulfides and an increase in
the odor threshold of the process  stream.  This is similar to the
process that would occur at  the  mill  site upon addition of these
liquors to the mill's sewer  system and treatment plant.

IDENTIFIABLE ODOR THRESHOLD  OF OTHER  PROCESS STREAMS

Noncondensible Line Condensates

     Two samples of condensate from a noncondensible gas handling
system were obtained  from a  544  metric ton/day (600 ton/day)
kraft mill with about 200 metric ton/day (225 ton/day) NSSC pulp.
The condensates were  from the  noncondensible line conveying kraft
batch digester combined with multiple-effect evaporator noncon-
densibles to the lime kiln.  The samples collected had identifi-
able odor thresholds  of 1.1x10*  and 1.1x10   as shown in Table 15.
These were the highest of any  single  process stream measured dur-
ing this survey, however, the  flow was minimal.

                               55

-------
Turpentine Decanter Underflow

     Nine analyses were run on turpentine decanter underflow  from
six different mills including two from the southeastern United
States,  The log average of the identifiable odor threshold was
7.2x10  with a miximum of 1.2x10  and a minimum of 2.9x10  as
shown in Table 16.

     This source had a flow varying from 124 liters/metric ton at
mill I to 650 liters/metric ton at mill H (30 gal/ton to 156
gal/ton).  Mill E, which had the highest dilution factor for the
identifiable odor threshold, was a West Coast mill which used
relatively fresh chips.  Mills A and B, southeastern mills/
tended to require a high dilution before reaching the EDCQ iden-
tifiable odor threshold.

Vapor Compression/Recompression Evaporator Streams

     One of the more recent innovations in black liquor evapora-
tion has been the use of vapor compression or recompression
evaporators as a first stage of black liquor evaporation.  These
systems usually use excess steam from another source.  The units
typically have two effluent streams, the steam condensate and
the vapor condensate.

     The vapor condensate stream typically required a high dilu-
tion to odor threshold.  The flow of this stream was about 250
liters/metric ton  (60 gal/ton) of pulp production.  As shown in
Table 17, the identifiable odor threshold of the vapor condensate
varied from 2.0x10  to 9.4x10  with a log average of 3.0x10 .

Pulp Cleaning Effluent

     Six samples containing decker filtrate were analyzed from
three different mills.  The results are presented in Table 18.
The samples were only collected from deckers following screen
rooms.
                                                        2
     The identifiable odor thresholds varied from 7.8x10  to
>4.2xlO .  Mills A and B were bleached kraft mills and mill E
was a linerboard mill.  The variations in the identifiable odor
thresholds noted at mill A, 7.8x10  for run 1 and >7.6xlO  for
run 2, could be attributed to a suspected increase in black
liquor carryover since the conductivity of the sample for run 2
was approximately four times that of run 1.  At mill B, the vari-
ation in the identifiable odor threshold of 9.6x10  to 8.1x10 ,
might have been partially attributed to the pine pulp on the
first decker and hardwood pulp on the second decker.

     Mill E was a kraft mill with some neutral sulfite pulping.
The kraft pulp following the three drum, four stage brown stock


                                56

-------
          TABLE 16.   IDENTIFIABLE ODOR THRESHOLD OF TURPENTINE DECANTER UNDERFLOW
01
-j

Run
1
2
3
4
5
6
7
8
9
10
11

Mill
ident.
A
A*
B
E
E
H
I
I
J
J*
J

No. of
panel .
6
6
6
8
8
6
8
8
6
6
7
Log
No. of
panelists
ident. odor
at all dil.
0
0
0
1
2
0
0
0
0
0
0
average
Maximum EDso
test dilution
dilution factor
(xlO D) (xlO D)
<36
120
870
84,000 1100
8,400 >1200
160
3.5
2.9
63
58
7.2
72
No. of
panel
"missed
low"
2
1
0
2
0
0
1
0
0
0
1

Min. dil.
at level
missed
(x!0~°)
20
20
-
120
—
-
1.2
-
_
—
1.2

Flow
1 /metric
ton


0.088

0.14
0.46

0.070


0.27


         Duplicate run - log average of duplicates used in column log average

-------
              TABLE 17.   IDENTIFIABLE ODOR THRESHOLD OF VCE CONDENSATES
Ul
00



Run
1
2
3
4
5
6
7

8
9
10


Mill
ident.
F
F
I
I*
I*
I
I*

J
J*
J


No. of
panel .
7
6
8
8
8
8
8
Log
6
6
7
Log
No. of
panelists
ident. odor
at all dil.
1
0
0
0
0
0
0
average
0
1
0
average
Maximum
test
dilution
(xlO~5)
840
—
—
-
—
—
-

_
8400
—

ED50
dilution
factor
(xlO~5)
22
110
12
24
16
2.0
5.3
7.4
540
940
8.2
30
No. of
panelists
"missed low"
0
0
0
0
0
0
0

0
0
0

Mm. dil
at level
missed
(xlO~5)
_
-
-
-
-
—
-

_
—
-


        Duplicate  or triplicate run-log average used in column log average

-------
     TABLE  18.   IDENTIFIABLE ODOR THRESHOLD OF DECKER FILTRATE
,
Run
1
2
3
4

5
6

No. of
panelists Cone, of ED5f)
Mill No. of ident. odor highest dilution
ident. panel at all dil. dilution factor
A
A*
B
B*

E
E

6
6
6
6

6
7
Decker
1 0.21
2 21
0
0

6 42
0
filtrate log average
0.0078
>0.76
9.6
0.0081

>42
18
0.84
No. of
panel .
"Missed
low"
0
0
0
0

0
1

Min. dil. Conduc-
at level tivity
"missed" ( mhos /cm)
670
2500
610 (pine)
325 (hard-
wood)
0,1400
0.58 0,1400


Second decker at same mill

-------
washer was diluted prior to a refiner, low density storage and
screening followed by a decker that was operated as a thickener.

Black Liquor Concentrator Condensates

     There are several sources of condensate generation within
the kraft pulping black liquor evaporation system.  One is con-
densate from the forced circulation concentrator.  These rarely
contributed more than about 400 liters per metric ton of pulp
production (100 gal/ton) to the total mill effluent and represent
condensate obtained from the last stage of liquor concentration.
Concentrators normally increase the liquor solids from about 50
to 65 percent.

     Seven samples were obtained from three mills and four dif-
ferent forced-circulation concentrators.  These sources included:
(a) mill B number 2 concentrator, (b) mill D number 2 concentra-
tor, and (c) mill I concentrator surface condenser condensate
and, (d) mill I concentrator condensate.

     The identifiable odor thresholds of all four of these pro-  ,
cess streams were low and similar ranging from 3.7x10  to 8.5x10 .
The log average of these twelve samples was 5.1x10  as shown in
Table 19.

Black Liquor Multiple-Effect Evaporator Total Condensate

     The combined multiple-effect evaporator condensate stream
had an average flow of 5,400 to 10,200 liters/metric ton (1,300
to 2,500 gal/ton or 520 to 750 gla/1,000 Ibs of black liquor
solids).  Sixteen samples of the combined multiple-effect evapo-
rator condensates were obtained from five different mills for
odor threshold analysis.  The results indicated that the identi-
fiable odor thrshpld varied from 1.4x10  to 3.1x10  with a log
average of 2.5x10 .  These results are shown in Table 20.

