MIDWEST RESEARCH INSTITUTE
            OCCUPATIONAL  EXPOSURES  FROM  BAGGING AND  DRUMMING  OPERATIONS
                                    FINAL  REPORT

                         EPA  Prime  Contract No. 68-02-4252
                              Work  Assignment  No.  10
                            MRI  Project No. 8810-A(01)

                                 September 9,  1987
                                        For

                       U.S. Environmental Protection Agency
                            Office of Toxic Substances
                           Field Studies Branch, TS-798
                              Washington, D.C.  20460

             Attn:  Mr. Michael R. Kalinoski, Work Assignment Manager
                    Dr. Kin F. Wong, Work Assignment Manager
                    Dr. Joseph J. Breen, Project Officer
                    Mr. William M. Burch, Project Officer

MIDWEST RESEARCH INSTITUTE


-------
            OCCUPATIONAL EXPOSURES FROM BAGGING AND DRUMMING OPERATIONS
                                   FINAL REPORT

                         EPA Prime Contract No.  68-02-4252
                              Work Assignment No.  10
                            MRI Project No.  8810-A(01)

                                 September 9, 1987
            Attn:
                    For

   U.S. Environmental Protection Agency
        Office of Toxic Substances
       Field Studies Branch, TS-798
          Washington, D.C.  20460

Mr. Michael R. Kalinoski, Work Assignment Manager
Dr. Kin F. Wong, Work Assignment Manager
Dr. Joseph J. Breen, Project Officer
Mr. William M. Burch, Project Officer
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY, MISSOURI 64110 • 816 753-7600

-------
          This document has  been  reviewed and approved for publication by
the Office of Toxic Substances, Office; of-'Pesticides and Toxic Substances,
U.S. Environmental Protection fA"gency? -; The^use^df-^lkraae names or commercial
products does not constitu.tSe^Agencyfendbrsemerit.or -recommendation for use.

-------
                                   PREFACE
          This report describes the study performed by Midwest Research Insti-
tute (MRI) under Work Assignment 10 of Environmental Protection Agency (EPA)
Contract No. 68-02-3938.  The objective of this work was the development and
validation of predictive models to be used in estimating the inhalation expo-
sure of workers involved in the bagging or drum filling of solids or in the
drumming of liquids.  The EPA work assignment managers were Dr. Kin Wong and
Michael Kalinoski.

          The task leader for Midwest Research Institute was Dr.  Chatten
Cowherd, Director, Environmental Systems Department.  The authors of this
report were Dr. Chatten Cowherd, Mary Ann Grelinger, and Dr. Gregory Muleski.
Mrs. Grelinger coordinated the laboratory studies and assisted in the plant
testing and experimental data analysis.   Dr.  Muleski was responsible for the
exposure model development.   Mr. Fred Bergman designed the low-flow wind tun-
nel and coordinated the plant testing and the chemical analysis of field and
laboratory samples.  Mr. Steve Cummins assisted in the collection of field
and laboratory samples and in data analysis.   Other MRI staff members who con-
tributed to this study were Fred Hopkins, Bonnie Carson, Mark Golembiewski,
Dr. George Scheil, David Steele, Frank Pendleton, and Dr.  Robert Hegarty.
                                        MIDWEST RESEARCH INSTITUTE
                                        Chatten Cowherd, Director
                                        Environmental Systems Department
Approved:
r"
     o4^ c&kZL*^ — if .
Jack Balsinger
Quality Control Coordinator
*aul C.  Constant
Program Manager

-------
                              TABLE OF CONTENTS
 I.         Summary ..........................   1

 II.        Introduction ........................   5

                A.    Background and Purpose ..............   5
                B.    Industry Characterization ............   5
                C.    Prior Modeling Approaches ............   6
                D.    Study Design ...................   7
                E.    Quality Assurance  ................   8

 III.       Conclusions  ........................   g

                A.    Plant Studies  ..................   9
                B.    Laboratory Drumming  Study ............   9
                C.    Model  Development  ................  10

 IV.        Plant Scale  Studies  ....................  11

                A.    Drumming Operation (Plant A) ...........  11
                B.    Bagging Operation  (Plant  B)  ...........  22

 V.         Laboratory Drumming  Study  .................  51

                A.    Emission Tests ..................  51
                B.    Dispersion  Tests .................  63
                C.    Exposure Tests ..................  72

 VI.       Model Development .....................   91

                A.    Modeling Alternatives  ..............   91
                B.    Modeling Approach  ................   93
                C.    Model Validation .................   97

 References ..............................  101

Appendix A - Industry Survey
Appendix B - Reports on Additional Plant Surveys
Appendix C - Laboratory Pilot Study
Appendix D - Quality Control Data
Appendix E - Data from Laboratory Drumming Tests

-------
                               LIST OF FIGURES
Number                             Title                              Page

IV-1      Plant A drumming line	12
IV-2      Sampling element for collection of hydraulic fluid mist
            and vapors	14
IV-3      Sampler deployment around the Plant A drumming operation.  .   15
IV-4      Overhead view of airflow at 1.5 m above drumming platform
            for Test A-l (Fluid 1)	20
IV-5      Overhead view of airflow at 1.5 m above drumming platform
            for Test A-2 (Fluid 2)	21
IV-6      Vapor concentrations (ug/m3) of Fluid 2 observed during
            Test A-2	23
IV-7      Mist concentrations (ug/m3) of Fluid 2 observed during
            Test A-2	24
IV-8      Plant B bagging facility	27
IV-9      Sampling elements for collection of sodium tripoly-
            phosphate particles 	   29
IV-10     Sampler deployment around the Plant B bagging operation .  .   30
IV-11     Overhead view of airflow at operator breathing height
            during Test B-l	35
IV-12     Overhead view of airflow at operator breathing height
            during Test B-2	36
IV-13     Overhead view of airflow at operator breathing height
            during Test B-3	37
IV-14a    Total suspended particulate concentrations of STP (ug/m3)
            for Test B-l	41
IV-14b    Respirable particulate concentrations of STP (ug/m3)
            for Test B-l	42
IV-15a    Total suspended particulate concentrations of STP (ug/m3)
            for Test B-2	43
IV-15b    Respirable particulate concentrations of STP (ug/m3)
            for Test B-2	44
IV-16a    Total suspended particulate concentrations of STP (ug/m3)
            for Test B-3	45
IV-16b    Respirable particulate concentrations of STP (ug/m3)
            for Test B-3	46
IV-17     Airborne particle size distributions obtained using
            cyclone/cascade impactor	47
V-l       Overhead view of pull-through flow tunnel 	   53
V-2       Full scale 55-gal drumming operation in flow tunnel ....   54
V-3       Comparison of emissions from top versus bottom drum
            filling of perc at nominal 25 gpm	60
V-4       Profile of methanol concentrations at 3 m sampling plane
            (200 fpm)	65
V-5       Profile of wind speeds at 3 m sampling plane for methanol
            test at 200 fpm	66
V-6       Comparison of vertical profiles of methanol  concentrations
            from a drum side release with wind speed of 200 fpm ...   67
V-7       Comparison of vertical profiles of methanol  concentrations
            from a drum side release with wind speed of 100 fpm ...   68
                                     vn

-------
                          LIST OF  FIGURES  (continued)


Number                             Title                               Page

V-8       Cross-sectional view of methanol concentrations (ppm)
            at 1 m downwind for a drum side release at 100 fpm,
            compared to downwind  facing worker exposure (ppm)  ....  69
V-9       Cross-sectional view of methanol concentrations (ppm)
            at 3 m downwind for a drum side release at 100 fpm,
            compared to downwind  facing worker exposure (ppm)  ....  70
V-10      Sampling locations for  upwind bung position 	  73
V-ll      Sampling locations for  side bung position	74
V-12      Example exposure testing	75
V-13      Effect of breathing height and  upwind distance on exposure
            to perc concentrations (ppm)	80
V-14      Effect of breathing height and  upwind distance on exposure
            to perc drumming emissions	81
V-15      Effect of wind speed and upwind distance on exposure to
            methanol drumming emissions at 1.5 m breathing height .  .  82
V-16      K contours for methanol (M < 100), 100 fpm air velocity
            and downwind worker orientation 	  83
V-17      K contours for methanol (M < 100), 200 fpm air velocity
            and downwind worker orientation 	  84
V-18      K contours for methanol (M < 100), 100 fpm air velocity
            and non-downwind worker orientation 	  85
V-19      K contours for methanol (M < 100), 200 fpm air velocity
            and non-downwind worker orientation 	  86
V-20      K contours for methanol (M > 100), 100 fpm air velocity
            and downwind worker orientation 	  87
V-21      K contours for methanol (M > 100), 200 fpm air velocity
            and downwind worker orientation 	  88
V-22      K contours for methanol (M > 100), 100 fpm air velocity
            and non-downwind worker orientation 	  89
V-22      K contours for methanol (M > 100), 200 fpm air velocity
            and non-downwind worker orientation 	  90
VI-1      The box (uniform mixing) model	92
                                     vm

-------
                               LIST OF TABLES
Number                               Title                            Page

IV-1      Plant Tests of Drumming Operation 	   18
IV-2      Representative Drum Fill Cycles	18
IV-3      Plant A Ventilation	19
IV-4      Mass Concentrations Observed During Test A-2 (185-min
            test period)	25
IV-5      Mass Concentrations Observed During Test A-2 (98-min
            test period)	26
IV-6      Sodium Tripolyphosphate Analysis Protocol 	   32
IV-7      Plant Tests of Bagging Operation	33
IV-8      Plant B Ventilation	34
IV-9      Mass Concentrations Observed During Test B-l	38
IV-10     Mass Concentrations Observed During Test B-2	39
IV-11     Mass Concentrations Observed During Test B-3	40
IV-12     Bulk STP Particle Size Distribution	48
V-l       Chemicals Used in Drumming Simulations	52
V-2       Emissions Test Data - Perchloroethylene	57
V-3       Emissions Test Data - Ethylene Glycol	58
V-4       Rates of Attainment of Maximum Drum Exit Concentration.  .  .   61
V-5       Drumming Emission Factors 	   62
V-5a      Comparison of Saturation Factors	63
V-6       Dispersion Test Log	71
V-7       Exposure Test Leg	76
V-8       Drum Emissions for Tests of Dispersion and Exposure ....   76
V-9       Special Exposure Studies	78
V-10      Effect of Worker Movement on Exposure  	   79
VI-1      Default Values for Equation 1	93
VI-2      Default K Values	95
                                      ix

-------
 I.  SUMMARY
          Under the Toxic  Substances  Control  Act,  the potential  exposure to
workers involved with  the  manufacture and processing of a new chemical  must
be estimated.  A  significant  portion  of worker exposure is thought to occur
during packaging operations,  in  particular,  drumming and bagging of liquids
and solids.  Currently, crude techniques are used to predict worker exposure
to emissions from  the  drumming  of liquids and  the  bagging or drumming of
solids.  Although  mathematical models have been developed previously to im-
prove  these  estimates, available data are  generally inadequate to verify the
models.

          The purpose  of this work assignment was to  develop  and validate
predictive models used to determine the inhalation exposure levels of workers
involved in  the  drumming  of liquids  and the bagging or drumming of solids.
The study consisted of four tasks.   An industry survey was conducted to char-
acterize industrial drumming  and bagging operations and to select chemicals
and facilities for study.   An initial laboratory pilot study was performed
to test emission  models,  dispersion theories, sampling strategies, and ana-
lytical methods.    A plant-scale  occupational  exposure study was  directed to
the collection of  specific data  on worker breathing zone concentrations re-
sulting from actual bag and drum filling operations.   Finally, a  second labo-
ratory pilot study was performed using a  drum  filling station to develop a
predictive model of inhalation  exposure, which was tested against data from
the plant-scale occupational exposure study.

          In the first laboratory pilot study, a walk-in laboratory flow tun-
nel was used to characterize the low-flow (i 2 m/s) dispersion of methane gas
and ammonium chloride  particles  at distances of 2 to  6 m  from the point of
release.  Several  release  configurations were  included:  single- and  double-
point  releases into an unobstructed  flow; point release near a wall; point
release past a side-draft hood;  and bung release from a 55-gal drum (with and
without a  close  fitting hood).   A  two-dimensional array of  samplers was used
to characterize dispersion patterns at various distances downstream.   The ob-
served dispersion  patterns  were  compared to existing dispersion theories and
then used  as the basis for a modified dispersion algorithm.  The predictive
capability of  the  resulting dispersion model was found to  be significantly
better than  the  other models tested  against the laboratory data.   The first
laboratory pilot study is presented in Appendix C of this report.

          The plant-scale  occupational  exposure study was performed on full
scale  packaging systems at two industrial plants.  At Plant A, low-volatility
hydraulic  fluids were  packaged in  55-gal drums;  and at Plant B, granular so-
dium tripolyphosphate  (STP) was packaged in 50- and 100-Ib bags.   Integrating
samplers were  deployed at  fixed  points  surrounding each work  station.   Pairs

-------
of samplers were colocated to determine measurement precision.  Personal sam-
plers were operated simultaneously by plant personnel.  Additional  data col-
lected during  each test  included  ventilation flow  speeds  and directions
across the work  station  and characterization of worker activity  (time and
motion).   Although hydraulic fluid vapors were found  in low concentration  at
the  drumming  station  (^ 100 |jg/m3), comparable  amounts  of mist were  also
present.

          The concentrations of particulate STP  in the vicinity of  the drum-
ming were much  higher (•>» 1,000 ug/m3)  with about 20% in the respirable size
range (i 10 umA).  Moreover, respirable  STP concentrations measured by the
personal  sampler  were considerably higher than  the  respirable particulate
concentrations measured at nearby sampling points.   This is believed to result
from the airflow which was  directed toward the back of the bagging  operator,
and which produced a  relatively stagnant air region immediately in front of
the operator.   Gently circulating air in this region is likely to have trans-
ported emissions from the  bagging  operation directly to the breathing zone.

          Subsequently a second laboratory pilot study was conducted using a
drumming station and  encompassed the measurement of  emissions, dispersion,
and operator exposure resulting  from a full-scale 55-gal  drumming operation
in the low-flow laboratory wind tunnel.   The second laboratory pilot study is
described in Section  V.  A  55-gal  drum with an added bottom drain was posi-
tioned on a platform  within the flow tunnel.  A 300-gal  reservoir was in-
stalled outside the tunnel  at a height of 2.13 m (7 ft) above the floor.  The
overhead tank was  fitted with chemically resistant  hosing  and a pump  for
transfer of the test  liquid to the 55-gal drum at a fixed rate controlled  by
a ball valve  at the  hose nozzle.   After each drum filling, the contents of
the drum were pumped  back to the overhead tank.  The  test liquids (methanol,
perchloroethylene, and ethylene glycol)  spanned  a vapor pressure range from
about 0.1 to 125 mm Hg at 25°C.

          Drumming emissions were measured by sampling and continuously ana-
lyzing vapor effluent from  the bung opening as  the  drum  was being filled.
Drumming emissions were  characterized for top versus  bottom filling and as a
function of drum fill  rate.   Top fill emissions exceeded bottom fill emissions
by a  factor of  two or more.  Because of the tendency  of perchloroethylene  to
froth, top fill  emissions  were found to exceed  the equivalent of saturated
vapor.

          Distributions of  airborne  concentration  and personal exposure for
fixed worker positions were measured in the vicinity of the drumming opera-
tion.  The presence of the  worker enhanced the exposure by drawing  the plume
upward on the lee side of the worker.  The highest exposure occurred with the
source in front of a  worker and the ventilation  airflow approaching from be-
hind the worker.

          A semi empirical  model  of drumming exposures was  developed using
(a) contours of  breathing height concentrations developed  from  the second
laboratory pilot study and (b)  worker time/motion data from plant-scale oper-
ations.   The model separates worker exposure into two components—background

-------
and plume contribution.  The background value is calculated by assuming com-
plete mixing of the plume in the plant ventilation air.  The plume contribu-
tion depends  on  emission rate  and worker positioning in  relation  to the
source and the ventilation  flow direction.   The validity  of  this modeling
approach was checked against the test data from the pi ant-scale drumming oper-
ation, with favorable results.

-------
II.  INTRODUCTION
     A.    Background and Purpose

          Under the Toxic Substances Control  Act  (TSCA),  a manufacturer or
processor must submit to  the  EPA a premanufacturing notice (PMN) for every
new chemical prior  to commercial  production.   The  Agency  has 90  days to  re-
view the PMN and to take any regulatory actions deemed  necessary.  As part  of
this review, the potential workplace exposure  associated with the manufacture
and processing of the new chemical  must be  estimated.

          Monitoring data on similar substances, structural  analogs,  or mathe-
matical  models compiled  from  experience  with  the  specific industrial opera-
tions are often utilized to  evaluate the level  of worker exposure.  The common
models generally are based on material  balances and diffusion transport con-
siderations.  However,  data  from the open literature are frequently not avail-
able to verify the accuracy  and precision of these  models.