     Run 4 was not included in the general data summary as that
sample was obtained during a period of multiple-effect evaporator
boil-out.  The high conductivity observed in that sample was of
some note, since it was a factor of almost 20 greater than the
conductivity of other similar samples measured during this study.

     The odor values obtained on condensates from mills A and B
were somewhat higher than those reported for the other mills, a
log average of thegidentifiable odor threshold of 1.1x10  as
compared to 1.9x10 .  When comparing the sample value7ranges, the
Southern mills A and B, varied from 3.8x10  to 2.2x10 , whereas
the West Coast mills varied from 4.3x10  to 3.1x10 .  The conduc-
tivity values did not indicate a difference in liquor carryover.
The difference may have been due to the difference in wood
species pulped.
                               60

-------
    TABLE  19.   IDENTIFIABLE ODOR THRESHOLD OF FORCED CIRCULATION CONCENTRATOR CONDENSATES
Run
1
2
3
4
5
6
7
8
9
10
11
12

Mill
ident.
B
B*
D
D*
D
I
I
I*
I
I*
I
I*

Source
#2 Cone. evap.
#2 Cone. evap.
#2 Cone. cond.
#2 Cone. cond.
#2 Cone. cond.
Cone. S.C. cond.
Cone. S.C. cond.
Cone. S.C. cond.
Cone. cond.
Cone. cond.
Cone. cond.
Cone. cond.

No. of Maximum
panelists test
No. of ident. odor dilution
panel. at all dil. (xlO~5)
6
8
8
8
7
8
8
8
8
8
8
8

0
0
0
0
1 8.4
0
0
0
0
0
0
0
Log average (all sources)
difStion
factor
(xlO~5)
0.56
0.24
0.44
1.2
1.0
0.21
1.2
<0.81
1.8
0.85
0.12
0.16
0.51
Min. Dil.
No. of at level
panelists missed
"missed low" (xlO~5)
1
0
1
0
0
0
0
2
0
0
0
0

0.058
—
0.058
-
-
_
-
0.12
_
-
_
-

Duplicate sample, only log average of duplicates utilized in mill log average

-------
                       TABLE 20.  IDENTIFIABLE ODOR THRESHOLD OF COMBINED MULTIPLE-EFFECT EVAPORATOR CONDENSATES
CTN
Run
1
2
3
4
5
6
7
8
9
10
11

12
13
14
15
16
17


Mill
ident.
A
B
B*
C+
C
C
C
C
C
C
C

D
D
E
E
E
E


No. of
panelists
No. of ident. odor
panelists at all dil.
7
6
7
7
6
5
9
7
6
6
7
Log
8
7
8
8
7
7
Log
Log
1
0
0
0
0
0
0
0
0
0
0
average mill C
1
0
2
1
0
0
average mill E
average, all mills
Maximum
test
dilution
(xlO~s)
4200
_
-
_
-
8.4
-
-
-
-
-

84
-
42
420
-
-


ED50
dilution
factor
(xlO~5)
220
<83
38
0.45
4.7
>4.3
4.4
4.3
89
310
43
' >17
6.2
1.4
>10
55
12
5.4
14
24
Min. dil.
No. of at level
panelists missed
"missed low" (xlO )
0
0 12
0
1 0.058
0
0
0
0
0
0
1 5.8

0
0
0
0
0
0


Conductivity
( mhos /cm)
270
240

•WOOO
390
260
260
350
-
-
-

330
430
485
290
200
255


                  Duplicate sample (same sample, same day) only log average of duplicates  utilized  in mill log average
                  Multiple-effect evaporator on boil-out, not included in mill log  average

-------
                     TABLE  21.   IDENTIFIABLE ODOR THRESHOLD OF MULTIPLE-EFFECT EVAPORATOR COMPENSATES PROM MILL A
U>


No. of
panelists
Maximum EDso
test dilution No. of
Min. dil.
at level
No. of ident. odor dilution factor panelists missed
Run
1

2
3
4
5
6
Source
#1 MEE & Cone.
cond.
#2 Body cond.
#3 Body cond.*
#3 Body cond.
MEE hot well
MEE comb. cond.
panelists at all di
comb.
7 3
7 0
7 0
7 0
7 0
7 0
1. (xlO~5) (xlO~5) "missed

0.011 0.00070 1
13 3
- 16 3
3.7 1
610 1
220 1
low" (xlO~5) (

0.000014
5.8
5.8
0.58
230
4200

Cond.
u mhos/cm)

10
300

310
40
270

*


Run
1
2
3
4
5
6
Duplicate of previous run (same sample, same
TABLE 22.

Source
#2 Cone. +
2,3,4,5 MEE
Hot well
(#6 vapors)
Comb. MEE cond

IDENTIFIABLE ODOR THRESHOLD
NO. of
panelists
No. of ident. odor
panelists at all dil.
6 1
8* 0
6 0
7* 0
6 0
8* 0
day)

OF MULTIPLE-EFFECT EVAPORATOR CONDEMSATES FROM MILL
Maximum ^Dgn
test dilution No. of
dilution factor panelists
(xlO~3) (xlO~3) "missed low"
1.1 0.089 0
0.58 1
690 2
510 0
83 2
38 0
Min. dil.
at level

B

missed Cond. Flow
(xlO~ ) ( mhos/cm) (gpm)
- n.d
0.058
120 410
-
12 240
"
•*•
.



460

                 Duplicate of previous run, collected simultaneously but analyzed  for odor  threshold  on different days
                 n.d. denotes not detectable

-------
TABLE 23.  IDENTIFIABLE  ODOR THRESHOLD OF MULTIPLE-EFFECT EVAPORATOR CONDENSATES FROM MILL C


Run
1
2
3
4


Source
#3 Body cond.




No. of
panelists
7
7
6
6
No. of Maximum
panelists test
ident. odor dilution
at all dil. (xlO~5)
0
0
0
1 0.42
Log average #3 body condensate
5
6
7
8
9
10
11
12
13
14
15
16
17
18

19
20
21
22
23
24
25
26

*
+
#4 Body cond.

#5 Body cond.

#6 Body cond.

#7 Body cond.

S.C. cond . *
S.C. cond . *
S.C. cond . *
S.C. cond . *
S.C. cond.
S.C. cond.
Log average
Comb . cond .







Log average
6
6
6
6
6
6
6
6
_
-
-
-
6
6
0
1 0.42
_ _
1 0.42
0
0
0
3 0.21




0
0
dilution No. of
factor
(xlO~5)
0.021
0.017
0.040
0.067
0.031
0.023
0.095
0.12
0.14
0.46
0.83
0.13
0.076
18
18
<25
2.2
12
5.9
panelists
"missed low"
0
0
0
0

0
0
0
0
0
0
0
0




0
0
Min. dil.
at level
missed Cond.
(xlO~5) (ymhos/cm)
_ _
- -
25
15

85
- —
25
-
130
20
125
20


1550
1650
1475
975
Approx .
flow
(qpm)
62




147

272

412

497



123



S.C, condensate <10
7+
6
5
9
7
6
7
7
0
0
2 8.4
0
0
0
0
0
0.45
4.7
>4.3
4.4
4.3
89
43
43
1
0
0
0
0
0
0
0
0.058 ^7000
390
260
_
-
_
_
-
620







combined condensate >11
Log average of daily results,
Multiple-effect
evaporator on
2 to 8 tests per day
boil-out, not included

in source

log average





-------
Components of.Multiple-Effect Evaporator Condensates

     Condensate  samples  were  collected at five different  loca-
tions around the multiple-effect evaporators at the Southeastern
mill A.  At this location a split liquor feed to bodies 4  and 5
was featured.  The  samples included the number 1 multiple-effect
evaporator condensate  mixed with the combined concentrator conden-
sates, the number 2  multiple-effect evaporator body condensate,
the number 3 body condensate, the multiple-effect evaporator hot
well, and the  combined multiple-effect evaporator condensate
which also contained the surface condenser condensate.  The
results from  these  samples, as shown in Table 21, indicated that
the largest portion of the odor could be attributed to those
condensates originating  from  the first stages of liquor evapora-
tion.  The identifiable  odor  threshold of the condensate through
the third body was  only  3.7x10 .  Addition of those condensates
from the fourth  and-fifth bodies increased the identifiable odor
threshold to  6.1x10  .

     Mill B featured a mid-body liquor feed.  The liquor entered
body no. 5 and flowed  sequentially to bodies 6,  4, 3, 2, and 1.
Condensate samples  were  collected from three different locations
in this multiple-effect  evaporator system.  The three samples
included one  containing  the number 2 concentrator and numbers 2,
3, 4, and 5 multiple-effect evaporator condensate, one from the
evaporator hot well which included the vapors from the number 6
body, and one  containing the combined condensates from the total
black liquor  evaporation system.  The data, as shown in Table 22,
indicated that the  condensates from the later stages of evapora-
tion had a minimal  identifiable odor threshold when compared to
those originating from the first stages of evaporation.  Atten-
tion is called to the  data for runs 1 and 2 as compared to runs
3 and 4.  The  results  indicated that the number 6 evaporator body
condensate had an identifiable odor threshold of  5x10 ,  whereas
those condensates from the latter stages of evaporation had an
identifiable  odor threshold of 6.0x10 .

     The multiple-effect evaporators at mill C were operated in a
strictly counter-current flow manner with liquor feed to body 7
and continuing sequentially to 6, 5, 4, 3, 2, and 1.  Samples
were collected from this system at number 3 evaporator body,
number 4 evaporator body, number 5 evaporator body, number 6
evaporator body, number  7 evaporator body, the surface condenser,
and the combined multiple-effect evaporator and concentrator
condensates.

     The results presented in Table 23 indicate that the primary
odor contribution was  from the vapors off the number 7 evaporator
body.  These  vapors were condensed and emerged as condensates
from the surface condenser which had an identifiable odor5thresh-
old log average  of  1.0x10  with variations between 2.2x10   and
about 2.5x10  .  The converse was also true, those condensates

                                65

-------
from the bodies at the latter end of the liquor evaporation
sequence contained much lower identifiable odor thresholds as
exemplified by the number 3 multiple-effect evaporator body
condensates.  This body had a log average identifiable.odor
threshold of 3.1x10  with a range of 1.7x10  to 6.7x10 .

      Multiple-effect evaporator condensate samples were also
obtained from mill D.  These multiple-effect evaporators were
mid-feed units.  The feed was to the 4th effect and then the
liquor progressed sequentially to the 5th, 3rd, 2nd, and 1st
effect prior to entering the concentrators.  Samples were
collected from fivfe different locations in this complex including
the number 2 evaporator body, the number 4 evaporator body, the
number 5 evaporator body, the surface condenser and the hot well
which contained the combined evaporator condensates.

     The results, as shown in Table 24, indicated that the first
stages of liquor evaporation had the highest identifiable odor
threshold dilution level.  This was exemplified in the values
obtained from the,surface condenser condensates which varied from
1.2x10  to 4.5x10 .  Additionally, those condensates from the
latter stages of evaporation had lower identifiable odor thresh-
olds as shown by the data obtained from the number 2 body and
number 4 body condensates.  The number 2 body condensate thresh-
old varied from 6.9x10  to 2.1x10  and the number 4 body conden-
sates varied from 1.8x10  to 1.3x10 .

     Multiple-effect evaporator condensates were analyzed from
mill E and the data is presented in Table 25.  At this location
the liquor feed was split to the evaporators.  The liquor entered
bodies 5 and 6 concurrently, then recombined and sequentially
entered bodies 4, 3, 2, and 1.  Samples were collected from six
different locations at this mill, including the number 2 body
condensate, the number 3 body condensate, the number 4 body con-
densate, the number 5 body condensate, the number 6 body conden-
sate, and the evaporator hot well which was the total evaporator
condensates.

     This set of samples did not follow the expected odor inten-
sity pattern as clearly as at the previous locations.  In an
attempt to understand what may have happened, the conductivity
of the condensates are also presented in the table.  The conduc-
tivities increased as a trend toward the liquor feed end, but not
significantly.  They did not explain the difference in behavior
of this set of samples when compared to the others.  This informa-
tion points to the need to evaluate each source independently.
The identifiable odor thresholds at the number 2 body varied from
 1.8x10  to 2.7x10 .  This increased for the hot well condensates
where the identifiable odor threshold varied from >lxlO  to
5.5x10 .
                                66

-------
     TABLE  24.   IDENTIFIABLE  ODOR THRESHOLD OF  MULTIPLE-EFFECT EVAPORATOR CONDENSATES FROM MILL D
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
Source
#2 Body cond.



#4 Body cond.


#5 Body cond.

S.C. cond.

Hot well
( comb . cond . )
No. of
panelists
8
8*
7
7*
8
7
7*
8
7
8
7
8
7
No. of
panelists
ident. odor
at all dil.
0
0
0
0
0
0
0
0
0
0
1
1
0
Maximum.
test
dilution

-------
                 TABLE 25.  IDENTIFIABLE ODOR THRESHOLD OF MULTIPLE-EFFECT  EVAPORATOR  CONDENSATES FROM MILL E
00
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
Source
#2 Body cond.


#3 Body cond.

#4 Body cond.

#5 Body cond.

#6 Body cond.

Hot well
(comb, cond.)
No. of
panelists
8
8*
8
8
8
8
8
8
8
8
8
.8
8
No. of
panelists
ident. odor
at all dil.
0
0
0
0
0
1
1
0
1
1
1
2
1
Maximum
test
dilution
(xlO~5)
_
-
-
_
-
4.2
42
_
4.2
420
8.4
42
840
ED50
dilution
factor
(xlO~5)
2.2
2.7
1.8
3.4
3.2
1.4
3.1
0.36
1.1
31
0.90
10
55
No. of
panelists
"missed low"
0
0
2
0
0
0
0
0
0
0
0
0
0
Min. dil.
at level
missed Cond.
(xlO~ ) (umhos/cm)
32
—
0.58 40
42
25
100
65
120
135
95
100
485
290
Flow
(gpm)
164


318

408

576

642

829


            Duplicate of previous run (same sample, same day)

-------
                  TABLE 26.  IDENTIFIABLE ODOR THRESHOLD OF MULTIPLE-EFFECT EVAPORATOR  CONDENSATES FROM MILL I
vo
Run
1
2
3
4
5
6
7
8
9
10
11
Source
#3 Body cond.
#4 Body cond.

#5 Body cond.

16 Body cond.

#7 Body cond.

S.C. cond.