          Currently under the PMN review process,  a simple mass balance model
is used to estimate worker exposure to  vaporous emissions  (Berman 1982, USEPA
1984).  The model is based on complete  mixing  corrected by a mixing factor  to
account for  incomplete  mixing.   For the packaging of solids,  no dispersion
model is  currently  used.   Instead the  OSHA permissible exposure limit for
nuisance dust (15 mg/m3  time-weighted average,  or  TWA)  is  used as a conserva-
tive estimate in the assessment.

          The purpose of this work assignment  is to develop and validate  pre-
dictive models  used to  determine the inhalation exposure levels of workers
involved  in  the  drumming of liquids and the bag or drum filling of solids.
Such models  must address  both the  release and the dispersion of emissions in
the immediate vicinity of the packaging equipment.

     B.    Industry Characterization

          The bagging and drumming  of  PMN chemicals takes place under a wide
range of  conditions that  influence  worker exposure during packaging.*  These
variables  include  degree of automation; employee  work practices; level  of
occupational  health awareness  (managers and  employees);  filling equipment
design, age,  and frequency of  maintenance; container  and closure  design;
plant layout  and ventilation;  effectiveness of emission  control  devices, if
present;  employee  use of personal  protective  gear;  unit  value  of  product
(lower product  loss rates,  including packaging emissions, will be permitted
*Bomberger et al.  (1983)  provides  a good, up-to-date source of information
describing typical  filling  technology  and available literature.   The reader
is referred to that for additional  references.

-------
if the product value is high); and product flammability.  Some volatile  liq-
uids and dust-prone solids are carefully controlled because of the fire  haz-
ard they represent in the absence of such controls.   In an effort to assemble
a broadly applicable data base and methodology for estimating worker exposure
in the bagging and drumming of PMN chemicals, only the most relevant of these
variables could be studied.

          1.  Drumming of Liquids

          In the  drumming  of PMN liquids, it  is  expected that both fully
automated and semiautomated  filling  practices would be commonly used.   The
product will be  either  top  filled (dispenser stays near the top of the bar-
rel, liquid splashes  freely  during filling)  or bottom filled (dispenser re-
mains submerged  inside the barrel, as much as possible, to minimize volatile
emissions).   Emission controls may or may not be present.   An operator may be
required to manually screw a cap on each barrel or wipe droplets off the tops
of the barrels (Bomberger et al.  1983).

          In addition to work practices and  controls,  exposure to a specific
chemical is dependent on  its inherent properties.  Vapor pressure is  a  key
physical/chemical parameter  for  any  analysis of emissions from the drumming
of liquids.   PMN chemicals have been found to exhibit a considerable range of
vapor pressures common to industrial chemicals in general.

          2.  Bagging of Solids

          In the bagging of PMN solids it is unlikely that production volumes
will  justify  fully automated packaging lines  in  many instances (Bomberger
et al. 1983).  However,  the  use of automated or semiautomated equipment re-
mains a  key variable  in predicting worker exposure.  The product will  typi-
cally be placed  into  bags (open top or valve)  by gravity feed, auger  feed,
or vibrator  feed equipment.   Emission controls may or may  not be present.
Frequently, an operator may  be  required to manually staple or  otherwise  seal
the bag tops.

     C.   Prior Modeling Approaches

          The models  available  for modeling  contaminant concentrations in an
industrial  environment may  be grouped into three types based on their theo-
retical  basis:   mass balance;  Gaussian  diffusion;  and aerosol mechanics.

          Mass balance models  are based on mathematical  accounting schemes
whereby all fractions of the material passing through the process under study
are quantified,  and  the total outputs are always equal to the total inputs.
These  approaches are  most applicable for gaseous  emissions  because vapors
follow air  currents  freely  and are not  influenced  by gravity.  Particulate
emissions consisting  of  fine particles—smaller than  10 to 20  |jm  in diameter
according to  Clement  Associates (1981)  and generally  smaller  than  5 urn in
diameter according to Bomberger et al.  (1983)—may also  be  modeled effec-
tively using mass balance equations.

          Gaussian dispersion models  are  not commonly employed in an  indoor
setting.  However, Schroy (1979) has discussed the use of this type of model.

-------
Such models are  based  on the Gaussian diffusion  of  particles as they move
away from a source,  and they predict the maximum downwind concentration of
the contaminant as a function of distance from the source.

          The  third model type, based on aerosol mechanics,  is exemplified by
Davis (1971).   Here, the modeling approach is  based  on  the  properties and
mechanics of airborne  particles,  such as particle generation, coagulation,
and settling.   Because  the  use of this method involves solutions of compli-
cated equations,  use of a computer is recommended.

          The  discussion  section  at  the  conclusion of  the Clement Associates
(1981) report  notes that mass balance equations are both the simplest and most
commonly used approach  for estimating exposure  to a substance  in an enclosed
space.   However, there  is no widely accepted methodology for this  type of
analysis.

          Berman (1982)  provides  methodology  for estimating workplace expo-
sure to PMN substances.  His approach is based on a collection of typical and
worst case approximations combined with easily understood equations and nomo-
graphs.   For bagging operations,  a typical concentration of 2  mg/m3 of mate-
rial (30% of which is  respirable) is assumed in the air surrounding the bag
packer and, if warranted, a worst case value of 30 mg/m3 (also 30% respirable)
is assumed for especially dusty conditions.   For drum filling of liquids, a
plug of saturated air equivalent to the barrel volume is assumed to be gener-
ated for every container filled.  The concentration and total  volume  of  con-
taminated air so released is  calculated  using an  equation which accounts  for
vapor pressure, temperature, and gram-molecular weight of the liquid, as well
as the  filling rate and filling mode (top  filling  versus  bottom filling).
Evaporation from spillage is ignored.  A typical ventilation rate for bagging
and drumming areas is assumed to be 85 mVmin (3,000 ft3/min) and a worst case
value is 14.2 mVmin (500 ft3/min).   For outdoor operations with only minimal
structural protection,  the average wind  velocity  of North America, 14.5  km/h
(9 mph), is assumed in place of a building ventilation rate.

          The validity  of this  methodology remains unknown, because  insuffi-
cient data are available for verification.   However,  the Berman report repre-
sents the most complete set  of  guidelines presently available  for estimating
workplace exposure  to  PMN chemicals during bagging and drumming activities.

     D.    Study Design

          The study  design  was comprised of four tasks.   An industry survey
(Task A) was conducted to characterize industrial drumming and bagging opera-
tions and to select chemicals and facilities for study.  An initial  series of
laboratory pilot studies (Task B) were performed to test emission models, dis-
persion  theories,  sampling  strategies,  and analytical methods.  The  plant-
scale occupational  exposure  study and the laboratory  drumming  study  (Task  C)
were directed to the collection of specific data on worker breathing zone con-
centrations resulting  from actual bag and drum  filling operations.   Finally,
the results of the laboratory drumming studies were used to develop a predic-
tive model of inhalation exposure, which was tested for validity against data
from the plant-scale occupational exposure study (Task D).

-------
     E.   Quality Assurance

          Consistent with  the Quality Assurance Program Plan,  specific  qual-
ity control checks were used throughout the experimental  phases of this study.
These are reported separately in Appendix D.

-------
  I.  CONCLUSIONS
          The conclusions derived from the plant scale bagging and drumming
studies and from the  laboratory drumming study are presented separately below.

     A.   Plant Studies

          1.   Drumming of Low  Volatility Hydraulic Fluids

          •    Low but detectable  concentration  levels (^ 100 ug/m3) were
               observed  only  when  the side draft  hood was not operated.

          •    Mist,  rather than  vapor concentrations, predominated at nearly
               all  sampling locations.

          •    Actual  drumming operator exposure was  limited  by the small
               percentage of time spent at  the drum filling apparatus.

          2.   Bagging of  Granular Sodium Tripolyphosphate (STP)

          •    Airborne  suspended  particulate concentrations  of STP were
               fairly uniform  with 20% or  less in the  respirable size range
               (< 10  urn aerodynamic diameter).

          •    Silt fraction  of  bulk STP  (i.e.,  portion of dry material
               passing a  200-mesh screen) was found to  be good determinant of
               relative emission  rates of STP.

          •    Enhanced  operator exposure  resulted from  stationary operator
               position  at  the  bagging  apparatus  and from ventilation air
               backflow.

     B.   Laboratory  Drumming  Study

          1.   Emissions

          •    Emissions  increase with vapor pressure of the drummed liquid.

          •    Emissions  of the  drummed  chemical (vapor plus mist) may exceed
               the value  calculated as saturated vapor.

          •    For the chemicals tested, topfill  emissions were found to ex-
               ceed bottomfill emissions  by a factor  of 2 or more.

-------
     2.   Dispersion

     •    Plumes of vapor/mist were  found  to sink because of:   (a)  en-
          trainment into drum wake with downward momentum; and (b)  higher
          density caused by higher molecular weight or lower temperature
          in relation to ambient  air.

     •    The bung position has little effect on downwind dispersion of
          plume.

     3.   Worker Exposure

     •    Highest worker  exposure was found  to  occur with source in
          front of  the worker and  ventilation airflow  from behind.

     •    Cross ventilation of drumming operation produced significantly
          lower exposures than rear ventilation.

C.    Model Development

     •    Previously developed models are  inadequate because either:
          (a) they do  not account for spatial variability of exposure;
          or (b) they  require such detailed physical/chemical data and
          plant-specific  input data  as  to  be impractical for assessing
          exposure to new chemicals.

     •    The proposed  model  accounts for  both spacial variability and
          the previously  uncharacterized  effect of worker  orientation
          while requiring limited physical  data.

     •    The proposed model was  shown to successfully predict the plume
          contribution  to worker  exposure  at the plant-scale drumming
          facility, using an  emission factor  developed from the labora-
          tory  study of drumming  emissions  in the  low-flow  wind tunnel.
                                 10

-------
 IV.  PLANT SCALE STUDIES
          The plant-scale occupational  exposure  study  was performed on full
scale packaging  systems at two plants.  At Plant A, low-volatility hydraulic
fluids were  packaged  in  55-gal  drums; and at Plant B, granular sodium tri-
polyphosphate (STP) was packaged in 50- and 100-lb bags.   Integrating samplers
were deployed at fixed points surrounding each  work station.   Pairs of sam-
plers were collocated  to provide for  determination of measurement precision.
Personal samplers were operated simultaneously by plant personnel.   Additional
data collected during  each  test  included ventilation  flow speeds/directions
across the work  station,  and characterization of worker activity (time and
motion).  The studies performed at Plant A and Plant B  are described in detail
in the sections immediately following.

     A.    Drumming Operation (Plant A)

          1.   Introduction

          The first plant-scale study was directed to the collection of spe-
cific data on worker breathing zone concentrations resulting  from  the drumming
of liquids.  These  data will  be  used to develop  and  refine  models  for the
assessment of inhalation exposure  for workers  involved  in  such  packaging oper-
ations.

          The testing was conducted at a plant facility where low volatility
hydraulic fluids  are packaged in 55-gal  drums  by  means  of  an  automated bottom-
fill  system.   This  facility  was judged to be representative of full-scale
drumming operations and associated  emission controls.   The testing took place
in April 1985.

          The layout of the central drumming  facility  is shown  in  Figure IV-1.
Typically 160 to  180 drums (55 gal) are filled in a  period of 4 to 5  h during
the day  shift.   The drumming operator spends most of the time inspecting and
labeling drums prior to filling.  When the  drum  filling apparatus  is  activated
by the operator,  the filling  lance  is  automatically  inserted  and lowered into
the drum.  As the  drum is filled,  the lance  is automatically retracted such
that the discharge end of the lance remains under the liquid level.   The ac-
tual  drum fill  time is  approximately 1 min.  After the  drum is  filled,  a bung
cap and  cover are installed by the worker using  self-powered crimping tools.

          The manual filling station for 5-gal drums in the central  drumming
facility is operated on an irregular schedule.  The  operator  spends about 50%
of his time  filling  the drum.   The remainder of the time  is spent getting
empty drums and placing the filled drums on a pallet.   Filling  of  each 5-gal
drum requires approximately 20 s.

          The ventilation  pattern  in the drumming area  is  influenced by wall
fans  which blow air across  the packaging line  from the  rear.  When operating,


                                      11

-------
                                                                         N
                             Manual
                             Fill
0     1      2
      m
Heater
                                                           OQCDO
                                             Drumming Platform
                                           Overhead •
                                             Fans
                                                              Wall
                Figure IV-1.   Plant A drumming  line.

-------
 the overhead heater fan  blows  air laterally in the direction shown on Fig-
 ure IV-1.   In addition, both drum filling stations are equipped with a local
 side draft hood, which pulls  air from around the  bung opening,  thereby cap-
 turing vapors discharged  by each drum as  it  is  filled.

           The hydraulic fluids  drummed at this  facility span  a  range  of vola-
 tility (vapor pressures at  22°C  from 5 mm Hg to < 0.01 mm  Hg).   In order  to
 provide for greater analytical  sensitivity,  the decision was  made to  test  the
 two most volatile  liquids  (referred  to as Fluid 1 and Fluid  2)  having  vapor
 pressures  at 25°C of 5.4  mm Hg  and <  2 mm Hg,  respectively.

           Preliminary  sampling  of organic vapors generated  by the drumming of
 Fluid  2 was conducted  by MRI on  February 7, 1985.  Integrated  samples  were
 collected  on glass fiber filters  followed by silica gel  tubes  (as described
 below)  over a period  of  281 min,  during  which  55  drums  (55-gal  capacity)  of
 Fluid  2 and 162 drums   (5-gal capacity) of Fluid 1 were  filled.   Analysis  of
 the four samples by gas  chromatography yielded total  mist  and vapor  concen-
 trations  of  0.060  to  0.090 mg/m3.   The three  samples  collected very near
 (<  1 m) the point of vapor release from the  55-gal drums produced nearly the
 same concentration as  the "background" sample collected 4 m away.  This uni-
 formity of concentrations was  judged to be the result  of secondary emissions
 associated  with minor  spillage in  the  general  area and/or the  operation of
 the 5-gal  drumming  station  nearby.

           2.  Experimental  Methods

          This  section  describes  the  sampling and  analysis  procedures used  in
 the primary testing of the  55-gal  drumming  station during April  of  1985.

                a.   Sampling  Procedure

                A  network  of 20 samplers was  deployed at  fixed points in the
 vicinity of the 55-gal drumming operations.   As shown in Figure IV-2,  each
 sampler consisted of a  filter holder with  a  20-mm  glass fiber filter followed
 by  a tube  containing activated 42/60  mesh silica  gel.   The tube  contained
 50  mg  of  silica gel in both a front and  back section.   Each silica gel  tube
 consisted of  a  front or main section and  a backup  section.  The  front and back
 sections were separated by an inert divider.   The  two fractions were recovered
 separately.   A  test to demonstrate the retention of the analyte by the silica
 gel  tube  is described  in  Appendix D.   Analysis of randomly selected  backup
 portions of two tubes  for each  analyte indicated levels  beneath  the limit  of
 detection.  The absence of detectable amounts of analyte  in the back half was
 taken  as an indication that breakthrough  had not  occurred.   Air was drawn
 through each  sampler by a suction manifold connected to a vacuum pump.   The
 use  of  25-ga. x 5/8-in. hypodermic  needles as critical  orifices  in each  line
 ensured a standard flow rate of 0.30 L/min (±5%) through each sampler.

               Most of  the samplers were positioned at  breathing height  (1.5 m
above  the  work  station platform, along the  front  and  rear boundaries,  as
shown  in Figure IV-3.   Four sets of samplers were colocated  to  provide for
determination of measurement precision.  Samplers  3 through 8 were positioned
to  profile  for  location of  the plume  centerline downwind of  the point of
vapor release from drum filling.   An operator personal  sampler and additional


                                      13

-------
                                            Critical
                                            Orifice
Figure  IV-2.
Sampling  element for collection of hydraulic
  fluid mist and vapors.
                           14

-------
A Sampler (Prefilter
+ Silica Gel Tube)
       Figure  IV-3.   Sampler deployment  around the  Plant A drumming  operation.

-------
colocated  samplers  (of the same  type  as  described above) were  operated  by
plant personnel during each test.

               All  samplers  were operated simultaneously durina periods  of
actual  drumming  rather than the  full  8-hr work  shift.  Even  though  the mea-
sured concentrations would not  represent  normal  workday values,  they would be
substantially  more  useful  for  purposes of exposure model  development and
validation.

               Because the  ventilation patterns  and secondary source activi-
ties might strongly influence the resulting concentration field  in the worker
zone, these  conditions  were defined with plant  personnel to  the extent pos-
sible before each test.   It was desirable to minimize secondary  source activi-
ties (e.g.,  operation  of the 5-gal  drumming  line) because  of the resulting
interferences with model  development.