No. of
panelists
7
8
7
8
7
8
7
8
7
8
7
No. of
panelists
ident. odor
at all dil.
0
1
0
2
0
0
-
2
0
0
0
Maximum
test
dilution

-------
     The last multiple-effect evaporator complex investigated was
that at mill I which featured a split liquor feed to the 6th and
7th effects with the flow continuing sequentially, to numbers 5,
4, 3, 2, and 1 effects.  Samples were collected from six differ-
ent sources for analysis, including the number 3 body, the number
4 body, the number 5 body, the number 6 body, the number 7 body,
and the surface condenser condensate.

     These samples, with the exception of number 7 body conden-
sates, very nicely follow the progression of identifiable odor
thresholds experienced with previous samples.  The results are
presented in Table 26.  The Ed5Q of the number 3 multiple-effect
evaporator body condensate was the lowest.  The two samples gave
values of 2.4x10  and >3.2xlO .  The highest odor threshold
values were from the surface condenser condensates and the two
samples gave values of 2.1x10  and 7.8x10 .  It was noted that
the condensate from the number 7 body had a lower identifiable
odor threshold than that from number 6 body.  It was determined
that the lower EDj._ was caused by lesser liquor carryover, since
the conductivity or the condensate from the number 7 body was
also lower than that from number 6.

     By carefully observing Tables 21-26, it can be seen that the
vapors from the first two liquor stages of evaporation required
more dilution to reach the identifiable odor threshold.  This is
in agreement with previous findings that indicated most of the
volatile constitutents in black liquor were removed in the first
few stages of the multiple-effect evaporators (62).

Digester Condensates

     The identifiable odor threshold of the digester condensate
streams varied as a function of the condensation method utilized.
At mill C, indirect heat exchangers for condensation were used,
whereas at the other four locations sampled, direct or jet con-
densers were utilized to reduce the temperature of the digester
blow gas prior to treatment.

     At mill C, the identifiable odor threshold of the digester
condensates had a log average of 2g9xlO  as shown in Table 27,
and ranged from 7.0x10  to >1.4x10 . The other four locations
used direct condensation for gas stream temperature control and
the condensates had a log average identifiable odor threshold of
4.8x10  and a range of 3.1x10  to 3.7x10 .

     It was apparent that the direct contact condensers, with the
resultant dilution water yielded condensates with an average
identifiable odor threshold that was lower than those from indi-
rect condensers.  It was assumed that the difference in identifi-
able odor threshold observed for the direct contact condensates
was due to the dilution provided.  The effluent volume associated
with direct condensers is variable and a function of the

                               70

-------
        TABLE 27.  IDENTIFIABLE ODOR THRESHOLDS OF DIGESTER CONDENSATES
-4
I-1

Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Mill
ident.
A
A
B
C*
C*
C*
C*
C*
C*
F
F
G
G
G
G

No. of
panel.
7
7
7
7
6
9
6
6
7
7
6
6
6
7
7

No. of
panelists
.ident. odor
at all dil.
2
0
0
0
0
0
1
2
0
0
0
0
1
0
0
Log average
Maximum
test
dilution
(xlO~5)
84
-
-
_
-
—
4200
8400
-
_
—
_
84
-
—

ED50
dilution
factor
(xlO~5)
1.6
3.0
11
110
70
480
550
>1400
200
7.8
365
2.5
7.6
1.4
0.31
24
No. of
panelists
"missed low"
2
0
0
0
1
0
0
0
0
0
0
0
0
0
0

Min. dil
at level
missed
(xlO~5)
0.12
—
-
_
12
—
—
—
-
_
-
_
—
—
-


         Indirect condenser

-------
individual design.  Effluent flows associated with indirect con-
densers are on the order of 1.3 liter/minute (0.35 gal/min) per
ton of production.

First Chlorination Stage Bleach Plant Effluent

     The first chlorination stage effluent from the bleach plant
had a relatively low identifiable odor threshold.  Seven runs
were performed on six different samples from three different
mills and the data is shown in Table 28.  The log average of the
identifiable odor threshold was 1.2x10  with a range of 8.1x10
to 1.4x10 .

     The information generated in other portions of this study is
of significance when interpreting the information on the relative
odor threshold level of this process stream.  When the odor
threshold levels were determined for the individual components of
these combined streams and a combined odor threshold calculated
it was found somewhat, if marginally, greater than the ddor
threshold of the actual combined stream.  For example, the odor
threshold levels of5the actual combined streams were 0.11x10 ,
1.1x10 , and 5.6x10  while the calculated odor thresholds were
15x10 , 2.5x10 , and 2.3x10  respectively as shown in Table 33.
These data indicate that there was some benefit in mixing these
odorous streams with the bleach plant effluent.

     The run 7  sample from mill C appeared to be an anomaly with
no reason apparent for the high value of 1.4x10  reported.  This
value was 2 orders of magnitude greater than other odor thresh-
olds found from this source during the study.

     It was noted that the odor perceived in the flask at the
identifiable odor threshold level for all bleach plant effluent
samples had a different characteristic than the previewed odor.
The characteristic chlorine odor of the concentrated sample was
not the odor perceived at the identifiable odor threshold level
by the panelists.  These observations indicated that the chlo-
rine odor was not significant or predominant at the threshold
level.

Mixtures of First Chlorination Stage Effluent and Digester
and Evaporator Condensates

     It was estimated that about 29,200 liters/metric ton (7,000
gal/ton) of flow could be attributed to first stage chlorination
effluents.  This flow is mixed with that from the multiple-effect
evaporators and the digester condensates at some mills.  The flow
from these two sewers is on the order of 10,000-11,000 liters/
metric ton (2000-2500 gal/ton).  The purpose of mixing is to
permit the residual chlorine in the bleach plant effluent to
react with sulfides and other odor causing constituents of the
                                72

-------
         TABLE 28.  IDENTIFIABLE ODOR THRESHOLD OF  1ST  CHLORINATION STAGE FILTRATE
OJ



Run
1
2
3
4
5
6
7


Mill
ident.
B
C
C
C*
C
C
C


No. of
panel.
8
7
6
6
9
7
6

No. of Maximum
panelists test
ident. odor dilution
at all dil. (xlO~5)
0
0
0
0
0
0
1 4.2
Log average
ED
dil
iO. .
ution
factor
(xlO~5)
<0.
<0.
0.
0.
0.
0.
1.
0.
0037
0038
00081
0011
013
013
4
012
No. of
panelists
"missed low"
2
3
0
0
0
1
0

Min. dil.
at level
missed
(xlO~5
)
0.000015
0.0017
-
—
—
0.0002
—





9



         Duplicate of previous run  (same  sample,  same  day)  combination  counted as
         one run in  log average

-------
condensate streams in an effort to reduce the odor attributable
to these effluents.

     At one location, mill C, mixing of the above process streams
was normal operation.  Eleven odor determinations were made on
ten samples from this combined stream.  One of these samples was
collected while the multiple-effect evaporators were operated in
the boil-out mode as is shown in Table 29.  The identifiable odor
threshold log average for these combined streams was 4.5x10  with
a range of 1.1x10  to 1.0x10 .  This compared with a log average
of 1.0x10  and a range of 8.1x10  to 1.4x10  for just the first
chlorination stage bleach plant effluent from this mill, an
increase of more than two orders of magnitude.

TABLE 29.  IDENTIFIABLE ODOR THRESHOLD OF BLEACH PLANT ACID SEWER
           COMBINED WITH CONPENSATES FROM MEE AND DIGESTERS AT
                              MILL C



Run
1*
2
3 +
4
5
6
7
S
9
10
11

No. of
panel.
7
6
6
8
6
7
9
7
6
6
7
No. of
panelists
ident. odor
at all dil.
0
0
0
0
1
0
2
1
1
0
0
Maximum
test
dilution
(xlO~b)
_
-
-
-
840
-
4.2
42
42
-
—
ED50
dilution
factor
(xlO~3)
0.20
11
4.7
4.1
100
4.5
0.11
1.1
5.6
6.5
12
No. of
panel.
"missed
low"
0
0
0
0
0
0
0
0
0
0
1
Min. dil
at level
missed.
(xlO '
_
-
-
-
-
—
—
—
—
—
1.2
      Log average                      4.5
*    MEE on boil-out, not included in average
+    Duplicate of previous run (same sample, same day), log
     average of both runs used in log average

Steam and Air Stripper Streams

     Six determinations and one duplicate were made on stripper
feeds from three different locations.  The results are shown in
Table 30.

     The feed to the three systems studied was quite varied and
included such sources as turpentine decanter underflows, digester
condensates and multiple-effect evaporator condensates.  The
identifiable odor thresholds of these feeds7had a log average of
3.8x10  with a range from 2.0x10  to 3.5x10 .
                               74

-------
                 TABLE  30.   IDENTIFIABLE ODOR THRESHOLD OF STRIPPER FEED
-4
U1



Run
1
2
3
4
5
6
7


Mill
ident.
F
F*
F
H+
H+
I
I


No. Of
panel.
7
7
6
6
6
8
7
Log
No. of Maximum
panelists test
ident. odor dilution
at all dil. (xlO~5)
0
0
0
0
0
0
0
average
ED50
dilution
factor
(xlO~5)
350
300
160
146
<24
2.0
7.8
38
Min. dil
No. of at level
panelists missed
"missed low" (xlO~5)
1 120
1 120
0
0
3 12
0
0


          Duplicate of previous run (same sample,  same day), log average used in column
          log average
          Feed to air stripper

-------
     The stripper feed at mill F was the multiple-effect evapo-
rator surface condenser condensates, the turpentine decanter
underflow and the vapor compression evaporator vent condensates
at about 2700 liters/metric ton (650 gal/ton).  At mill H the
stripper feed included the turpentine decanter underflow and con-
densate from the number 5 body of a five body mid-feed multiple-
effect evaporator set for a total flow of about 1600 liters/met-
ric ton (about 400 gal/ton).  At mill I the feed to the stripper
included turpentine decanter underflow, stripper vacuum pump seal
water, concentrator and evaporator surface condenser condensates
and vapor compression evaporator condensates for a total flow of
about 1300 liters/metric ton (300 gal/ton).  These were typical
of some of the most odorous streams identified.

     Six determinations were made on air and steam stripped
materials from three different mills and the data is presented
in Table 31.  The sample from mill H followed an air stripper,
whereas the other four samples followed steam strippers.  This
air stripper was a temporary installation that received about ,
190 liters/minute (50 gpm) of feed condensate and about 22.7 m /
min (800 cfm) of air.  The air to liquid feed ratio was about 40
to 1.  Odor threshold reduction across the air stripper was about
50 percent.  Since this was a temporary unit, it is suspected
that the unit was not optimized and these results may not have
been representative of the capability of these systems.

     Table 31 indicated the identifiable odor threshold of the  ,
product, from the two steam strippers had a log average of 1.4x10
to 2.8x10 .  The odor thresholds of the bottoms from the steam
strippers at both sites were similar despite the difference in
the feed odor thresholds.  This indicated that the odorous
constituents in the effluent streams that were removable by steam
stripping had been volatilized.

SUMMARY OF PROCESS STREAM IDENTIFIABLE ODOR THRESHOLDS

     The range and/or log averages of the identifiable odor
thresholds have been summarized and presented in Table 32.  These
values indicate that the most odoriferous streams, as a class,
were those that were termed in-process liquor steams with a high
pH and high sodium sulfide content.  These included white liquor,
black liquor, weak wash, and green liquor.  Excluding the weak
black liquor samples, these streams had a combined log average
identifiable odor threshold of 2.0x10  with the highest being a
green liquor measurement at 6.3x10  and the lowest being another
green liquor sample at  2.6x10 .

     The various condensate streams originating from handling
black liquor had the next highest identifiable odor thresholds,
having a combined log average of 3.2x10 .  The streams included
in this classification were the turpentine decanter underflow,
                               76

-------
TABLE 31.  IDENTIFIABLE ODOR  THRESHOLD  OP  AIR AND STEAM STRIPPED MATERIAL



Run
1
2
3
4
5
6


Mill
ident.
F
F
H*
H*
I
I


No. Of
panel.
7
6
6
6
7
7
Log
No. of Maximum
panelists test
ident. odor dilution
at all dil. (xlO~5)
1 4.2
0
0
0
0
0
average
ED50
dilution
factor
(xlO~5)
0.28
0.18
60
35
0.12
0.072
0.14
Min. dil
No. of at level
panelists missed
"missed low" (x!0~5)
0
0
1 12
3 12
0
0


Air  stripper  material  not included in column log average

-------
              TABLE 32.    SUMMARY  OF  IDENTIFIABLE ODOR THRESHOLDS FOR SOURCES EVALUATED
oo

No. of
Source trails
Non-cond. line condensate
White liquor
Weak wash
Green liquor
Black liquor
Turpentine decanter underflow
Air stripper product
Stripper feed
VCE (VRE) vapor condensates
MEE combined condensate
Digester condensate
Decker filtrate after pulp
cleaners
Steam stripper product
1st chlorination stage filtrate
2
2
2
3
2
10
2
6
6
15
15
7
4
7

* Maximum
11000
410
400
630
0.31
>1200
60
350
712
310
>1400
42
0.28
13
ED5Q (xlO 5
Log average
_
-
-
-
-
72
-
38
30
24
24
0.84
0.14
0.012
)
Minimum
1100
195
90
<26
0.010
2.9
<35
2.0
3.3
1.4
0.31
0.0078
0.072
0.00081

* Duplicate runs on same sample
+ Temporary unit, not optimized
indicated as one
trial


-------
the vacuum compression  evaporator condensates,  the digester
condensates, and the  multiple-effect evaporator condensates,  in
descending order with respect to identifiable odor thresholds.

     Process streams  with low dilution ratios to the  threshold
level and being within  a factor of ten of each  other  included the
decker filtrates,  the brown stock washer filtrates, the  steam
stripper product,  and chlorination stage filtrates.

ADDITIVE NATURE OF EFFLUENTS (ODOR BALANCE)

     An effort was made to determine if odor thresholds  of  vari-
ous process streams were additive or could be simulated.  While
it was established that odor thresholds of these effluent samples
could not be predicted  from the reduced sulfur  analysis, the
odor threshold may be additive.  If odor thresholds are  additive,
it would be possible  to predict the impact of a process  change on
total effluent odor threshold.   Seven trials were run including
five designed to simulate bleach plant effluent that  received
digester and multiple-effect evaporator condensates.  Two trials
were run to simulate  a  total unbleached kraft mill effluent.