               During each sampling  period, a warm-wire anemometer (Hastings-
Raydist  air  meter)  and smoke traces were  used to  determine the  ventilation
flow pattern  in  the vicinity of the work station.   Also the  volumetric flow
rates at all major ventilation  inlets  and outlets of the central  drumming fa-
cility were  determined  to provide for  calculation  of  overall air exchange
rate.

               Worker positioning and  movement  during each sampling period
was documented by determining the relative amounts of time spent  in  each work
zone.  Worker  activity was  keyed to the  number  of drums  filled  during each
time increment,  which was  monitored separately.  Preliminary analysis of
worker positioning  was used to  define  the work zone, for  purposes of sampler
placement as described above.

               b.   Analysis Procedure

               After a test, each exposed sample collection medium was removed
from the sampling system, capped, and  placed in  a  refrigerated container  for
return to  the  laboratory  for analysis.  After arrival at  the  laboratory,  the
silica gel  tubes were broken and the front and backup sections transferred to
1-dram vials.  The  filters,  already transferred to autosampler vials at the
end of each test,  did not require transfer.

               Samples were extracted by adding 1.0 ml  of acetone, accurately
measured, to each vial  and  allowing the vials to  sit for 30  min with  occa-
sional  swirling.   Aliquots of the extracts from the tubes were transferred to
autosampler vials for  analysis  by gas chromatography.   The filters were re-
moved from the extracts prior to analysis.

               The silica gel  and filter extracts were analyzed separately by
gas chromatography.   The analytical  parameters  of the method were as follows.
                                      16

-------
                          Gas Chromatographic System

          Instrument:  Varian 3700 gas chromatograph
          Detector:  Flame ionization
          Column:  DB-5 capillary, 15-m x 0.25-mm ID, fused silica
          Injection mode:  Splitless for 45 s
          Carrier gas:  Helium
          Carrier gas linear velocity:  30 cm/s
          Splitter flow rate:  100 cc/min
          Inlet temperature:   250°C
          Detector temperature:  300°C
          Column oven temperature program:  40°C for 1 min, then 40 to 250°C
            at 10°C/min
          Data handling:  Nelson analytical data system

               Analyte concentrations in the extracts were determined by com-
parison of the electronically integrated peak areas of chromatograms obtained
from injection of  standard solutions with  peak  areas obtained  from  injection
of the sample extracts.   Response factors obtained from injection of standard
solutions of  each  analyte  (1-100 ug/mL)  were found to  vary  ±20% (RSD) for
Fluid 1 and ±25%  (RSD)  for Fluid 2 over the range of 1 to 100 |jg/mL.   These
response factors were considered invariant,  and the averages  were  used to
calculate analyte  levels  in  the samples.   The  method  has  shown a limit of
detection (LOD) of  0.5  ug/sample (based  on a 1-mL extraction  volume) which
corresponds to a  discernible peak above  the noise level of the  instrument.
The limit of  quantisation (LOQ)  has been set at 2 ug/sample (based on a 1-mL
extraction volume) which is 4 x LOD.

               Masses found on the filter and in the silica gel were combined
to determine  total  mass  sampled.   Concentration values were calculated by:

     Concentration fufl/m3} -   Total  mass  collected (uq) x 1.000 (L/m3)
     Loncentrat10n (ug/m ) -  Sam1
                             Samp11ng rate (L/min) x Samp1ing time (min)

          3.  Test Results

          Two tests  of  drumming were performed as summarized in Table IV-1.
Limitations  in  the drumming frequency of  the  two more volatile  hydraulic
fluids prevented additional testing of these fluids.

          As indicated in Table IV-2, the automated drum fill time for either
test fluid  was  approximately  1 min.   Over 99% of  the  weight of fluid was
added in the fast fill mode at a rate of approximately 60 gpm.   The lance re-
ceded from  the  drum  during filling such that the discharge (bottom) end re-
mained below the fluid level in the drum.

          During these tests of drum filling,  the operator was observed to
spend about  two- thirds of  the  time at the  south  end of the work platform
labeling the drums and placing  loose caps on them.  Approximately 30% of the
time was spent  crimping  the caps on the drum openings after filling.  Only
about 3% of the time was spent  in front of the drum filling apparatus, posi-
tioning the drum and  activating the automated filling operation.
                                      17

-------
     Table IV-1.  Plant Tests of Drumming Operation
No. of Test
Test Date Substance 55-gal drums duration
drummed filled (h:min)
A-l 4/23/85 Fluid 1 100 drums 3:05
A-2 4/25/85 Fluid 2 59 drums 1:38
Air
temperature
(°F)
82
70
     Table IV-2.  Representative Drum Fill Cycles
                                    Fluid 1     Fluid 2
Downward. movement of lance begins
Fill starts
Fast fill ends
Slow fill ends
Lance retracted from drum
Total fill weight
Fast fill total
Slow fill total
0:00
0:03
0:54
0:57
1:00
475 Ib
472 Ib
3 lba
0:00
0:04
1:06
1:14
1:16
454 Ib
451 Ib
3 lba
In half-pound increments.
                          18

-------
               a.  Ventilation Patterns

               The air velocity distribution at breathing height (1.5 m above
work platform)  is  shown  in Figures IV-4 and IV-5 for Tests A-l and A-2, re-
spectively.  Air speeds at worker breathing heights (1.5 m above the drumming
platform) were  generally  low  (< 50 fpm) with a forward component across the
work platform.  The characteristic upward component was due to (a) the higher
velocity air stream above the work station created by the overhead wall fans,
and  (b)  the  location of  primary  air  exhaust vents in the  ceiling of the
building.  The  lateral  (northerly) component of the air motion at the south
end  of the work platform was caused by the  southerly ambient wind entering
door openings along the south wall of the building.

               A study of total workroom ventilation indicated good agreement
between  the  total  air inflow and  outflow of about 12,000 cfm.  As shown  in
Table IV-3,  this volumetric flow  rate  produced about 26 air changes per hour
within the drumming facility.
                      Table IV-3.  Plant A Ventilation3
              Inflow total                     12,711 ftVmin

              Outflow total                    10,789 ftVmin

              Ventilation rate                 11,750 ftVmin
                (inflow-outflow average)

              Total room volume                29,900 ft3

              Free room volume                 26,910 ft3
                (90% of total room volume)

              Room ventilation turnover rate   26.2 air changes
                                                 per hour

              aVentilation study conducted on April 26, 1985.
               b.  Concentration Patterns

               All of  the  Fluid  1  analyte masses collected on the prefilters
(mist) and in the silica gel traps (vapor) during Test A-l were below the 2-pg
limit of  detection.   Based on a sampling rate  of 0.30  L/min and a sampling
duration  of  185  min,  the minimum detectable  concentration corresponding  to a
2-ug sample  mass is  36 pg/m3.  Thus,  if  the  analyte mass was equally present
as mist and vapor, the total analyte concentration could have been as high as
72 pg/m3.  The total  analyte concentration  measured by the plant laboratory
on the operator  personal sampler for Test A-l was 150 pg/m3.

                                      19

-------
           Sampler (Prefilter
           + Silica Gel Tube)
                   ••
ro
o
                      L
                                                                  I
                                   Wind Vector
                                   = 50 fpm
                                    Down,
      « SL
                               12
                                *— I anr^o F
i
                 
-------
ro
             Sampler (Prefilter
             + Silica Gel Tube)
                                                                                   t
                                                                      Wind Vector
                                                                       = 50 fpm
            I
Up
 i@ 180 fpm  t f
••	•' '•
                  Up
           Up
                                                  Variable
                                               QQOQQ
                                         Lance Fill Position
               Variable
               10 fpm
                                             Up
                                        I
                                                        Up
                                                                  Up
                                                               Up
                                                                                                       ' J
                                                                      Note: Strong flow of air entering
                                                                      drumming room from south door
                                                                      (upper right corner of diagram).
                                   r
                        Overhead
                           Fans
                   Figure IV-5.  Overhead view of airflow at 1.5 m  above drumming platform
                                           for Test A-2 (Fluid 2).

-------
               Unlike the  case for  Fluid  1,  detectable concentrations of
Fluid 2 analyte were measured at several sampling locations in Test A-2.   The
distributions of analyte concentrations in the form of  vapor and mist are
shown in Figures IV-6 and IV-7, respectively.   Analyte concentrations of mist
exceeded concentrations of  vapor  at nearly all of  the  sampling locations.
The highest  mist concentrations  occurred at sampling locations downwind of
the drum filling apparatus,  as would be expected.  Except for  the sampling
points near  the drum filling apparatus, the vapor concentrations were below
the detectable minimum of about 76 ug/m3 (assuming a 0.30 L/min sampling rate).
Although the  limit  of detection  for the Fluid 2 analyte mass was also 2 ug,
the minimum  detectable concentrations  were larger  than those for Fluid 1
(Test A-l)  because  of the  shorter sampling duration (98 min) for Test A-2.
The total analyte concentration measured by the plant laboratory on the oper-
ator personal sampler for Test A-2 was 240 ug/m3.

               The higher Fluid 2 concentrations measured during Test A-2 are
thought to  be due primarily to the fact that the local  side draft hood at the
drumming station was inadvertently not operated by plant  personnel.   Even
though the 5-gal drumming  station was active during Test A-l along with the
55-gal drumming line, both side-draft hoods were in operation throughout this
earlier test.  Analyte masses  and  concentrations for Tests A-l  and A-2 (both
prefilter and silica gel  trap values)  are  given  in Tables IV-4 and  IV-5.

     B.   Bagging Operation (Plant B)

          1.   Introduction

          The second  plant-scale  study  described  herein was directed to the
collection  of specific data on worker breathing zone concentrations resulting
from the bagging of solids.   These data will  be used to develop and  refine
models for  the assessment of inhalation exposure for workers involved in such
packaging operations.

          The testing was conducted at a plant facility where granular sodium
tripolyphosphate is packaged in 50 and 100 Ib bags.   This facility was judged
to be representative of full-scale bagging operations and associated emission
controls.  The testing took place in April of 1985.

          Sodium tripolyphosphate (STP) is incorporated into synthetic deter-
gent granular formulations  as  a  "builder" or sequestering agent.   It serves
to eliminate  interference  with the detergent action by the calcium and mag-
nesium ions  (hardness)  in  the water used  in the wash solution.  STP  may be
used in powder, prill, or granule form.

          At  Plant B, 40,000 to 60,000  Ib of tripolyphosphate are bagged on a
two-spout autopacker  (see  Figure  IV-8)  in a period of about 6 hr during the
day shift.    Two grades of  product  are packaged—a food  grade  and a technical
grade.  Data  on  the product texture as provided by plant personnel indicate
that  these  powders have silt  contents  (i.e.,  portions  passing a 200-mesh
screen) of only 1  to 2%.   According to  plant personnel, worker  exposure data
based on periodically collected  personal samples indicate maximum exposure
levels of 2 to 3 mg/m3.
                                      22

-------
oo
                  • Sampler (Prellller
                    -» Silica Gel Tube)
<69
                                                                                                                <80
             <74
                          Figure IV-6.  Vapor  concentrations (yg/m3)  of Fluid 2 observed
                                                during Test A-2.

-------
                    • Sampler (Pretlller
                      + Silica Gel Tube)
  169
              108
ro
                                                                            Trap Only (Vapor and Mist)
                           Figure  IV-7.
Mist concentrations (yg/m3)  of Fluid 2 observed
       during  Test A-2.

-------
     Table IV-4.  Mass Concentrations Observed During
                    Test A-l (185-min test period)
Sampler
location
ID
1
2a
2b
3
4
5
6
7
8a
8b
9a
9b
10
11
12
13
14
15a
15b
16
17
Sampler
flow rate
(L/min)
0.301
0.300
0.307
0.299
0.271
0.274
0.276
0.279
0.271
0.300
0.285
0.309
0.276
0.258
0.275
0.288
0.300
0.275
0.273
0.307
0.280
Total analyte
(UQ)
Mist
(filter)
< 2a
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Vapor
(trap)
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Mass
concentration
(ug/rn3)
Mist
< 36
< 36
< 35
< 36
< 40
< 40
< 39
< 39
< 40
150b
< 38
180b
< 39
< 42
< 39
< 38
< 36
< 39
< 40
< 35
< 39
Vapor
< 36
< 36
< 35
< 36
< 40
< 40
< 39
< 39
< 40
< 38
< 39
< 42
< 39
< 38
< 36
< 39
< 40
< 35
< 39
bAnalyte masses of < 2 pg are below the LOQ.
 Plant laboratory analysis for combined filter and trap.
                            25

-------
      Table IV-5.  Mass Concentrations Observed During
                     Test A-2 (98-min test period)
Sampler
location
ID
1
2a
2b
3
4
5
6
7
8a
8b
9a
9b
10
11
12
13
14
15a
15b
16
17
Sampler
flow rate
(L/min)
0.296
0.272
0.294
0.296
0.260
0.264
0.270
0.286
0.264
0.296
0.274
0.294
0.272
0.252
0.265
0.272
0.286
0.264
0.264
0.300
0.274
Total analyte
(UQ)
Mist
(filter)
4.9
2.9
4.0
2.3
4.6
3.5
< 2
a
2.2
3.8
2.9
< 2
< 2
2.5
< 2
< 2
3.1
3.2
2.9
Vapor
(trap)
< 2a
< 2
< 2
2.4
2.1
< 2
2.6
3.2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Mass
concentration
(ug/m3)
Mist
170
110
140
79
180
140
< 76
b
85
< 210b
140
240C
110
< 81
< 77
94
< 71
< 77
120
110
110
Vapor
< 69
< 75
< 69
83
82
< 77
98
110
< 77
< 74
< 75
< 80
< 77
< 75
< 71
< 77
< 77
< 68
< 74
.Analyte masses of < 2 pg are below the LOQ.
 Silica gel trap only.
 Plant laboratory analysis for combined filter and trap.
                          26

-------
ro
             Fine
             Particle
             Collection
             Barrel
Coarse
Particle
Collection
Barrel
                                                                                      Control
                                                                                      Panel
                           Pallet
                           Conveyor
                                                                28  Inch
                                                                Overhead
                                                                Fan
                                                                  J
                                                                                                   1
                                                                                                   m
                                         Figure  IV-8.  Plant  B bagging  facility.

-------
          The operator, seated  in  front of the autopacker during its opera-
tion, slides a  bag  onto one of two  spouts  and then activates the flow of
powder.   Some visible  emissions can  be  observed as the bag is filled.  Addi-
tional emissions occur when each  self-sealing bag is tipped from the spout
onto the conveyor below.

          The ventilation airflow across the packaging equipment is  strongly
influenced by the 28-in. overhead fan to the rear of the operator seat.  The
local side-draft hood  behind the  spouts for capture of dust emissions does
not  induce a  significant  air draft  in  front of  the  spouts near the worker
breathing zone.

          2.   Experimental  Methods

          This section describes the sampling and analysis procedures used in
the testing of the bagging station in April  of 1985.

               a.  Sampling Procedure

               A network  of 24 samplers (12 pairs)  was  deployed at fixed
points in  the vicinity of the bagging  operation.  As shown in Figure IV-9,
both  elements of the sampling pair collected particulate on 37-mm Millipore®
filters (0.2 pm  pores) in  cassettes.  One sampler in each  pair was fitted
with  a  Dorr-Oliver 10-mm  nylon cyclone preseparator.   At a  flow  rate  of
2 L/min, this cyclone  has a 50% cut point of about 3.5 pm aerodynamic diameter
for  measurement  of  respirable particulate concentrations.  The other sampler
in each pair was used  to measure total  suspended particulate concentrations.

               A standard  flow  rate  of 2.0 L/min (±5%) through each filter
cartridge was ensured  throughout substrate buildup by maintaining sonic flow
through a critical orifice.  A 20 ga. x 1 in.  hypodermic needle served as the
critical orifice in each  line connecting a sampler to the suction mainfold.

               Most of the samplers were positioned at breathing height along
the boundaries of the  bagging equipment, as shown in Figure IV-10.   Two sets of
samplers were colocated to provide for determination of measurement precision.
Samplers 3a and 3b were positioned for location of the plume centerline down-
wind  of  the points of particulate  release.   An operator personal sampler
fitted with  a 10 mm cyclone (of the  same  type as  described above) was op-
erated by plant personnel during each test.

               Further particle size characterization was provided by a Sierra
Instruments Model 230  high-volume (20 cfm) five-stage cascade impactor with a
cyclone preseparator.   This equipment was operated near the point of release
to measure particle size  distribution in the range of 1 to 15 umA.   In this
study, the particle size fractions < 15 pmA are referred to as inhalable par-
ticulate.   Finally, a  RAM-1 aerosol analyzer was used periodically to monitor
the  time  dependence of particle concentrations near the bagging operation.

               All  integrating  samplers were  operated simultaneously during
periods of actual  bagging  rather than the full 8-hr  work shift.  Even though
the measured STP concentrations may  not represent normal workday values, they
are  substantially more  useful for purposes of model refinement and validation.


                                      28

-------
                           Respirabje Dust|Sampler
                                                       a
            • Nylon Cyclone Presepafator
Figure IV-9.
Sampling elements  for  collection of sodium
  tripolyphosphate particles.
                          29

-------
                                                Cassette Sampler Pair

                                             O  Cyclone/Cascade Impactor
Figure  IV-10.
Sampler deployment around the Plant B bagging
  operation.
                            30

-------
               During each sampling  period,  warm-wire anemometers and smoke
traces were used to determine the ventilation flow pattern in the vicinity of
the work  station.   Also,  the volumetric flow  rates  at all major building
ventilation inlets  and outlets were  determined to provide  for calculation of
overall air exchange rate.

               b.  Analysis Procedure

               At the end of each particulate sampling period, the cartridges
with the  Millipore® filters  were removed from the sampling system.   Plastic
stoppers were replaced on the cartridge inlets and outlets, and the cartridges
were placed in  protective containers for shipment to the  laboratory.  After
each day  of testing,  the cyclones were washed with  distilled water  and  air
dried.

               The  exposed Millipore®  filters  in their respective cassettes
were returned to the MRI  laboratories to be analyzed for phosphate ion with a
Technicon AutoAnalyzer system.  The  specific reagent preparation and analysis
procedures were  adapted  from those  found in  EPA  Method 365.1  for  phosphorous
in water and wastes (EPA-600/4-79-020).

               In the Method 365.1  analysis  scheme,  sulfuric  acid hydrolysis
converts  the  polyphosphates  to  the orthophosphate  form for  reaction  with
ammonium  molybdate  and  antimony  potassium  tartrate.   The antimony-
phosphomolybdate complex  formed  by  this reaction is reduced to an intensely
blue-colored  complex  by  ascorbic acid.  The  color  is proportional  to the
phosphorous concentration and is measured with  the  Technicon AutoAnalyzer.

               The  following modifications to Method 365.1 were incorporated:

          1.   The  system was standardized with  sodium tripolyphosphate  (STP)
               from the fill  line.   The STP was  hydrolized following steps 8.2
               through 8.2.3 of  Method 265.1.   Standardization  with  STP  from
               the  fill  line permitted the results  to be  presented  directly
               as ug of STP.

          2.   Brig-35  (nonionic wetting  agent) was not  used  in  the system
               because it interfered with the analysis.

          3.   Acid wash  water was   not used because it interfered with  the
               analysis.

          4.   The  stable solution  (Note 1:  under step 7.5) was  not used be-
               cause  it  proved unsatisfactory.   The  single mixed  reagent was
               used.

          5.   The  sample rate (Figure  3 of method)  was changed to 0.6 mL/min.

          6.   The  mixed  reagent rate was changed to 0.4  mL/min.

          7.   The  waste  rate was changed to 0.8 mL/min.

          8.   The  delay  coil  in the heating bath was  changed from 13  turns  to
               22 turns  to increase the development  time.

                                       31

-------
               The phosphate  analysis  of each participate sample was initi-
ated by  removing  the filter from the  cartridge and placing the  filter  (ex-
posed side  up)  in a bottle with a diameter  slightly  larger than the  filter
diameter.   After  30  mL  of  deionized water was added,  the  solution was boiled
for 10 to 15 min to a volume of 20 mL.   The filter was removed from the solu-
tion and washed  three  times with deionized water, with the wash water added
to the  sample.   After  1 mL of 11 N sulfuric  acid  solution was added  to the
sample solution,  it  was gently boiled  for 30  to 40 min to obtain a volume of
less than 10 ml.   Then  the solution was cooled and made to 10 mL in a volu-
metric flask and then refrigerated.   This procedure is detailed  in Table IV-6.
           Table IV-6.  Sodium Tripolyphosphate Analysis Protocol
                        (based on EPA Method 365.1 for phosphorous)


1.   Place plugs in  both  ends of filter cassette and air blow off any par-
     ticulate from outside  of cassette.   Do this in an area which cannot be
     contaminated.

2.   Open cassette and observe filter condition.

     a.   If filter  is  torn,  then place both backup  filter and  Millipore®
          filter in  ~ 60 mL Pyrex® bottle.   Otherwise place only Millipore®
          filter in each uniquely labeled filter bottle.

     b.   Rinse front half  of cassette interior with DI water (three small
          rinses) and add rinse  water to  filter bottle.   Add  DI H20 as
          necessary  to  attain approximately 30  mL total volume for each
          filter sample.

3.   Take water blanks  from  unused cassettes to check handling  and DI water
     contamination.    (Minimum of  one  DI  water blank per day in  addition to
     field blanks.)

4.   Boil the solutions in Pyrex® bottles (containing the filter/rinse water)
     for about 10-15 min to reduce volume to ~ 20 mL.

5.   Remove filter from bottle  and rinse filter with DI water using funnel.
     Rinse volume should  be  added to bottle to total  approximately 30-40 mL
     volume.   Discard filter.

6.   Add 1 mL of 11N H2S04 to the 30-40 mL sample volume.

7.   Again, distill  off water to obtain a sample volume of  less  than 10 mL.
     Allow the sample to cool  to room temperature.   Pour the concentrate into
     a 10 mL volumetric flask.  Dilute the sample to 10.0 mL volume.   Transfer
     the sample back to the original  Pyrex® container,  cap, and  store under
     refrigeration for  subsequent analysis  using the Technicon AutoAnalyzer.


aAll glassware  should  be washed  with  hot  1:1 HC1 and triple rinsed with
distilled H20.


                                      32

-------
               The sample  analysis  was  initiated by bringing the samples to
room temperature.  The sample and container were weighed and recorded as value
(A).  One drop of phenolphthalein indicator was added to each sample solution,
and the  solution taken to the end  point with  IN  sodium  hydroxide solution.
The container was then reweighed with the value recorded as (B), and the solu-
tion transferred  to  a  clean  container.  The original container  was  dried and
weighed and recorded as value (C).

               The total mass of sodium tripolyphosphate in the original sam-
ple was calculated as follows:
          Total ug STP  =
                        _  B-C
                           A-C
                    ug/mL value from Technicon •  10 mL
Mass concentrations were calculated by:


     r«n^Qn+M+,-«n f„f,/m3^ -   Total mass collected (pq) x 1,000 (L/m3)
     Concentration (ug/m3) - Sampling rate (L/min) x sampling time (min)
          3.  Test Results

          Three tests of drumming were performed as summarized in Table IV-7.
The  first  two  tests  (B-l and B-2) involved the packaging of technical grade
STP, and the third test (B-3) involved food grade STP.
                Table IV-7.  Plant Tests of Bagging Operation
Test date
4/23/85
Packaged
substance
Sodium tri poly-
Packaging
operation
80 100- Ib bags
Test
duration
(hr:min)
2:11
Air
temperature
77
4/24/85
4/25/85
phosphate (tech-
nical grade)

Sodium tripoly-
phosphate (tech-
nical grade)

Sodium tripoly-
phosphate (food
grade)
and 200 50-Ib
bags

514 50-Ib bags
736 50-Ib bags
3:19
2:54
63
78
                                      33

-------
               a.  Ventilation Patterns

               The air velocity  distribution  at operator beathing height is
shown in Figures IV-11 through IV-13 for Tests B-l, B-2, and B-3, respectively.
The overhead fan at the rear of the operator produced the primary flow pattern
with speeds of several hundred feet per minute.   This primary flow stream drew
air from either  side  and blew it  across  the  packaging equipment as shown.

               A study of total workroom ventilation indicated good agreement
between the total  air inflow and  outflow of  about  23,000 cfm.   As  shown  in
Table IV-8, this volumetric  flow rate  produced  about 15 air changes per hour
within the packaging facility.
                      Table IV-8.  Plant B Ventilation3
            Inflow total                           21,583 ftVmin
            Outflow total                          24,200 ftVmin

            Ventilation rate                       22,892 ftVmin
              (inflow-outflow average)

            Total room volume                      98,925 ft3
            Free room volume                       89,032 ft3
              (90% of total room volume)

            Room ventilation turnover rate = 15.4 room changes/hr
            aVentilation study done April 26, 1985.
               b.   Concentration Patterns

               Analyte masses  and  calculated concentrations for Tests  B-l,
B-2,  and B-3  are  given  in  Tables  IV-9,  IV-10, and IV-11,  respectively.
Analysis  of  the  samples  obtained at location 3b and  the  personal samples was
performed by the plant laboratory.

               For each test, the concentration pattern was found to be fairly
uniform  (as shown in Figures IV-14  through IV-16), with only a small fraction
of  suspended particulate concentration consisting of  STP  particles in the
respirable size  range  (^ 10 umA).   The airborne particle size distributions
for  the  three  tests,  as determined from the high-volume cyclone/cascade im-
pactor data, are plotted in  Figure  IV-17.
                                      34

-------
        Wind Vector
        = 200 fpm
      a
                                  t
                                   ! 900 fpm
                        Up
 t
                               • Filling •
                               Spouts
                                  Operatoc/Seat
                                                              Variable
Cassette Sampler Pair

Cyclone/Cascade Impactor
A
                                       28 Inch
                                      Overhead
                                         Fan
      Figure  IV-11.  Overhead view of airflow at operator
             breathing height during Test B-l.
                             35

-------
Wind Vector
= 200 fpm
^ r-
c

N
vup
•
rn




<
i
i
4
i
k
Down
i
i
Filling
Sn/*\i i+o


-




\
Up
•
| —



"1

      Up @ 60 fpm


•  Cassette Sampler Pair

   Cyclone/Cascade Impactor
                                    Operator Seat
                                    t
                            illo:
                                                                Up;Variable
A
                                         28 Inch
                                        Overhead
                                           Fan
      Figure IV-12.   Overhead view of airflow at operator
              breathing  height during Test  B-2.
                               36

-------
 Wind Vector
 = 200 fpm
         a
                                     i
                           \
                                   500 fpm


                                      Down
                                  • Filling •
                                  Spouts
                                     Operator Seat
                                                                 UP
          UP
T
Cassette Sampler Pair

Cyclone/Cascade Impactor
A
                                      28 Inch
                                     Overhead
                                        Fan
         Figure IV-13.   Overhead view of airflow at operator
                breathing height during Test B-3.
                              37

-------
     Table IV-9.  Mass Concentrations Observed During Test B-l
                    (131-min sampling period)
Sampler
ID
1
1- Cyclone
2
2-Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4-Cy clone
5
5-Cyclone-
6
6-Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8-Cyclone
9
9-Cyclone
« -, d
Cyclone-
Cascade Impactor6
Sampler
flow rate
(L/min)
2.00
2.0a
^l
2.0a
1.90
2.0a
2.00
2.0
1.95
2.0a
1.90
2.00
2.00
1.95
2.10
2.05
2.00
2.05
1.95
2.05
1.95-
2.20
566*
566
Total analyte
as Na5P307 (pg)
386
19
239
131
328
b
114c
< 35C
258
55
697
98
421
23
158
45
301
16
370
30
442
< 3
55; 700,
60,500
STP
concentration
(M9/m3)
1,500
70
940
500
1,300
b
440
< 130
1,000
210
2,800
370
1,600
90
570
170
1,200
60
1,400
110
1,700
< 10
740
800
a
 Estimated value.
"Sample destroyed.
 .Plant laboratory analysis.
 Particles > 15 pmA; 134-min sampling time.
 Particles < 15 pmA including background; 134-min sampling time.
                                38

-------
     Table IV-10.  Mass Concentrations Observed During Test B-2
                     (199-min sampling period)
Sampler
ID
1
1- Cyclone
2
2-Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4-Cyclone
5
5-Cyclone
6
6-Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8-Cyclone
9
9-Cyclone
« , b
Cyclone
Cascade Impactor0
Sampler
flow rate
(L/min)
2.00
1.95
1.95
2.05
1.95
1.95
2.00
2.10
2.00
1.95
1.95
2.20
2.00
1.95
2.00
2.05
2.05
2.05
1.95
2.05
2.05
1.95
566
566
Total analyte
as Na5P307 (pg)
513
86
286
67
303
44
298a
< 35a
598
23
845
< 3
687
51
713
16
566
49
7,480
61
1,400
201
168,000
78,400
STP
concentration
(|jg/m3)
1,300
220
740
160
780
110
750
< 80
1,500
60
2,200
< 10
1,700
130
1,800
40
1,400
19,000
150
3,400
520
1,600
750
.Plant laboratory analysis.
 Particles > 15 jjmA; 184-min sampling time.
 Particles < 15 umA including backbround; 184-min sampling time.
                                   39

-------
      Table IV-11.  Mass Concentrations Observed During Test B-3
                      (174-min sampling period)
Sampler
ID
1
1-Cyclone
2
2- Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4- Cyclone
5
5-Cyclone
6
6- Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8- Cyclone
9
9-Cyclone
r* , fa
Cyclone
Cascade Impactorc
Flow rate
(L/min)
2.10
2.00
2.10
2.10
1.90
1.90
2.00
2.20
1.95
2.00
1.95
2.05
2.05
1.95
1.95
2.00
2.00
2.05
1.95
2.05
1.95
2.05
566
566
Total analyte
as Na5P307 (ug)
794
156
2,520
65
722
247
483a
< 35a
927
174
2,800
188
1,530
145
693
105
997
307
1,020
71
1,050
110
64,800
61,800
STP
concentration
(ug/m3)
2,200
450
6,900
180
2,200
750
1,400
< 90
2,700
500
8,200
530
4,300
430
2,000
300
2,900
860
3,000
200
3,100
310
1,300
1,200
.Plant laboratory analysis.
 Particles > 15 umA; 88-min sampling time.
 Particles < 15 umA including background;  88-min sampling time.

-------
1450
                                       1470.
                                                    940
                                                               1010
                  1730
                                                                    1530
                                                               Bagging Spout
                                                      •  Cassette Sampler

                                                      O  Cyclone/Cascade Impactor
           Figure IV-14a.  Total suspended particulate concentrations of
                            STP (jjg/m3)  for Test B-l.
                                        41

-------
110
                                                                210
                                                                    7401
                                                                Bagging Spout
                                                       •  Cassette Sampler With
                                                          Cyclone Precollector

                                                      O  Cyclone/Cascade Impactor

                                                         aSample destroyed

                                                         blnhalable Particulate,
           Figure IV-14b.   Respirable particulate concentrations  of STP
                             (ug/m3) for Test B-l.
                                       42

-------
                                                      1500
                                              •  Cassette Sampler Pair

                                             O  Cyclone/Cascade Impactor

                                                aSample Contaminated
Figure  IV-15a.  Total  suspended particulate concentrations of STP
                 (ug/m3)  for Test B-2.
                              43

-------
150
                                                                 60
                                                                     750'
                                                                Bagging Spout
                                                          Cassette Sampler With
                                                          Cyclone Precollector
                                                       O Cyclone/Cascade Impactor

                                                         alnhalable Particuiate
          Figure IV-15b.   Respirable particulate concentrations of STP
                            (|jg/m3) for Test B-2.

-------
               fin
                                        2170.
                                                                2730!
                  3100
3020
                                                                     2540
                                                                Bagging Spout
                                                       •  Cassette Sampler

                                                       O  Cyclone/Cascade Impactor
        Figure IV-16a.   Total suspended particulate concentrations of STP
                          (jjg/m3) for Test B-3.
                                       45

-------
200
                                        450
                  310
                                                    180
                                                                 500
                                                                     1240*
                                                                Bagging Spout
                                                       •  Cassette Sampler With
                                                          Cyclone Precollector

                                                       O  Cyclone/Cascade Impactor

                                                         alnhalable Paniculate
         Figure IV-16b.   Respirable  particulate concentrations of STP
                           (|jg/ni3) for  Test B-3.
                                      46

-------
                     WEIGHT % GREATER THAN STATED SIZE
       999 99.8  99 98  95  90   80 70 60 50 40 30  20   10  52  1 050201
100



50


20
^ 10
(0

O
K
o
S S
oc
in
<
5
UJ y
PARTICL




05


0.2
0.1





























































































































I

|
Jb






















J
i

















B-2-



y
/
/
/ ^
/
/
-B-3

















•M
/




/
u
f



















r
t
/
//
/






















y
//.


























f
V
«-


























p
-I


























3-1















































































































































































































10




50



10







2





05


U £
01
                     WEIGHT % LESS THAN STATED SIZE
Figure IV-17.  Airborne particle size distributions  obtained  using
                 cyclone/cascade impactor.
                                47

-------
In the  relatively  stagnant  flow  region  immediately  in  front  of  the  operator,
it is  likely  that  gently  circulating  air  transported emissions  from the  bag-
ging operation directly to the breathing zone of the operator.