     The first three  runs,  as shown in Table 33,  were performed
to investigate the effects of various modes  of  sample mixing  on
the identifiable odor threshold.  Those samples designated  "Syn-
thetic I" were made by  adding 15 mis of chlorination  stage  bleach
plant effluent, 1  ml  of multiple-effect evaporator combined con-
densates, and 0.5  ml  of digester condensates,  individually  to the
odor-free water.   This  solution was then sequentially diluted as
necessary.  Those  samples designated as "Synthetic II" were made
by diluting the chlorination stage effluent,  the multiple-effect
evaporator combined condensates, and the digester condensates,
individually to 2000:1  prior to mixing together in the propor-
tions as required  to  become the stock solution.   Those identified
by "Synthetic III" were prepared by adding 30 ml of chlorination
stage effluent to  2 ml  of multiple-effect evaporator  combined
condensate and 1 ml of  digester condensate,  and then  transferring
1 ml of this solution to the first dilution  stage of  2000 ml.
This sample was then  sequentially diluted as necessary.

     In the first  two sets of data all three dilution techniques
Synthetic I, II, and  III, appeared to give similar results.   How-
ever, in the third set  of data the results appeared to be incon-
gruous with the previous information.  There was no identifiable
reason for the discrepancy noted in the third run.

     For the fourth and fifth data sets,  digester and multiple-
effect evaporator  condensate were diluted in odor free water.
The purpose was two-fold, (a)  to investigate the ability to
simulate a complex process stream by adding  only the  compounds
of major impact to the  identifiable odor threshold, and  (b) to
determine the effect  of the chlorine on the  identifiable odor

                                79

-------
                                   TABLE 33.  ADDITIVE  NATURE OP ODOR FROM MILL PROCESS STREAMS
CD
O
Data
set Run
I 1
2
3
4
5
6
7

II 8
9
10
11
12
13
14

III 15
16
17
18
19
20

IV 21
22
23
No. of
Source panel .
MEE comb. cond.
el- stage filtrate
Dig. cond.
Acid sewer .
Synthetic I 2
Synthetic II ,
Synthetic JII
Calculated
MEE comb. cond.
C12 stage filtrate
Dig. cond.
Acid sewer ,
Synthetic I _
Synthetic II ,
Synthetic III
Calculated
MEE comb . cond .
C12 stage filtrate
Dig. cond.
Acid sewer
Synthetic ,
Synthetic III
Calculated
MEE comb. cond.
Dig. cond ^
Simulated acid sewer
9
9
9
9
9
9
9

7
7
7
7
7
7
7

6
6
6
6
6
6

6
6
6
No. of Maximum
panelists test
ident. odor dilution
at all dil. (xlO~5)
0
0
0
2
0
0
0

0
0
0
1
1
0
0

2
1
1
1
2
1

0
2
1
__
-
-
4.2
-
-
-

„
-
-
42
51
-
-

420
4.2
4200
42
51000
84000

_
8400
840
ED,.- Min. dil.
dilution No. of at level
factor panelists missed
(xlO~5) "missed low" (xlO~5)
4.4
0.013
480
0.11
6.7
16
72
1.1
3.4
3.5
3.1
2.5
89
1.4
550
5.6
> 7500
2400
23
310
> 1400
50
0
0
0
0
0
0
5 0.58

_ _
1 0.00029
3 12
0
0
0
0

0
0
0
0
0
1 120

0
0
0
                                                                                                   (continued)

-------
                                                           TABLE  33.   (Continued)
oo


Data
Set
V



VI




No. of Maximum EDso
panelists test dilution
No. of ident. odor dilution factor
Run Source panel. at all dil. (xlO~5) (xlO~5)
24 MEE comb. cond. 7 0 - 43
25 Dig. cond. .7 0 - 200
26 Simulated gcid sewer 7 0 31
Calculated 23
27 Decker water 7 0 18
28 MEE S. cone, hot well 7 0 - 5.4
29 Total mill effluent 7 0 16
30 Simulated THE 7 0 - 2.8
Calculated 7.2

No. of
panelists
"missed low"
1
0
0

1
0
0
1

Min. Dil.
at level
missed

-------
thresholds.  The results indicated that the calculated and di-
luted odor thresholds were the same number.  This showed that the
effects of dilution on condensate streams could be calculated.

     A test was also performed on a simulated total unbleached
kraft mill effluent.  This run was performed by combining screen
room decker filtrate, multiple-effect evaporator and concentrator
combined condensate with odor-free water in the proper propor-
tions to simulate the total mill effluent.  In this instance,
5 ml of odor free water, 1.3 ml of evaporator and concentrator
condensates, and 3.7 ml of screen room decker filtrate were com-
bined.  The results are presented as data set VI and indicated
that the impact of an odorous effluent from the kraft pulping
process on the total mill effluent odor threshold could be
calculated.

     The calculated "identifiable" odor thresholds presented for
each set of data was calculated in the manner shown in the
following example:

     For Data Set I

          One volume of multiple^effect evaporator conden-
     sates at an "IOT" of 4.4x10  and one-half volume of
     digester condensates at an "IOT" of 480x10  in fifteen
     volumes of first chlorination stage bleach plant
     effluent at an "IOT" of 0.13x10

      Ix(4.4xl05) + 0.5x(480xl05) + 15x(0.13xl05)  ,,. , n5
                    (1 + 0.5 + 15)                 isxiu

S UMMARY

     The data generated in this portion of the study indicated
that within the limits of the test procedure, the identifiable
odor threshold associated with effluents from kraft pulping
process streams can be determined and duplicated using head
space analysis for the identifiable odor and the triangle testing
technique.  The work also showed that the effect of one process
stream on the total mill effluent odor level could be simulated
using a weighted average calculation.
                               82

-------
                         REFERENCES
 1.  Chereraisinoff, P.N., Young, R.A., (eds.)/ Industrial Odor
     Technology Assessment, Ann Arbor Science Publishers, Inc.,
     Ann Arbor, Michigan  (1975).

 2.  Rankin V., Harvey Aluminum, 266 F. supp. 169 (D. Ore.)
     (1963).

 3.  Waid, K. R., "Enforcement of Nuisance Regulations for Odor
     Control: Texas Experience," 70th Annual Meeting of APCA,
     Toronto, Ontario (June 1977).

 4.  Feldstein, M., Storey, T.E., Mancott, T.S., "The Control of
     Odor by Emission Limitations," 70th Annual Meeting of APCA,
     Toronto, Ontario (June 1977).

 5.  Dravnieks, A., "Measuring Industrial Odors," Chemical
     Engineering -Deskbook Issue 81 (22) 91-95 (October 21,
     1974).

 6.  Dravnieks, A., "Odorous Air Pollutants - Sampling in the
     Field and Sensory/Analytical Evaluation in the Laboratory,"
     Illinois Institute of Technology Research Institute,
     Chicago, Illinois.

 7.  Duffee, R A., "Appraisal of Odor Measurement Techniques,"
     Battelle Memorial Institute - Columbus Laboratories,
     Columbus, Ohio (1967).

 8.  Cederlof, R., Edfors, M.L., Frieberg, L., Lindvall, T.,  "On
     the Determination of Odor Thresholds in Air Pollution
     Control - An Experimental Field Study of Flue Gases from
     Sulfate Cellulose Plants," JAPCA, Ij6 (2) 92-94  (February
     1966).

 9.  "Evaluation of the Use of Humans in Measuring the Effective-
     ness of Odor Control Technology at the Source," NCASI
     Atmospheric Quality  Improvement Technical Bulletin No. 56
     (September 1971).