               The concentrations  observed  during Test B-3 are  significantly
higher than the  values  measured for Tests  B-l and  B-2.  This resulted from
the higher fraction  of  fines in food grade STP compared to  technical grade
STP,  as detailed below.

          c.    Bulk Particle Size Distributions

               Three representative samples of  the  bagged STP were obtained
on the 3 days of testing at Plant B, to be analyzed for particle size distri-
bution.   These samples  were dry sieved, yielding the results shown in Table
IV-12.   Also  shown are  dry  sieving  results  from plant  laboratory analysis of
earlier samples of food grade and technical  grade STP.


               Table IV-12.   Bulk STP  Particle Size  Distribution
Cumulative percent
Mesh
No.
10
16
20
30
40
50
60
80
100
140
200


Particle
diameter
(urn)
> 2,000
> 1,190
> 850
> 590
> 425
> 297
> 250
> 180
> 150
> 106
> 75
< 150
< 75
Plant
Food
grade
0
0.13
3.6
17
-
57
-
79
87
-
~
13
""
B analyses
Technical
grade
0.002
0.016
4.5
21
-
67
-
86
91
-
~
9
™

Technical
grade
(Test B-l)
-
-
2.6
32
61
-
89
95
97a
99
99
2.9a
0.71
MRI analyses
Technical
grade
(Test B-2)
-
-
-
30
56
-
85
93
96a
98
99
4.1a
0.79

Food
grade
(Test B-3)
_
-
-
17
38
-
72
84
89a
95
97
II3
3.1
   Value  obtained  by  interpolation  using  log  normal  distribution.
                                     48

-------
               As indicated in Table IV-12, the MRI analysis of the technical
grade STP yielded a  smaller weight fraction of  finer  particles  (particles
smaller than 150 |jm in physical diameter).   If the process for the manufacture
of technical grade STP  is  not as closely  controlled as for food grade STP,
significant differences in bulk size distributions could occur.  The MRI and
plant analyses of the  food grade STP were comparable for the < 150 pm size
fraction.
                                     49

-------
V.  LABORATORY DRUMMING STUDY
          In the first laboratory pilot study, a walk-in flow tunnel was used
to characterize the low-flow  (^  2 m/s) dispersion of methane gas and ammonium
chloride particles at distances of 2 to 6 m from the point of release.   The
release configurations included:  single- and double-point releases into an
unobstructed flow, point release near a wall, point release past a side-draft
hood, and bung  release  from  a 55-gal drum (with and without a close fitting
hood).  A two-dimensional array of samplers  was  used  to  characterize disper-
sion patterns at  each downstream distance.   The  observed dispersion patterns
were compared to  existing dispersion theories  and  then used  as  the  basis  for
a modified dispersion algorithm  which balanced realism against ease of appli-
cation.  The  predictive  capability of the resulting dispersion model was
found to be  significantly better  than  the  other models  tested against the
laboratory data.   The first  laboratory pilot study is not further discussed
here.  (For details refer to  Appendix C.)

          The second  laboratory pilot study was conducted  using  a  drumming
station at MRI  and encompassed the measurement of  emissions, dispersion,  and
operator exposure  resulting  from  a  full-scale  55-gal  drumming  operation  in
the  low-flow wind tunnel.  A  55-gal  drum  with an added bottom drain was posi-
tioned on a  platform within  the flow tunnel.   A 300-gal reservoir was in-
stalled outside the tunnel  at a  height  of 2.13 m (7 ft)  above the floor.  The
overhead tank was fitted with polyethylene hosing  and a  pump for  transfer of
the  test  liquid to the  55-gal  drum at a fixed  rate  controlled by the hose
nozzle.  After  each  drum filling,  the  contents  of the drum  were pumped back
to  the  overhead tank.  The test liquids (ethylene  glycol, perchloroethylene,
and  methanol)  spanned a vapor  pressure range  from about 0.1 to 125 mm Hg at
25°C, as indicated in Table V-l.  The portions of the second laboratory study
having to do with  emissions,  dispersion,  and exposure are described separately
in  the sections immediately following.

      A.   Emission Tests

          1.  Experimental Apparatus

          The  test facility  used  for this  study was the pull-through  flow
tunnel  constructed in MRI  Field Station Building  No. 1.  As shown in  Fig-
ure V-l,  the cross section  of the tunnel was  2.7  m wide by  2.1 m high.   The
centerline  airflow velocity  was adjustable  over a  range of  mean  values from
0.5 to 2.2  m/s (0.9 to  4.9 mph).   The flow tunnel is described  further  in
Appendix C.

          A  standard 55-gal  steel  drum was positioned on a platform within
the flow tunnel,  as  illustrated in Figure V-2, such that the top of the drum
was 1 m above the tunnel floor.  The drum was equipped  with a  bottom outlet
consisting of  1-1/4  in.  cast iron pipe connected by  a flange to the center  of
the drum bottom.   A lever-lock  removable lid provided for easy access to the
 interior  of  the drum.

                                      51

-------
                               Table V-l.   Chemicals Used  in  Drumming Simulations
en
ro
Chemical
Name
Perchloroethylene
Methanol
Ethylene glycol

Formula
C2C14
CH40
C2H602

Molecular
weight
166
32.0
62.1

Surface tension
(dynes/cm @ 20°C)
31. 7b
22. 6C
48.4

Viscosity
(cP @ 20°C)
0.798 @ 30°C
0.55
19.83
0.0198
Vapor pressure3
(Torr @ 25°
18.5
125
0.098

C)




 Vapor pressures at other temperatures may be calculated using the Antoine Equation:
logloP = A - B/(C + t)

where P is the vapor pressure of the compound in mm of mercury (Torr) and where t  is  the
temperature in degrees centigrade.
                     Perchloroethylene
                     Methanol
                     Ethylene glycol
Antoine equation constants
ABC
7.02003 1415.49 221.0
7.87863 1473.11 230.0
8.2621 2197.0 212.0
Temp.
range (°C)

-20 to +140
25 to 112
           In  contact with vapor.
           In  contact with air.

-------
                                OVERHEAD VIEW OF PULL—THROUGH

                                           FLOW TUNNEL
cn
co
                           -4m-
                                             •8 m
o
                                                                   55-Gal Drum
                                                                   Wind Flow
                       Figure V-l.  Overhead view of pull-through flow tunnel.

-------
                        Vent:
in
-£>
                       Figure V-2.   Full  scale 55-gal  drumming operation in flow tunnel

-------
          A  300-gal  steel  reservoir was  installed  outside  the tunnel at  a
 height  of 2.13 m (7 ft) above the floor.   This overhead tank was fitted with
 a  cross-linked polyethylene  hose for transfer of  the  test liquid to the
 55-gal  drum.   A model  26 Gasboy vane type pump,  with direct drive connection
 to  a  115 V AC explosion proof  motor,  provided for liquid  delivery  to  the
 55-gal  drum  at up to  30 gal/min.  The pump was also  used to  return the  con-
 tents of the drum back to  the  overhead  tank.   A series of stainless steel
 full  flow ball valves  and innerconnecting chemically resistant hose was  used
 to  direct the  liquid flow and control its  rate of flow.  The  internal diameter
 of  the  hosing  was  3.18 cm (1.25  in.)  upstream  of  the pump and 2.54 cm
 (1.0  in.) downstream of the pump.

          The  liquid delivery apparatus  provided for either  top filling or
 bottom  filling of the  drum.   The top fill  nozzle  consisted  of a 6-in.  length
 of  1-in.  pipe  connected by a 90-degree elbow to a section of  hose fitted with
 a ball  valve for  control of flow  rate.  The bottom  fill nozzle  consisted of a
 36-in.  length  of pipe  connected  in  the  same  manner.   During drum filling,
 liquid  volume  was monitored  by the  totalizing flowmeter connected to  the
 pump.   A  thermometer was installed  in the flow  line  at the pump outlet  for
 monitoring of  liquid temperature.

          The  primary  analytical device for these studies was the Beckman  402
 total hydrocarbon analyzer.   To measure emissions,  a Teflon sampling line
 (1/8-in.  I.D.  by 13 to 20 ft in  length)  drew vapor effluent from the bung
 opening  surrounding the fill  nozzle as the drum  was  being  filled.  A 3-in.
 long  pipe was  screwed  into the  bung to provide an  extension  for prevention
 of  air  infiltration into  the  bung effluent.  The  sample was delivered to the
 analyzer  at  a  rate of  2.2 to 3.6 L/min such that the residence time  in  the
 sampling  line was between 0.5 and 1.5 sec.

          The  output of the analyzer was  fed to an  EPSON HX-20  microcomputer
 for sequential  calculation  of 5-sec  average analyte concentrations.   A Heath
 continuous recorder with a chart speed of  1 in/min produced a time concentra-
 tion profile of the emissions.

          During  an  emissions test, the  drum  fill  rate was  held constant
 over  the  entire  filling period.  The shut-off time corresponded either  to a
 certain liquid height on the sight tube or a reading of 53.6 gal on the  volume
 meter, which was  reset to zero prior to each test.  The fill time was recorded
with  a  stop  watch,  and the liquid temperature in the fill   line was recorded
 at the beginning and end of the fill.

          In the  course  of  the  experiments, tests of emissions from an  ini-
tially "dry" drum, and  from a just emptied ("wet") drum were performed.   Prior
 to  the  tests of perchloroethylene emissions from a dry drum, the drum was
cleaned using  a  cotton mop  and a  stream  of  compressed  air.  Then a rubber
 stopper (to which was  attached  a long wire for removal  purposes) was placed
 in  the  drum drain opening.  All  tests of ethylene glycol emissions were  per-
 formed with a  dry drum created by inserting a clean  polypropylene bag into
the drum.  The bag  was pressurized  to set against  the  drum walls, and then
the open  end of the bag was wrapped over  the  rim of the drum  and the lid
affixed.
                                      55

-------
           2.   Test Results

           Emission tests were performed for two chemicals, perchloroethylene
 (perc) and ethylene glycol ; methanol was not tested because of the potential
 explosion hazard.   Top and bottom fill  modes  and three fill  rates ranging from
 10 gal/min to 27 gal/min were included in the experimental design.  In addi-
 tion,  both the filling  of a dry drum and  a wet  drum were tested.   A  summary
 of drumming emission tests is shown in Tables  V-2 and V-3 for  perchloroethyl-
 ene and ethylene  glycol, respectively.

           For each drum filling  test,  the  emission factor (mass  loss  per  drum
 fill)  was determined as follows.   After the 5-s  integrated concentrations from
 the Epson microcomputer were averaged over the total fill time,  then the re-
 sulting mean  volumetric concentration  (ppm) was  corrected for  decay in  sensi-
 tivity of the Beckman 402  over  time.   Next  the adjusted mean  concentration
 value  was converted to a mass concentration  value (mg/m3), and then combined
 with  the volume of  air  displaced from the drum by the fluid.   Finally, the
 total  chemical  loss  per drum fill  was  calculated as  follows:
L =
                                    CV V0 MW)/1Q6 VM
where:     L  = chemical mass  loss per  fill operation, g

           Cy = mean of 5-s  integrated  chemical concentrations observed
               in air drawn from drum  during fill operation, ppm (or L of
               chemical vapor  in 106 L of air)

           VQ = drum volume  filled (equal to the volume of displaced air
               from drum),  gal

           1^ = molecular weight of drummed chemical, g/mole

           VM = molar volume of chemical vapor emitted from drum at ambient
               temperature and pressure, L/mole

           and k = 3.785 L/gal

           During the top  filling of  perc, a considerable amount of  froth was
observed to collect on the liquid surface.   This was also present during bot-
tom filling, but to a lesser extent.   Observed perc concentrations were above
the  saturation concentration  of about 25,000 ppm  (volumetric) at  25°C.   A
sampling line tip with internal diameter of 3 mm, 7 mm,  or 11 mm was selected
to provide for near-isokinetic  sampling of perc at  the respective fill rate.
The production of mist from the perc froth is believed to account for concen-
trations as high as  38,000 ppm at the drum bung opening during fill cycles.
Ethylene glycol  produced much less froth and presumably  less mist.

           For purposes of  separating vapor  emissions from mist emissions,  a
10 mm Dorr-Oliver  nylon  cyclone precollector  with 50% cutpoint  of about
3.5 umA was attached to the inlet of the sampling line for about half of the
ethylene glycol  tests.
                                      56

-------
            Table V-2.  Emissions Test Data  -  Perch!oroethylene
Date
8/1/85
8/1/85
8/1/85
8/2/85
8/2/85
8/2/85
8/2/85
8/2/85
8/2/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/6/85
8/5/85
8/5/86
8/6/85
V6/85
8/6/85
8/9/85
8/9/85
8/9/85
8/9/85
8/9/85
Run
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Initial
drum
condition
Dry
Dry
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Dry
Wet
Wet
Wet
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Filling
mode
Top
Bottom
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Bottom
Bottom
Fill rate
(gal /mi n)
24.8
25.7
25.5
25.5
25.5
25.5
24.4
24.4
25.5
25.5
25.5
25.5
25.6
24.4
24.4
9.8
10.3
9.1
11.2
11.2
10.3
11.9
11.4
11.2
11.2
11.7
11.4
22.4
16.8
16.8
16.3
16.7
Fluid
temp. (°C)
21
22
22
25
26
26
27
27
27
23
24
24
25
26
23
27
28
29
29
30
31
27
32
33
28
29
29
29
29
30
31
32
Total
mass
loss
(g)
42.0
18.5
46.5
45.0
19.1
19.5
30.8
31.1
32.4
25.2
42.2
41.8
47.8
46.2
47.4
50.0
51.2
50.0
52.4
52.8
54.7
16.4
13.8
11.5
31.8
34.7
31.7
43.6
44.9
46.4
22. 0
25.0
Drum
emission
rate3
(mg/sec)
324
148
369
357
151
155
234
236
257
200
335
331
380
351
360
152
164
141
182
184
175
60.7
48.9
40.0
111
126
112
304
235
242
112
130
Based on 53.6-gal total fill volume.
                                     57

-------
                                   Table V-3.   Emissions Test Data - Ethylene Glycol
en
00
Date
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Initial
drum
condition
Dry
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Filling
mode
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Top
Top
Top
Fill rate
(gal/min)
27.1
27.0
27.2
26.1
27.2
27.5
26.5
26.3
26.3
26.5
26.5
26.1
26.0
26.0
9.7
10.5
9.8
9.6
10.6
9.7
10.2
11.3
10.4
22.4
20.9
Fill
volume
(gal)
53.8
53.9
53.9
53.6
53.9
54.0
53.9
53.9
52.1
52.6
53.9
53.9
52.4
52.8
53.9
53.9
52.8
52.6
53.9
53.9
53.9
52.6
52.4
53.9
53.9
Fluid
temp. (°C)
22
22
22
22
22
23
23
23
23
23
23
24
24
24
17
18
18
20
20
20
21
21
21
21
21
Cyclone
(Y/N)
N
N
N
N
N
Y
Y
N
Y
Y
N
N
Y
Y
N
N
Y
Y
N
N
N
Y
Y
N
N
Total mass
loss
(g/drum)
a
a
a
2.0
2.2
b
b
b
1.9
1.9
0.5
0.39
0.40
0.33
0.31
0.22
0.21
0.22
0.20
1.1
1.1
1.0
1.0
1.1
1.1
Drumming
emission rate
(mg/s)
a
a
a
16.2
18.5
b
b
b
16.0
16.0
4.3
3.1
3.3
2.7
0.93
0.71
0.65
0.67
0.66
3.3
3.5
3.6
3.3
7.6
7.1
     .Test not used in summary;  no bag in drum.
      Sampling inlet with cyclone "contaminated"
with fresh air; very irregular chart trace.

-------
          The mist generated  in  both perc and ethylene glycol tests  is be-
lieved to  have consisted  mostly of particles greater than  50 |jm.   These
particles would have a tendency to settle out near the source.

          Continuously recorded  concentration-time  profiles  of the dry drum
emission tests are presented  in  Appendix D together with information on ob-
served maximum 5-sec mean  concentrations and the  integrated mass  lost during
each drum  fill.   Figure V-3 presents a characteristic comparison  of profiles
for top  fill  versus  bottom fill  of  perc.  The top  filling of  a drum  causes
exit stream  concentrations to rise quickly with  time as  the nozzle sprays
liquid into  the  drum,  resulting  in considerable  splashing and mixing with
air.  During  a bottom  fill, the  spray of liquid produced  by the lance at the
bottom of  the drum  is  quickly submerged causing concentrations to rise more
slowly with time.

          To  determine whether the concentration-time profiles of ethylene
glycol were  similar  to those  of perc over similar  fill times, the 5-s con-
centration values at different points into the fill cycle were normalized  to
the respective maximum concentration values.   A summary of this analysis is
presented  in  Table V-4.  For  the top-fill tests of  perc at the fast (25 gal/
min) fill  rate,  emissions  had reached 91.4% of the  maximum recorded 5-s con-
centration value  at  25% total fill  time (about 30-35 s).  In the case of
ethylene glycol,  however,  only  33%  of  the  concentration maximum had been
reached  at 25% of the fill  time.   The opposite  occurred during  the fast
bottom-fill  tests which showed a much more pronounced concentration increase
for the ethylene glycol than  for perc.   Similar effects were also observed in
the slow fill (11 gal/min) tests.  The bursting of perc bubbles from  the froth
and consequent injection of mist into the upper part  of the  drum  is believed
to  have  caused the quick increase  of concentrations observed throughout perc
top filling tests.  During perc bottom filling, however,  less  froth is created;
fewer bubbles are available to burst causing mist; and consequently, much slower
rates of concentration increases are seen in the early stages of a fill cycle.

          Table V-5 presents  mean  emission rates  and  emission  factors for  39
tests of dry  drum filling.   A comparison of top versus bottom  filling of perc
shows that  top filling produces  from 2  to 4  times the emissions than from
bottom filling.  The emission factor decreases as the drum fill rate  increases.
However, for the bottom filling  of perc, emission factors for  the two faster
fill rates  exceeded  that  of  the slowest fill  rate.  As  indicated in Table
V-5a, the percent of vapor saturation values for top and  bottom fill with perc
agrees favorably  with  values  presented by EPA (1985) for cargo tanks.  How-
ever, the values for larger vessels presented by EPA (1985) and Schrag (1985)
are substantially smaller.

          The  ethylene glycol emission  factors are  believed  to have been  in-
flated because of a higher volatility contaminant  in the liquid.  This un-
identified contaminant is  believed to have been  added  to the original two
55-gal drums  of ethylene glycol through contaminated transfer  hoses.   Maximum
concentrations in the range  of  6,000  to 8,000 ppm v/v were observed even
though the  saturation  vapor concentration of ethylene glycol  is known to be
less than  135 ppm at 25°C.   As  can  be  seen  in Table V-5, Tests 4 and 5 of
ethylene glycol  produced  raw  emission factors of 2.04 g  and 2.15 g,  respec-
tively, while tests 24 and 25 under  similar conditions produced emission fac-
tors of 1.06  g and 1.14 g, respectively.  This reflected  the loss of  volatile
contaminant  as the ethylene glycol was repeatedly handled during the  testing.

                                      59

-------
            30000
         a
         a
            20000
         
-------
   Table V-4.  Rates of Attainment of Maximum Drum Exit Concentration
                                             Percentages of
                                    maximum observed concentrations
                                 at given percents of total fill time3
                                  25%             50%             75%


A.  Top fill at 25 gal/min
      Perchloroethylene       91.4 (3.6)      96.7 (1.7)      95.4 (4.1)
      Ethylene glycol         33.4 (13.1)     57.0 (16.3)     83.7 (12.6)
        Both                  56.6 (30.2)     71.9 (23.2)     88.1 (11.7)

B.  Bottom fill at 25 gal/min
      Perchloroethylene       16.5 (10.8)     46.9 (10.4)     80.6 (7.5)
      Ethylene glycol         52.8 (6.3)      65.6 (9.4)      83.6 (9.8)
        Both                  34.6 (20.2)     56.2 (13.6)     82.1 (8.8)

C.  Top fill at 10 gal/min
      Perchloroethylene       91.9 (2.6)      96.6 (1.0)      97.9 (0.6)
      Ethylene glycol         48.5 (1.5)      73.7 (2.5)      90.0 (2.8)
        Both                  67.1 (21.6)     83.5 (11.5)     93.4 (4.5)

D.  Bottom fill at 10 gal/min
      Perchloroethylene        0.8 (0.3)      13.5 (11.3)     57.4 (15.4)
      Ethylene glycol         42.8 (22.0)     49.5 (9.2)      67.5 (7.0)
        Both                  27.0 (26.8)     36.0 (20.1)     63.7 (12.0)

E.  Top fill at 17 gal/min
      Perchloroethylene       96.2 (0.9)      98.9 (0.1)      99.3 (0.2)

F.  Bottom fill at 17 gal/min
      Perchloroethylene       16.0 (1.0)      44.5 (6.9)      78.9 (3.2)


 25%, 50%, 75% values are from the nearest 5 s mean concentration value.
Values in parentheses are standard deviations.
                                    61

-------
                                     Table V-5.   Drumming Emission Factors1
Chemical name
Perchl oroethyl ene





Ethyl ene glycol







Filling Nominal
mode' "fill
rate (gpm)
Top 11
17
25
Bottom 11
17
25
Top 11

25

Bottom 11

25

No. of
tests
3
2
6
3
2
4
2
2
.2
2b (4)
2
3
2
2
Mean emission
rate (mg/s)
180
238
337
50
121
164
3.4
3.4
16.0.
17. 4b
0.66
0.77
3.0
3.7
Mean emission
factor
(g/drum fill)
53.3
45.6
43.6
13.9
23.5
20.6
1.01
1.11
1.88.
2.10b
0.22
0.24
0.36
0.46
Corrected for
contaminants
(g/drum fill)






0.95
0.96
1.28
1.10
0.18
0.20
0.28
0.34
Vapor
and
mist
X
X
X
X
X
X

X

X

X

X
Vapor






X

X

X

X

a
 Dry drums only.
 Tests 4 and 5 only;  tests 24 and 25 show significantly less emissions.
See text.

-------
               Table V-5a.   Comparison of Saturation Factors
Operation
Present
study
(drums)
Section 4.4 of
AP-42
Cargo. Marine
Tanks0 Vessel sc
Shroy (1985)
Barges
 Top (Splash)
   Fill                    1.20         1.45

 Bottom  (Submerged)
   Fill                    0.46         0.50     0.2 - 0.5     0.06 - 0.23
 ?  Geometric means for perchlorethylene tests (dry drums only).
    Suggested values for loading petroleum liquids into clean cargo tanks.
 c  Suggested values for loading petroleum liquids other than gasoline and
      crude oil.
    Filling of clean barges with acetone and benzene.   Values in table represent
      maximum saturation factor reported for fill  operation.
          A separate  column  in Table V-5 presents ethylene  glycol  emission
factors adjusted  for  the loss of contaminant.  The mean emission factor for
tests 4 and 5, for example, was reduced to 1.10 g, the same value as the mean
emission factor  for tests  24  and  25.   Intermediate test values were also re-
duced based on  the assumption that top filling produced four times the con-
taminant loss of bottom filling.  With this adjustment, test-to-test compari-
sons can be made.

          Top filling emissions for ethylene glycol are shown to be from 3 to
5 times the emissions from bottom filling, a slightly higher ratio than for
the perc tests.   In three of the four sets of tests with a 10 mm nylon cyclone
precollector  (50% cutpoint of 3.5 urn AED), vaporous  emission factors were
slightly lower than comparable factors for vapor plus mist.

     B.    Dispersion Tests

          1.   Experimental Apparatus

          For the  tests  of plume dispersion of emissions  from  the drum, a
steady state  emission rate was  created by  bubbling air through a half-filled
drum.  The bottom fill lance (nozzle) from the  emission  tests,  with wire
screen covering the discharge end, was used to introduce air into the liquid.
A Bell and Gossett 1/2 hp vacuum pump forced air through a 1 gal surge tank
followed by an American Charcoal V-l test meter and into the lance.  The bub-
bling rate of 3.0 cfm corresponded to a drum fill time of 2 min and 23 sec,
which was close to the fast fill rate used in the emission tests.  The airflow
rate was determined from successive dry gas meter readings  separated by a
fixed time interval.
                                      63

-------
           The  wind  tunnel  flow rate  was  adjusted  to  nominal  values  of 100  fpm
 or  200 fpm,  by opening windows  to  reduce the vacuum in  the building which
 housed the tunnel.   A Hastings  warm wire anemometer was  used to  measure air
 velocity  profiles during the dispersion tests.   Before or after each disper-
 sion  test, the anemometer was  used  to check  the  centerline  airflow at 1,  2,
 3,  5,  6,  7 and 8  m  from  the  tunnel inlet.

           The  drum  center was  located at a distance of 4 m from the tunnel
 inlet.  Two  bung positions were used in the testing—upwind (facing the tun-
 nel inlet) and side  (facing  the  tunnel wall).

           A  rectangular  array  of 72  sampling  points  with  a 0.25 m spacing  was
 used  to determine the analyte  concentration  patterns  at distances  of 1 and
 3 m downwind from the center of  the  drum.   The inlet to the  air sampling line
 was held  in  succession at  each  grid  point for 5 to 18 readings of 5-s  concen-
 trations  following a 15-s  equilibration  period.   Background  concentration  was
 determined by  sampling at  an upwind  position  I m  above the floor  and  2  m down-
 stream from  the tunnel inlet.

           The  temperature  of the liquid  was  measured  at  least once  during
 each dispersion test.  For the tests  of  perchloroethylene, the latent  heat of
 the inlet air  (heated by  the  pump)  tended to balance  the cooling effect of
 evaporation  so that the  liquid  temperature remained closed  to the  air tem-
 perature  in  the tunnel.   However,  in the case of  methanol,  the substantially
 greater heat absorbed by evaporation  caused the liquid to cool several  degrees
 centigrade below  the ambient temperature.

           2.    Test Results

           A  summary  of the 18 dispersion tests using perc  and methanol is
 shown  in  Table V-6.   The propose of these tests was  to characterize the emis-
 sion plume shape  as a function of test chemical properties, tunnel flow rate,
 and bung  position.

          An example array of  concentrations  observed at  a downwind  distance
 of 3 m during  the side  bung release  of  methanol  emissions  is shown  in Fig-
 ure V-4.   Corresponding  wind speeds  are given in Figure V-5.  For this test
 (Test  14) with a mean 199 fpm tunnel  centerline flow rate, concentration ranged
 up to  500 ppm  with  an observed  background level of 6 ppm.  The center of the
 plume  was near the  tunnel  centerline  about  0.75 m above the  floor.   A slight
 drift  to  the right  side  (looking downwind) was observed  only at  the  0.75  m
 height; otherwise the horizontal concentration pattern was symmetrically dis-
 tributed  reflecting the  dominant effect of the drum wake on plume dispersion.

          Both perc and,  to a lesser extent, methanol emissions were found  to
 sink to the tunnel floor, as illustrated in Figures V-6 and V-7 which compare
vertical  profiles of methanol  concentration  from side  releases  at  nominal
tunnel flows of  200  fpm  and  100 fpm.  The cross-sections of the plume for  a
100-fpm flow are  shown in  Figures V-8 and V-9, for downwind  distances  of 1 m
and 3  m,  respectively.
                                      64

-------
                                                                Chemical:  Nctlianol
                          , Tunnel Ceiling
                               -
:>„
;.>
^x}
x£
S
^6
Left ^
Tunnel -'^4
Wall >£
Xx3
-:2
.' i
:-;
f
1 meter ,

7

15 7 9
18 70 27
14 149 391 127

9 497 302
11 16 46 225 424 286
10 39 42 260 423 260
13 38 108 217 299 244
215
32 73 210 250 251
1 i i 1 i 1 1
p 1-2 3 4 ; ; 5- --- 6 ;
^ • ^ s~ s _' UUIIK ruaEI &UII I «BftUU
Sampling Plane: 3 met
Test Date: 8/29/85

Concentration (ppm)
.
... |
18 ^ R'9ht
• • • x-^x Tunnel
/> Wall
44 ^x
177 57 14 ^>
164 84 -x^-
196 56 46 '.-'^
s'_
140 27 T-''-'
L i A
> 7 -'--8 9 - .
                           Tunnel Floor
Figure  V-4.   Profile  of methanol  concentrations at 3  m

                sampling plane  (200 fpm).

-------
cn
O1
                                               - 2 ' . '  ' 3
4  ;>j>x^  --6/-'>7

  funnel Floor
                                                                                                     Chemicali  Hcthanul



                                                                                                     Bung Position:  Side



                                                                                                     Sampling Plane:  3 meters



                                                                                                     Test Date:  8/29/85





                                                                                                        Wind speed (m/s)
                                    Figure  M-5.  Profile of  wind speeds at 3 m sampling

                                                     plane for methanol  test at  200 fpm.

-------
            o
            UJ
                                                           3m Downwind
                                 0.4     0.6

                                  C (h)/Cmax
                                (along tunnel
Figure V-6.   Comparison of vertical profiles of methanol concentrations from a drum
               side release with wind speed of 200  fpm.
                                       67

-------
              CD
              ui
                             1m Downwind
                                                          3m Downwind
                           0.2
  0.4     0.6
  C(h)/Cmax
(along tunnel
0.8
Figure  V-7.  Comparison of vertical profiles of methanol concentrations from a  drum
              side release with wind speed of 100 fpm.
                                         68

-------
Left
Tunnel -^ 4
Wall
                                                 Worker Exposures
                                                             •
                                               330           138    14
Right
Tunnel
Wall
                                           Tunnel Floor
  Figure V-8.   Cross-sectional view of methanol  concentrations  (ppm) at 1m downwind for a drum side
                 release at 100 fpm, compared to downwind facing worker exposure (ppm).

-------
                                                    Worker Exposure
                                                        •     •
                                                       168   73
                                                                                  //Right
                                                                                      Tunnel
                                                                                      Wall
                              .<  "4' 
-------
                        Table V-6.  Dispersion Test  Log



Date
8/20/85



8/21/85
8/22/85

8/23/85


8/28/85

8/29/85

8/30/85





Test
no.
1
2
3a
4a
5
6
7b
8b
9
10
11
12
13
14
15
16
17
18


Chemical
name
Perc



Perc
Perc

Perc


Methanol

Methanol

Methanol



Mean
fluid
temp.
(°C)
20
21
22
24
25
23
20
23
22
21
25
19
24
21
27
23
23
20
Nominal Tunnel
centerl ine
flow (fpm)
flow (fpm)
200


100
100
200

100


200

200

100





Bung
position
Upwi nd


Upwi nd
Upwi nd
Side

Side


Upwi nd

Side

Side

Upwi nd

Downwi nd
sampling
distance
(m)d
1
2
3
1
1
1
3
1
1
3
1
3
1
3
1
3
1
3
.Test aborted and pump moved out of tunnel.
 Exit airflow from drum of 2 ftVmin; on all
cairflow was 3 ftVmin.
 As measured downwind from the drum center.
other dispersion tests, exit
          This plume  sinking  behavior  is compatible with the high molecular
weight of perc  (166)  but not with the  molecular  weight of methanol (32).
However, the relatively high rate of vaporization of methanol resulted  in the
cooling  of  the emission  stream  released from the drum, which  produced a
slightly negative plume buoyancy.  As evidence of the cooling, the lower half
of the drum exterior was observed to collect condensation as methanol  testing
proceeded.   This effect was not observed with either perc or ethylene  glycol.
With about 2 gal of methanol vaporized during a typical  evening of tests, the
total vaporation heat  loss  approximated 1,400 kilocalories.   This  loss was
somewhat offset by  the latent heat of  the  air supply (29° to 35°C) to the
bottom of the test drum in relation to the tunnel  air temperature.   The lowest
methanol temperature observed during the dispersion and exposure experiments
was at the end of Test  12 where  17°C was measured.  The corresponding tunnel
air temperature was 22°C.
                                      71

-------
           Concentration  and wind speed values  measured  at the 3 m distance
 were used to calculate mass flux values for comparison with the steady state
 emission rates.   These  data are presented  in Appendix D which  discusses
 quality control  data.

           Appendix  E  contains plotted concentrations of perc and methanol at
 1  m  and 3 m  downwind  of  the drum  center.   In  addition, for each  concentration
 distribution,  there is an  accompanying  plot of  array wind  speeds.

      C.    Exposure  Tests

           1.   Experimental  Apparatus

           For  the tests  of worker exposure,  a steady state emission rate was
 created in exactly  the same manner as for  the dispersion tests described  above.
 Moreover,  the  drum was located at the  same  position,  and  the same two bung
 positions were used.

          Unlike  the  dispersion tests,  concentrations  for  the exposure tests
 were  measured  only at a breathing height  of  1.57 m for a standard worker
 height  of 1.73 m.   Two arrays of  sampling  points  (worker positions) were  uti-
 lized,  one corresponding to the upwind  bung position and the other  to  the side
 bung  position.  These are  illustrated in Figures  V-10 and V-ll.

          The  two bung positions, upwind  and  side,  simulated two airflow di-
 rection  scenarios.  The  upwind bung position results  when the wind flow is
 directed  toward  the worker's back at  the  drumming station, and the side bung
 position  results  when the wind flow is  across the drumming operation from the
 s i de.

          2.  Test  Results

          The  14  exposure  tests are summarized  in Table  V-7.   Exposure tests
 differed  from  dispersion tests  in that  the sampling line inlet was  placed in
 the breathing zone  of a "worker."  In Figure V-12, the worker positioned him-
 self  near the drum  with the wind  flow directed to his back (upwind bung posi-
 tion).   In  this  photo,  the air hose  used  to  supply fresh  air to  the worker
 during periods of testing is not  shown.