10.  Stern, A.C., ed., Air Pollution, Vol 2, 2nd Ed., Academic
     Press, New York, New York.
                               83

-------
11.  Huey, N.A., "Field Operations Activity Abatement Program,"
     61st Annual Meeting of APCA, St. Paul, Minnesota (June 23-
     27, 1968).

12.  "Fundamentals of Odor Perception - Their Applicability to
     Air Pollution Control Programs," NCASI Atmospheric Quality
     Improvement Technical Bulletin No. 54 (June 1971).

13.  Kniebes, D.V., Chisholm, J.A., Stubbs, R.C., "Odors and
     Odorants: The Engineering View," ASHRAE, New York, 27
     (1969).

14.  Berglund, B., Berglund, U., and Lindvall, T., "On the
     Principal of Odor Interaction," Reports from the Psycholog-
     ical Laboratories, University of Stockholm, Sweden 318 (12)
     (February 1971).

15.  Myddleton, W.W., "Cosmetic and Toilet Preparations,"
     Manufacturing Chemist, 35 71-72 (April 1964).

16.  Bartley, S.H., Principles of Perception, Harper, New York,
     New York (1958)

17.  "On Sensory Evaluation of Odorous Air Pollution Intensi-
     ties," NCASI Atmospheric Pollution Technical Bulletin No 50
     (September 1970).

18.  Moskowitz, H.R., "Quantitative Assessment of Odor Intensity
     and Odor Noxiousness Through Psychophysical Sealing," 68th
     Annual Meeting of APCA, Boston, Massachusetts (June 15-20,
     1975).            *>

19.  Friedrich, H.E., Benforado, D.M., "Utilization of the Odor
     Panel Technique for Evaluating Odorous Industrial Emissions,'
     66th Annual Meeting of APCA, Chicago, Illinois (June 24-28,
     1973).

20.  Duffee, R.A., Wahl, J.P., Marrone, W., Nader, J.S., "Defin-
     ing and Measuring Objectionable Odors," International
     Pollution Engineering Conference, Philadelphia, Pennsylvania
     (October 22-25, 1973).

21.  Prince, R.G.H., Ince, J., "Measurement of Intensity of
     Odor," Journal Applied Chemistry, £ 314 (1958).

22.  Benforado, D.M., Rotella, W.J., Horton, D.L., "Development
     of an Odor Panel for Evaluation of Odor Control Equipment,"
     JAPCA, 19 (2) 101-105 (February 1969).

23.  Wilby, F.V., "Variation in Recognition Odor Threshold of a
     Panel," 61st Annual Meeting of APCA, St. Paul, Minnesota
     (June 1968).

                                84

-------
24.   Leonardos, G., Kendall, D.A.,  Barnard, N.J.,  "Odor Threshold
     Determination of 53 Odorant  Chemicals,"  61st  Annual Meeting
     of APCA, St.  Paul, Minnesota (June  23-27,  1968).

25.   Dravnieks, A., "Development  of a  Dynamic Dilution Forced-
     Choice Triangle Olfactometer and  of a Sample  Storage System
     for the Odor  Unit Control Determination  of Rendering Plant
     Exhausts," IIT Research Institute,  Report  No. IITRI-C8217-2,
     Chicago, Illinois  (February  1973).

26.   "Odor Control by Scrubbing in the Rendering Industry Adden-
     dum," Environmental Protection Technology  Series, No. EPA-
     600/2-76-00900, U. S.  Environmental Protection Agency,
     Research Triangle Park, North Carolina  (September 1976).

27.   Dravnieks, A., Prokop, W.H., "Measurement  of  Ambient Odors
     Using Dynamic Forced-Choice  Triangle Olfactometer," 68th
     Annual Meeting of APCA, Boston, Massachusetts (June 15-20,
     1975).

28.   Dravnieks, A., Prokop, W.H., "Source Emission Odor Measure-
     ment by a Forced-Choice Triangle  Olfactometer," 66th Annual
     Meeting of APCA, Chicago, Illinois  (June 24-28, 1973).

29.  "Instructions for Dynamic Triangle  Olfactometer," ITT
     Research Institute, Chicago, Illinois  (February 1974).

30.  Stevens, R.K., O'Keefe, A.E., "Modern Aspects of Air Pollu-
     tion Monitoring," Analytical Chemistry,  £2 (2) 143A-149A
     (February 1970).

31.  Stevens, R.K., "Review of Analytical Methods  "for the Measure-
     ment of Sulfur Compounds in  the Atmosphere,"  National Air
     Pollution Control Administration, Dept.  of Health, Education,
     and Pfelfare  (now part  of EPA).

32.  Natusch, D.F.S., Thorpe, T.M.,  "Element  Selective Detectors
     in Gas Chromatography," Analytical  Chemistry, 45 (14) 1184A-
     1194A  (December 1973).

33.  Stevens, R.K., Hodgeson, J.A.,  "Application of Chemilumines-
     cent Reactions to  the  Measurement of Air Pollutants,"
     Analytical Chemistry,  45_  (4) 443A-449A  (April 1973).

34.  Brody, S.S.,  Chaney, J.E.,  "Flame Photometric Detector,"
     J. of Gas Chroma tography, _4  42-46 (February 1966).

35.  Greer, D.G.,  Bydalek,  T.J.,  "Response Characterization of
     the Melpar Flame Photometric Detector for  Hydrogen Sulfide
     and Sulfur Dioxide," Environ. Sci.  and Tech., 1_  (2) 153-155
     (February 1973).
                                85

-------
36.  Sugiyama, Suzuki, Y.,  Takeucho, T., "Characteristics of S^
     Emission Intensity," J. of Chromatography, 77 309-316
     (1973).

37.  Mizany, A.I.,  "Some Characteristics of the Melpar Flame
     Photometric G.C. Detector in the Sulfur Mode," Journal of
     Chromatographic Science, 8^ 151-154 (March 1970).

38.  Stevens, R.K., Mulik,  J.D., O'Keeffe,  A.E., Krost, K.J.,
     "Gas Chromatography of Reactive Sulfur Gases in Air at the
     Parts-per Billion Level," Analtyical Chemistry, 43 (7) 827-
     831 (June 1971).

39.  Blanchette, A.R., Cooper, A.D., "Determination of Hydrogen
     Sulfide and Methyl Mercaptan in Mouth Air at the Parts-per-
     Billion Level  by Gas Chromatography,"  Analytical Chemistry,
     £8 (4) 729-731  (April  1976).

40.  Farwell, S.O., Bamesberger, W.L.,  Adams, D.F., "Determina-
     tion of PPB/PPT Levels of Sulfur Compounds by Cryogenic
     Collection and GC/WCOT/FPD Methods," PNWIS-APCA Annual Meet-
     ing, Spokane,  Washington (November 14-16, 1977).

41.  Stevens, R.K., O'Keeffe, A.E.,  Ortman, G.C., "Absolute
     Calibration of a Flame Photometric Detector to Volatile
     Sulfur Compounds at Sub-Part-per-Million Levels," Environ-
     mental Science and Technology,  3^ (7) 652-655 (July 1969).

42.  deSouza, T.L.C., Bhatia, S.P.,  "Development of Calibration
     Systems for Measuring  Total Reduced Sulfur and Sulfur Diox-
     ide in Ambient Concentrations in the Parts per Billion
     Range," Analytical Chemistry, £8 (14)  2234-2240 (December
     1976).