          The  relatively constant emission rate from the drum enabled  steady
 state exposure tests to be  efficiently done.   However,  it is necessary to re-
 late  these  results  to the   nonsteady state emissions tests  described in Sec-
 tion A.   A  summary  of emission rates used both in tests of dispersion and
 exposure is given in Table V-8.

          The emissions  from bubbling air  through perc at  25°C produced con-
 centrations of about  35,000 ppm at  the bung hole, which corresponds closely
 to the  maximum concentrations observed  in  top filling at the fast  fill  rate.
 Since top  filling produced  fairly uniform  concentrations (~  30,000  to  37,000
ppm)  over the  latter  75% of the drum  fill  cycle,  the concentrations measured
 in the  perc  exposure  tests represent well the concentrations expected to be
 inhaled by a worker.   For  the remaining 25%  of the top filling  cycle, for
bottom  filling, and for  ethylene  glycol tests,  it is necessary to apportion
observed  exposure concentration  values by referring to  emission rates in
Tables V-3 and V-5.

                                      72

-------
                                                      Wind
                                                      Flow
50 Centimeters
  Figure V-10.   Sampling locations  for upwind bung position.
                             73

-------
                     oo
    Wind
    Flow
                 e   •    •
                                             -l	1-
                                                 1 Meter
Figure V-ll.   Sampling locations  for side  bung  position.
                           74

-------
Figure V-12.   Example exposure testing.



                  75

-------
                           Table  V-7.   Exposure Test  Log


Date
8/21/85


8/23/85
8/29/85


9/2/85


9/13/85






Date
8/13/85
9/2/85


9/13/85


Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table V-8.


Time
1503
1813
1820
2336
1830
1834


Chemical name
Perc


Perc
Methanol


Methanol


Ethylene glycol



Drum Emissions for


Chemical
Perc
Methanol
Methanol
Methanol
Ethylene glycol .
Ethylene glycol

Fluid
temp (°C)
25
24
23
21
26
24
20
28
23
21
21
21
21
21
Tests of Di
Fluid
temp.
(°C)
25C
28
28
20
21
21
Nominal tunnel
centerline

Bung
flow (fpm) positions
100
200
200
100
200
200
200
100
100
200
200
200
100
100
spersion and Exposure
Mean 5-s
concentration (ppm)
Raw Corrected
34534 34819
126636 120937
117686 120628
100958 101968
1381 1519
13626 1505
Upwind
Upwi nd
Side
Side
Side
Upwi nd
Upwi nd
Upwi nd
Side
Upwi nd
Upwi nd
Side
Side
Upwi nd

Emission
rate0
(mg/s)
326
216
216
187
5.40
5.35
.Based on checks of the Beckman THC analyzer with bag standards.
 Based on 3.0 ft3/min air supply to drum bottom.
 .Estimated value.
 Believed to be contaminated with volatiles.
ewith cyclone on sampler inlet.
                                        76

-------
          Special studies were conducted within the 14 exposure tests to mea-
sure the effect  on  exposure of:   (a) worker orientation  to  wind flow; (b)
worker movement; and (c) breathing height.   Table V-9 presents mean 5 s chem-
ical concentrations observed in studies conducted in an upwind bung situation
with the "worker" always facing downwind.  By moving only slightly while fill-
ing a drum, the  worker  can  lower  his exposure  by a  factor of  from 3  to 45  as
illustrated in Table V-10 for methanol.  This  effect  results  from the mixing
of the contaminant stream with clean background air.

          As indicated  in Figures V-13 and V-14, worker height between 1.4
m (5 ft 2 in.) and  1.5  m (5 ft 8 in.)  appears to have little effect on perc
exposure near the drum,  but the exposure is sensitive to wind speed.   At dis-
tance of about  0.5  m  from the bung, the effect of  breathing  height  is more
pronounced.   The effect of  worker distance and wind speed on methanol expo-
sure is shown in Figure V-15.

          The major conclusion from the  studies of  breathing  zone concentra-
tions is that worker orientation  to the wind flow  is  a  significant parameter
in predicting worker exposure.   Much  higher analyte concentrations were ob-
served in the breathing zone when the worker  faced downwind than when the
worker faced upwind or  faced crosswind.  This  unexpected  test result is dis-
cussed further in the next section on model development.

          For purposes  of exposure model  development  to be  discussed in
Section VI, isopleths of normalized breathing  zone  concentrations for worker
positions near  the drum during  filling  operations  have been prepared.  A
total of eight  figures  (Figures  V-16 through V-23)  are  provided, correspond-
ing  to  two  simulants  (perchloroethylene and methanol),  two  airspeeds (100
and 200 fpm), and two worker orientations  (facing  downwind and  facing either
upwind or crosswind).

          The units for the  normalized isopleth values  are seconds per cubic
meter (s/m3).   The value 0.1 s/m3 is used to represent the edge of the plume.
If these values  are multiplied by the  emission rate,  then the isopleths rep-
resent lines of constant concentration.  For example, if the emission rate is
5 mg/s, then the curves  shown in Figure V-16 correspond to 0.5, 5, and 50 mg/m3.

          All  airflows  are  from  the top of  the  page to the bottom.   It is
recommended that the  100-fpm  curves  be used  for  most indoor operations.

          As a  rule of  thumb,  it is suggested that the methanol curves be
used for drumming of liquids with molecular weights of less than 100 and that
the  perchloroethylene  results  be used for  larger  molecular  weights.  This
recommendation  is based on  the observed settling of denser vapors away from
the breathing zone.

          Finally, it is important to note that worker motion may often be as
important as orientation in determining exposure.   The results presented  in
these graphs are based  on motionless workers and thus represent highest poten-
tial exposure.   As noted in  the previous section, slight motion of the worker
may  reduce  observed concentrations  by as  much as  an order  of magnitude.
                                      77

-------
                                      Table V-9.  Special Exposure Studies
00
Date
8/21/85



8/29/85
8/29/85

9/2/85

Time
2055
2120
2126
2135
2316
2318
2319
2039
2040
Tests were of upwi
20 cm directly upwi
Chemical
Perc
Perc
Perc
Perc
Methanol
Methanol
Methanol
Methanol
Methanol
Temp.
24
24
23
23
20
20
20
~ 24
~ 24
nd bung emissions at
nd from bung opening.
Mean 5-s Situation
concentration (ppm)
Raw
695
683
720
128
5209
382
1859
4147
90
mean tunnel
Corrected
806
792
835
148
5339
392
1905
4396
95
centerl ine

Facing downwind and standing @ 0, 20a
Facing downwind and standing @ 0, 20a
Facing downwind and sitting @ 0, 20a
Facing downwind with nostrils at
drum level @ 0, 20
Facing downwind @ 0, 30 and stationary
Moving between -40, 30, and +40, 30
with large steps while facing downwind
Moving between -40, 30, and +40, 30
with small steps while facing downwind
Facing downwind and standing @ 0, 30C
Moving between -40, 30C and +40, 30
flow of 209 fpm. Position 0, 20 refers to
     Tests were of upwind bung emissions at mean tunnel  centerl ine flow of 188 fpm.
    £o 30 cm directly upwind of drum and ±40 cm to side.
     Tests were of upwind bung emissions at mean tunnel  centerl ine flow of 104 fpm.
    to 30 cm directly upwind of drum and -40 cm to side.
Position ±40, 30 refer

Position -40, 30 refers

-------
   Table V-10.   Effect of Worker Movement on Exposure
                (30 cm upwind,  methanol)
Worker movement
(side-to-side)
Stationary
Slow
Exposure
Wind speed: 100 fpm
7,980
-
(ppm)
200 fpm
5,410
1,740
Rapid
  Continuous movement                    90
  Large steps                             -         380
                            79

-------
                   2
            •1.5m—•
            • 1.4m— •
                   2
Breathing
Height
         Wind Flow
         200 fpm
22   330   480    690
15    75    90    720
                   2     2     2     2    140
            •1.0m—•     •     •     •     •
                          Release
                          Point
                                                                55 gal Drum
                                                  Perchlorethylene
                  100    80    60    40     20     0
                         Dlstahce.UpwInd from Source .(cm)
      Figure  V-13.   Effect of breathing height and upwind
                        distance-on exposure  to perc
                        concentrations  (ppm).
                                    80

-------
                                                    T800
100
80        60        40        20
 Distance Upwind from  Drum (cm)
 Figure  V-14.  Effect of breathing height, wind speed and
                upwind distance on exposures to perc
                drumming emissions.
                            81

-------
                                                          -I 8000
                 60,                  30
            Distance Upwind from Drum (cm)
Figure-V-15.
Effect-of wind  speed and upwind
-------
Figure V-16.
K contours  for methanol  (M<100), 100 fpm air velocity and  downwind
  worker orientation.
                                      83

-------
Figure-V-l-7.  K-contours-for methanol (M<100), 200 fpm air velocity  and  downwind
                worker orientation.
                                        84

-------
                                                                             A* r-
                                                                             F/o*/
Figure V-18.   K contours for methanol  (M<100), 100 fpm air velocity and non-downwind
                worker orientation.
                                         85

-------
                                         'fir
Figure V-19.
K contours for methanol (M<100), 200 fpm air velocity and non-downwind
  worker orientation.
                            86

-------
Figure V-20.   K contours  for perchlorethylene (Mi 100), 100 fpm air velocity and
                downwind  worker orientation.
                                      87

-------
                                                                         A
                                                                          s n €.
                                                           2.OO fpm  Air
                                                                            Air
                                                                            f?/o*/
Figure V-21.   K-'contours for- perch!orethylene  (M£100),  200 fpm-air velocity and
                 downwind worker orientation.
                                         88

-------
                                                             / /<
                                                               
-------
I	//7J
                                                                           /7
Ai
                                                                       ir
                                                                             /4/y
                                                                             /e/c»«v
Figure \l-23.   K contours for perch!orethylene (Mi 100), 200 fpm air velocity and
                 non'-downwind- worker  orientation.
                                          90

-------
VI.  MODEL DEVELOPMENT
     A.   Modeling Alternatives

          Several  models  have been proposed  for  the  purpose of assessing
worker  exposure  to bagging and drumming operations.   These have ranged from
models which are very simple to apply to those which  are not only complicated
but also restricted to the actual ventilation pattern  considered.   Most models
currently used  to  estimate workplace concentrations are based on the concept
of mass  balance  (Figure  VI-1)  possibly with additional  features such  as in-
complete mixing or time-dependent release rates.

          For the  drumming of liquids  at room temperatures, the  following
equation has been suggested for use (Office  of Toxic  Substances  1984):

                              r  _ 0.6  f V P° r
                              Cv	T	                       Ec»-  I
where:    GV = concentration, ppm

           f = saturation factor for loading,  dimensionless
           k = mixing factor, dimensionless

          P° = vapor pressure, atm.

           Q = ventilation rate, ftVmin

           r = loading rate, drums/hr

           V = volume of container,  cm3


          The values used  in the above  formula are taken from Table VI-1,
unless other values are known.   The  values in  the  table  are representative of
worst case and typical  case.   Equation 1 assumes  ideal  gas behavior and the
absence of chemical reaction.

          The worst case represents  a situation of general ventilation at low
rates in the workplace.  The typical case takes into account  higher ventila-
tion rates and more complete mixing.   The worst case  is  assumed when no infor-
mation is known about the workplace  controls (Office  of  Toxic Substances 1984).

          However,  the primary  limitations of  any mass  balance model result
from the neglect of  spatial  variation of concentration  and the influence of
worker position,  orientation, and motion on exposure.  It is important to note
that even though other  models  take  spatial variation into account, the fact
that a worker presents a substantial obstruction to the  generally  low venti-
lation flows requires that  worker location and orientation be considered in
                                     91

-------
                       BOX MODEL
              Room Volume
                  V
Volumetric
Air Flow
                               Mass Emission
                               Rate
                                Source
                         ^r =  G-CQ
                         at.
               Figure VI-1. The box (uniform mixing) model
                                92

-------
                  Table VI-1.  Default Values for Equation
Term
V
f
r
k
QC
Drums
2 x 105 cc (55 gal)
Worst
case
1.0
30/h
0.1
500 cfm
Typical
case
1.0
20/h
0.5
3000 cfm
Tank
1.9 x 107
Worst
case
1.0
2/h
0.1
3000 cfm
trucks
cc (5000 gal)
Typical
case
1.0
2/h
0.5
d
Tank
7.6 x 107 cc
Worst
case
1.0
1/h
0.1
3000 cfm
cars
(20,000 gal)
Typical
case
1.0
1/h
0.5
d
 Adopted from "A Manual for the Preparation of Engineering Assessments." a
 manual prepared by the Chemical Engineering Branch, Economics and Technology
 Division, Office of Toxic Substances, EPA, Washington, D.C.   20460.   August
b!984.
 Cans and bottles are assumed to be similar to drums except for the volume.
cUse 19,000 cm3 (5 gal) for volume of a can.
 If Q is known for a particular scenario, substitute it into the formulae
 .above.
 Typical operations take place outdoors.   Berman (1982) suggests Q = 23,400 V,
 where V is wind speed in mph.
assessing inhalation exposure.   Because it is almost impossible to standardize
worker motions, the model presented below yields exposure estimates correspond-
ing to an absolutely motionless worker.  As noted earlier, slight motions may
reduce breathing height concentrations substantially.

     B.    Modeling Approach

          The modeling approach adopted in this study attempts to incorporate
the relative simplicity of the mass balance models together with the features
of spatial  variation  and worker position necessary to provide more reliable
estimates of exposure.  Results of the exposure tests described in Section V.C
have been combined  in  the K contours given in  that section.  These  isopleths
(in units of s/m3) correspond to breathing zone concentrations resulting from
a unit emission rate  of  1 mg/s.  Thus, when the values are multiplied by the
emission rate  (in mg/s)  for  a particular application, the isopleths become
lines of constant concentration (in mg/m3) for that application.
                                      93

-------
The  basic steps in the modeling procedure are as follows:

1.   Estimate mass emission rate, G (mass/time),  by


                No.  of drums x loss per drum

                    Averaging time period



2.   Estimate ventilation rate,  Q (air volume/time).

3.   Apply  the  uniform  mixing  model  (Figure VI-1) to obtain the
     steady  state concentration, C   = G/Q.  This value represents
     the  background concentration-to which is added the plume con-
     tribution  due  to filling.

4.   Select  the appropriate set  of K contours,  using the criteria
     below:

     Molecular  weight,  M - Use methanol contours if M is less than
          100 and perchloroethylene results otherwise.

     Airflow rate - The use of the 100-fpm. contours is  recommended
          for most  indoor  applications.   The  200-fpm results may be
          used  for  outdoor operations or-in cases  of indoor filling,
          with  ambient  winds aiding  ventilation.(as might  be the
          case  during the  summer months).

5.   Conduct a  time-motion study of the worker noting time spent in
     each location  (as well  as  orientation to drum) during fill cycle.
     Average over several  cycles.

6.   Using the  appropriate K contours, develop a weighted K factor,
     K, by
                         n
                        i-i *i K'
                 K=    Ji-^-

                          i-i'  V
     where:     t.. = time spent in i-th location

               K.J = K factor for specific orientation in i-th

                      location
                            94

-------
           7.   Determine  the  plume contribution resulting  from  drumming or
               bagging  by multiplying K (in s/m3)  by G (in mg/s).   Call this
               result C ..
                       d

           8.   Add  C    to C .  to obtain the  total estimate  of  breathing zone
               concentration C*.

           9.   Multiple  C* by  1.25  m3/h to estimate  inhalation exposure.

           Because PMN assessments  are often  made before any drumming  line has
been  designed,  default  values for K  have  been provided for typical  filling
operations  in  Table VI-2.   The  modeling procedure  using these  default values
follows  the same steps  as above, except that steps  4 through 6 are no longer
needed.
                        Table VI-2.  Default K Values

Filling operation3
Automated

Manual


Airflow
Side
Rear
Side
Rear
K factors (s/m3)c
M < 100d M £ 100d
0.2 0.1
0.6 0.6
0.3 0.1
3 3
a.,, . .
• All indoors.
 lt«^IU\*IVW WW YV\J I I\C I .
 As noted in the text, the resulting breathing zone estimates should be
 considered as "worst case" in that no worker motion is considered.  See
ddiscussion in text.
 M represents molecular weight.
                                      95

-------
           The default values were  developed  using time-motion  data  gathered
 during  this  program.   These data are given below:
                                Percentage of time at station
                                  Automated          Manual
                Station             fill               fin

           Drum  fill                   5                 25

           Drum  close                25                 30

           Drum  setup/label           65                 40


           Five  percent  of  total  time spent  away  from  drum
           line.