43.  Bruner, F., Ciccioli,  P., DiNardo, F., "Further Developments
     in the Determination of Sulfur Compounds in Air by Gas
     Chromatography," Analtyical Chemistry, 48 (1) 141144
     (January 1975).

44.  deSouza, T.L.C., Lane, D.C., Bhatia, S.P., "Analysis  of
     Sulfur-Containing Gases by Gas-Solid Chromatography on a
     Specially Treated Porapak QS Column Packing," Analytical
     Chemistry, 4J (3) 543-545 (March 1975).

45.  Pearson, C.D., Hines,  W.J., "Determination of Hydrogen
     Sulfide, Carbonyl Sulfide,  Carbon Disulfide, and Sulfur
     Dioxide in Gases and Hydrocarbon Streams by Gas Chromato-
     graphy/Flame Photometric Detection," Analytical Chemistry,
     .42 (1) 123-126  (January 1977).
46.
"Analysis of H2S and S02 in the ppb Range," Bulletin 722C,
Supelco, Inc., Bellefonte, Pennsylvania (1975).
                               86

-------
47.  Bremner, J.M., Banwart, W.L. ,  "Identifying Volatile S
     Compounds by Gas Chromatography, "  Sulphur Institute Journal,
     1£ (1)  (1974).

48.  Weurman, C., "Comparison  of  Two  Methods  for  the  Measurement
     of Odors," VDI-Ber,  226  135-9  (1957).

49.  Cormack, D. , Dorling,  T.A.,  "Comparison  of Techniques  for
     Organoleptic Odor-Intensity  Assessment," Chem.  Ind. , 21  857-
     61 (1974).

50.  Dravnieks,  A., O'Donnell, A.,  "Principles and Some  Tech-
     niques  of High Resolution Headspace Analysis,"  J. Agr. Food
     Chem.,  19  (6)  1049-1056  (1971).

51.  Acree,  T.W.,  Butts,  R.M., Nelson,  R.R.,  Lee,  C.Y.,  "Sniffer
     to Determine  the Odor of Gas Chromatographic Effluents,"
     Analtyical  Chemistry,  £8 (12)  1821-1822  (October 1976).

52.  Sarkanen,  K.V.,  Hrutfiord, B.F., Johanson, L.N., Gardner,
     H.S.,  "Kraft  Odor,"  TAPPI, 5_3 (5)  766-783  (May 1970).

53.  Nishida,  K. ,  Honda,  T. ,  "New Apparatus on  the Measurement of
     Odor Intensity," 68th Annual Meeting of  APCA, Boston,
     Massachusetts (June 10-15, 1975).

54.   Polgar, L.G.,  Duffee, R.A.,  Updyke, L.J.,  "Odor Character-
      istics of Mixtures of Sulfur Compounds Emitted From the
     Viscose Process," 68th Annual Meeting of APCA,  Boston,
     Massachusetts (June 10-15, 1975).

 55.   Droege, H.F., "A Study of the Variation of Odor Threshold
      Concentrations Reported in the Literature,"  60th Annual
      Meeting of APCA, Cleveland,  Ohio  (June  11-16, 1967).

 56.   Caron, A.L.,  "Odor Threshold of Field Treated Kraft Pulp
      Mill Condensates," NCASI Internal Memo  (July 30, 197 J).

 57.   "Effect of Pulp and Paper Mill Effluents on the Taste  and
      Odour of Water and Fish," Project Report 12-1 to Domtar  Fine
      Papers, Ltd., CCWPAR  Secretariat, Canadian Forestry Service,
      Ottowa, Canada  (Progress to March 31, 1971).

 58.   Thibodeaux, L.J., Parker, D.G., "Desorption Limits of
      Selected Industrial Gases and Liquids from Aerated Basins,
      76th National Meeting of AIChE, Tulsa,  Oklahoma  (March  11,
      1974).
      trial Wastewater,"  172nd Meeting  of ACS  San Francisco
      California  (August  29-September 3, 197b).
                                 87

-------
60.   Standard Methods for the Examination of Water and Wastewater,
     (14th ed.),  APHS,  AWWA,  WPCF (1975).

61.   Fischer, R.A., Yates, P., Statistical Tables, 50-52, Hafner
     Publishing Co., New York, New York (1948).
                             /
62.   "Characterization of Various Condensate Streams in the
     Kraft Process," NCASI Stream Improvement Technical Bulletin
     No. 310 (April 1978).
                               88

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 REPORT NO.
 EPA-600/2-79-117
                        3. RECIPIENT'S ACCESSION NO.
 TITLE AND SUBTITLE
 Method for  Determining Potential
 Selected Kraft  Process Streams
Odor Contribution of
5. REPORT DATE
  June  1979
                                     issuing  date
                        6. PERFORMING ORGANIZATION CODE
 AUTHORIS)
 Michael E.  Franklin
 Andre L.  Caron
                        8. PERFORMING ORGANIZATION REPORT NO.
 PERFORMING ORGANIZATION NAME AND ADDRESS
 National Council of the  Paper Industry
                        10. PROGRAM ELEMENT NO.
  for Air  and Stream Improvement,
 260 Madison Avenue
 New York, NY 10016
Inc.
                                                              1BB610
                        11. CONTRACT/GRANT NO.
                          R-804646-01
12. SPONSORING AGENCY NAME AND ADDRESS
  Industrial  Environmental  Research Laboratory
  Office  of Research and  Development
  U.S.  Environmental Protection Aqency
  Cincinnati, Ohio  45268
                        13. TYPE OF REPORT AND PERIOD COVERED
                          Final ;7/20/76 - 6/19/78
                        14. SPONSORING AGENCY CODE
                          EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
          The  objective of this  project was to define the  potential  odor contribution
  of selected  process streams  in the kraft industry that are  routinely sewered.  A
  procedure was  suggested that can  be used for this purpose.
          Use  of a  dynamic olfactometer
  determined by  complete volatilization
  unsuccessful.   No correlation  between
  tion in the  qas stream as measured  by
  obtained.
      and odor  panels  to measure odor thresholds
      of the  sample  or stripping of the sample were
      odor threshold and reduced sulfur concentra-
      qas chromatoaraphic techniques could be
          Odor  panels were employed  usina the head space analysis  and the forced-choice
  trianqle technique.  It was shown  that identifiable odor  threshold values were more
  reproducible  and judged more meaningful than simple odor  threshold values.

          It was  shown that independent of mixing techniques,  odor intensities of kraft
  mill process  streams were additive.   This was demonstrated  for an acid sewer contain-
  ing first chlorination stage effluent, digester condensates,  and multiple-effect
  evaporator condensates; and for a  total mill effluent with  multiple-effect evaoorator
  condensates,  decker water, and odor-free dilution water used  as  make-up.  	
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
a.
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                                                                         c.  COSATI Field/Group
  Odors
  Odor control
  Pulping
  Sulfate pulping
                                        13B
18. DISTRIBUTION STATEMENT

  RELEASE TO PUBLIC
           19 SECURITY CLASS (ThisReport)
            ' UNCLASSIFIED
               I.NQ^F,
                                              20. SECURITY CLASS (Thispage)
                                                 UNCLASSIFIED
                                                                         22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE
                                                                     * U.S. GOVERNMENT PRINTING OFFICE 1979-657-060/5344
                                             89

-------