           To  derive  the default  values,  typical  work  station  locations  for
the two  types of fills were placed on the K contour plots  given  earlier.   In
addition,  two fill line  orientations  were  considered—airflow parallel to the
line (i.e., worker not facing downwind) and  airflow perpendicular to the line
(i.e., flow  from behind the worker).   The default values for R were then ob-
tained for the  four cases  using  the  procedure presented earlier as steps  5
through 7.

          As  an  example  of  the  use of the default R  values, consider a hypo-
thetical  drumming line which fills 290 drums with a liquid of molecular weiqht
of 200 over an 8-h period.
               1.  If an emission factor of 0.3 g/drum (mean value for ethyl-
          ene glycol) is used, then the average emission rate over the 8 h is
          given by

                    r _ 290 drums   n 0  , .      _   ,
                    G -- g-^ - x 0.3 g/drum = 3 mg/s


          Step 2.  Current ventilation plans call for Q = 12,000 cfm (5.7 m3/s)

          Step 3.  Using the uniform mixing model,


                         Css • £r%k - "I MWV

          Steps 4 to 6.   The use of an automated fill line with airflow from
          behind the  worker is  planned.   The appropriate  K  value (from
          Table VI-2) is 0.6 s/m3.

          Step 7.   The additional contribution due to drumming  is

                         Cd = KG = 1.8 mg/m3

               8.   The  estimate  of total  breathing  zone concentration is

                         C* = Css + Cd =  2.3 mg/m3

                                      96

-------
      C.  Model Validation

          Validation  of the proposed modeling technique  requires a complete
 set  of  independent data obtained  from a plant-scale bagging or drumming oper-
 ation.  Because no such set  is  found in the  literature, the plant-scale test-
 ing  reported  earlier in this document was done  in order  to provide data  for
 model validation.  Both bagging and drumming operations were tested as speci-
 fied by the scope of work, with the idea that emission factors for these  types
 of operations  could  be developed  independently.   Since the literature search
 yielded few (if any) appropriate  emission factors for  these operations, it be-
 came necessary to develop a  limited set of experimental values for this study.
 Because of  the need  to limit the  cost  of  this  phase  of the experimental  pro-
 gram, it  was  decided to develop  emission factors for  the  drumming operation
 only.

          1.  Plant A  Estimates

          Data  from  the drumming operation  at  plant A are suitable  for  a
 limited validation of  the modeling approach.   Note that data from run A-l are
 not  suitable because all sample masses were below the 2 ug limit of detection.

          No reliable  estimate  for  the vapor pressure  for fluid 2 could  be
 provided.   As  a result, it  was  not possible  to use the vapor pressure to  es-
 timate  an emission  level for run A-2.   Because of the low volatility of the
 fluid,  the mean emission factor measured for ethylene glycol (0.3 g/drum) was
 used in the validation  process.

          Using this  emission  factor,  the emission rate during test A-2 was
 found as

               G = 59 drums x 0.3 g/drum -=• 98 min = 3 mg/s

 As stated in  Section IV-A,  the  worker_spent only 3% of the time in front of
 the  fill station.   Thus, the weighted K value is

               K = 0.97 x (0 s/m3) + 0.03 x (1 s/m3)  = 0.03 s/m3

 Note  that a multiplier of 1 s/m3  (rather than 10 s/m3) has been used to ac-
 count for the  fact that only about 3 s per cycle was spent at the fill loca-
 tion, and thus no strong buildup  of vapors was possible.   Also, because this
 discussion is  meant to partially validate the model  using plant data,  a value
 of 0 s/m3 is  used  rather than  0.1 s/m3 for  locations  other than the  fill
 station.

          The  plume  contribution  resulting  from drumming  is then  given  by

                    Cd = KG = 0.03 s/m3  x 3 mg/s  = 90 ug/m3

As noted in  the text, an operator  exposure value  of  240 ug/m3  was measured by
a personnel  sampler.    Because samplers  15a through 17 in  Table IV-5  can be
considered as  measuring upwind concentrations,  it is  seen that the  background
 level is about 110 to 180 ug/m3.   Thus,  the  estimated drumming contribution
of 90 ug/m3  compares  very favorably  with the  measured value of 60 to  130  ug/m3.


                                      97

-------
           The background  estimate  for test A-2 using the modeling procedure
 would be

                          Css = G/Q

                              = 3 mg/s -r 11,750 cfm

                              = 540 ug/m3

 which is approximately 3  to 5 times higher than observed.  Note that results
 from samplers 3 through  7 suggest some stratification of concentration with
 height,  presumably because of  the  relatively high density of the hydraulic
 fluid vapors.   As  a result,  use of the uniform mixing model  to  estimate back-
 ground levels may  be  conservatively high when applied to  the  drumming  of
 heavier  liquids.

           2.   Plant B  Estimates

           A  previously determined emission factor  of  0.095  kg/Mg  (0.19 Ib/ton)
 for  bagging  of urea (USEPA,  1981) was  used to estimate the  uncontrolled emis-
 sion rate at plant B.   Furthermore,  a collection  efficiency of 80% was  as-
 signed to the local hood,  resulting in an uncontrolled  emission  factor of
 0.019 kg/Mg  (0.03  Ib/ton).  Finally, based on  particle size  ratios measured
 by  both  the cascade impactor  and cassette samplers,  14% of  the  controlled
 emissions  was  assumed  to  consist of respirable  particulate  (^ 3.5  urn in aero-
 dynamic  diameter).

           The  following estimated emission rates,  G,  are  found  using the data
 of Table-IV-7:
                                      Controlled
                                 emission  rate  (mg/s)
Run
B-l
B-2
B-3
Suspended
particulate
20
18
30
Respirable
particulate
2.8
2.6
4.2
          Time and motion  studies conducted at the plant indicate that roughly
60% of  the  operator's time is spent at the autopacker.   Thus,  the weighted K
value is

                K = 0.4 x  (0  s/m3)  + 0.6 x (1 s/m3) = 0.6 s/m3

The multipliers  (0  and 1  s/m3) are the  same  as  those used in the drumming
estimates.
                                      98

-------
          The predicted plume  contribution  arising from bagging is given by
KG.   For the three runs,

Run
B-l
B-2
B-3
The background concentration
model are given by G/Q, where
Predicted plume
contribution (ug/m3)
Suspended Respirable
particulate particulate
12,000 1,700
11,000 1,600
18,000 2,500
estimates calculated from the uniform mixing
Q = 22,900 cfm:
Predicted background (ug/m3)
Measured background
from Samplers 6, 7a
Run
B-l
B-2
B-3
Suspended Respirable
particulate particulate
1,800 260
1,700 230
2,800 390
concentrations were found by averaging the concentrations
, 7b, and 8:
Run
B-l
B-2
B-3
Measured background Cug/m3!
Suspended Respirable
particulate particulate
1,200 110
1,600 110
3,000 450
The background values  predicted  by the uniform mixing model agree well with
the measured values for the two particle size fractions.
                                      99

-------
          To assess  the  model's  performance  in predicting the plume contribu-
tion to worker exposure,  only  respirable particulate concentrations are avail-
able.  Respirable  particulate concentrations of 1,890 and  1,200 ug/m3 were
measured with  the  operator personal samplers  for  Runs B-l  and  B-2,  respec-
tively.  (No value was reported  for Test B-3.)

          The plume  contributions of 1,700  and 1,600 ug/m3 predicted by the
modeling technique  compare well  with  the measured  contributions of 1,800 and
1,100 ug/m3, after subtracting measured background concentrations.
                                      100

-------
 REFERENCES
Berman  DW.   1982.  Methods for  estimating  workplace  exposure to PMN sub-
stances.   Clement Associates,  Inc., Arlington,  VA.   Prepared  for  U.S.
Environmental  Protection  Agency,  Economics  and  Technology  Division,
Washington, DC.

Bomberger DC,  Boughton  R,  Endlich R,  Fowler D, Gikis B, Ludwig F, Wilhelm D,
Witham C.  1983.   Industrial process profiles to support PMN review:   filling
of  drums  and  bags,  draft.   SRI International, Menlo Park,  CA.  Prepared  for
U.S. Environmental Protection Agency, Office of Toxic Substances, Washington,
DC.  Contract No. 68-01-6016, Technical  Directive 69.

Clement  Associates.   1981.  Mathematical models  for estimating workplace
concentration  levels:   a  literature review.   Clement  Associates,  Inc.,
Washington, DC.  Prepared for U.S. Environmental  Protection Agency,  Economics
and Technology Division, Washington, DC.

Davis RJ.  1971.  A simple  model  for  the estimation of aerosol concentration
in a closed vessel.  Am Ind Hyg Assoc J 32:603-609.

USEPA.   1984.   A manual  for the preparation  of engineering  assessments.   Pre-
pared by  the Chemical Engineering  Branch, Economics and Technology Division,
Office of Toxic Substances, Washington,  DC.

Schroy JM.  1981.  Prediction  of workplace  contaminant  levels,  pp.  190-206.
Symposium Proceedings  on Control  Technology  in the  Plastics and  Resins Indus-
try, Atlanta, GA, February 27-28.   Cincinnati,  OH:   National  Institute  for
Occupational  Health  and  Safety.   NIOSH Publication  NO. 81-107.

Schroy JM.  1985.  Dynamic  conditions in filling drums.    Presented at the
1985 Summer National Meeting of the American Institute of Chemical Engineers
Seattle, WA.

USEPA.   1981.   Urea manufacturing  industry  -  technical  document.  EPA-450/
3-81-001.  Office of Air Quality Planning  and Standards,  Research Triangle
Park,  NC.

USEPA.   1985.   Compilation  of  air pollutant  emission  factors (AP-42).  Vol  I,
Section 4.3.   Research Triangle Park,  NC.
                                      101

-------
                                 APPENDIX A

                               INDUSTRY SURVEY


                              TABLE OF CONTENTS

                                                                         Page

I.        Introduction	     A-l

II.       Industry Survey	     A-l

               A.   Information Search Methodology 	     A-l
               B.   Significant Documents Identified 	     A-2
               C.   Industry Characterization Summary	    A-12

III.       Arrangements for Plant Visits 	    A-13

               A.   Industry Contacts	    A-17
               B.   Obstacles Encountered in Plant Visit
                     Arrangements 	    A-17

IV.       References	    A-19

V.        Annotated Bibliography of References on File	    A-20

VI.       Packaging Association Phone Survey Approach and Results .  .    A-39

-------
 I.   INTRODUCTION

           Under the Toxic Substances Control Act of 1976 (TSCA), a manufac-
 turer  or  processor must  submit to the U.S.  Environmental Protection Agency
 (EPA)  a premanufacturing notice  (PMN) prior to commercial production.  The
 EPA  has 90 days to review the PMN and to  take any necessary regulatory actions
 deemed necessary.  During this review, the  potential workplace exposure asso-
 ciated with the manufacture and  processing  of the new chemical must be esti-
 mated.

           This project seeks to  develop and validate predictive models which
 may  be used in the estimation of inhalation exposure of workers involved  in
 the  drumming of liquids  and the  bag or drum filling of solids.  The study de-
 sign is comprised of four tasks, the first  of which is an industry survey to
 characterize drumming and bagging operations and to select the chemicals and
 facilities  for subsequent study.

           This appendix  provides results  of the literature-based industry
 survey and  efforts to identify and gain access to plant sites suitable for
 testing.   Two related items are  included  here:  an annotated bibliography of
 the  references on file (Section  V) and information on a packaging association
 survey (Section VI).


 II.  INDUSTRY SURVEY

          The industry survey involved literature search and literature re-
 view activities as well  as personal  contacts with agencies,  companies, and
 associations.

     A.   Information Search Methodology

          Computerized literature searches  of National  Technical  Information
 Service (NTIS), Chemical  Abstracts,  and a few other data bases were conducted.
 Hard copies of all promising references were obtained and evaluated.   See
 Section V for an annotated bibliography of  all references on file.

          A manual review of the Applied Science and Technology Index for 1981
 and  succeeding years  was  also conducted to  check for relevant data in trade
 publications.  All promising references were obtained and evaluated,  and those
 found  to be useful were placed on file.

          Contacts with knowledgeable persons and groups were conducted by
phone and,  in a few cases,  in person.   Individuals  and  groups contacted in-
cluded Thomas Cooper  and  others  at NIOSH;  Jon Volkwein  at the Bureau  of Mines;
Garrity Baker,  Thomas  Goldberg,  and  Doug Fratz at the Chemical  Manufacturers
Association, the Synthetic  Organic Chemical  Manufacturers Association, and
the Chemical Specialties  Manufacturers  Association,  respectively;  and repre-
sentatives of 13 packaging  associations.   Section VI describes  the packaging
association phone  survey  approach and results.
                                     A-l

-------
      B.   Significant Documents Identified

           Only a few publications which discuss or quantify emission  levels
and worker exposure in drumming and bagging activities were located.  These
publications  were products of federally funded research  for the EPA,  NIOSH,
and the  Bureau of Mines.

           1.   EPA Studies

           Several  useful  EPA reports have been obtained  and reviewed.  These
are described below.

           Bomberger et al.  (1983) discuss the  filling technology commonly em-
ployed in  both the bagging and drumming of solids  and the drumming of liquids.
Typical  equipment and modes of operation are described in detail (i.e., bag
and drum designs,  container closure methods, conveying and filling equipment
configurations,  and emission control  systems).   However, these authors empha-
size  that  it  is  difficult to make general  statements on  industry-wide packag-
ing practices by chemical  manufacturers,  especially emission control practices.

           According to Bomberger, equipment purchased for the packaging of
industrial  chemicals  represents a minor fraction of the  packaging equipment
market.  The  bulk of  this  machinery is  designed and sold for use with foods,
cosmetics,  and Pharmaceuticals.   Hence,  the original design of such equipment
usually  does  not include  toxic emission controls.  The frequency and effec-
tiveness of on-site machinery modification to  achieve dust or volatile emis-
sion  control  by  chemical  manufacturers  are,  therefore, dependent on in-plant
recognition of potential  industrial  hygiene problems and on the ability of
plant personnel  to install  and maintain emission control devices.

           Bomberger also  discusses  estimated emission rates and the resulting
workplace  air concentrations during bag and drum filling of chemicals.  Bag
filling  emissions  have not  been well  characterized to date.   Existing studies
quantify workplace concentrations or  emission  rates at individual  locations
for a very  limited group  of materials,  and only certain parameters (and not
of a proprietary nature)  are reported.   Consequently, it is difficult to
determine  how representative the workplace concentrations and emission rates
which appear  in  the literature may  be,  relative to the wide range  of packaging
environments  which occur  in the chemical  manufacturing industry.   Bomberger
concludes  that there  are  insufficient data to  expand or revise the estimates
in the previous  EPA literature which  assume  levels of 2 mg/m3  around typical
semi-automatic filling stations  without  sophisticated dust control  mechanisms
in place and  30  mg/m3  around very dirty  operations—both of which  also assume
that 30% of the  airborne material  is  respirable  (Berman 1983).

          Bomberger's  discussion of volatile emissions from the drumming of
liquids  is based  largely on data from the  petroleum refining industry because
no other data  were  available.   Table A-l  provides estimated emission rates for
drum filling  of  a  light liquid with a relatively high vapor pressure (benzene)
and a heavy liquid  with a  relatively  low  vapor  pressure (cyclohexanol) under
                                     A-2

-------
   Table A-l.  Estimated Drumming Emissions From Light and Heavy Liquids'
Material
                               Emission rate to workplace environment (q/h)
  High1
Moderate
  Low"
Light liquid (benzene)
  drum vent
  fitting6        f
  chronic wet spot

       Total
2500 (85%)
 390 (13%)
  34 ( 2%)

2924
125 (41%)
144 (47%)
 34 (12%)

303
 42 (19%)
144 (65%)
 34 (16%)

220
Heavy liquid (cyclohexanol)
drum vent
fittings
chronic wet spot^
Total
33 (35%)
60 (64%)
1 ( 1%)
94
2 ( 4%)
51 (94%)
1 ( 2%)
54
1 ( 2%)
51 (96%)
1 ( 2%)
53
b$ource:  Bomberger et al.  1983.
 Top filling with no collection of vapors.  No routine maintenance of
 fittings, which are not inside a hood.  200-cm2 wet spot, which is not
 inside a hood.
 Top filling with 95% collection of vapors.   Routine maintenance of
 fittings, which are not inside a hood.  200-cm2 wet spot, which is not
 .inside a hood.
 Bottom filling of drum with 95% collection of vapors.   Routine main-
 tenance of fittings, which are not inside a hood.   200 cm2 wet spot,
 which is not inside a hood.
fThree valves, one pump, two rotating or swivel fittings.
 A mass transport coefficient of k = 0.12 representing quiescent
 evaporation was used.
9A mass transport coefficient of k = 0.10 representing quiescent
 evaporation was used.
                                     A-3

-------
three  sets  of filling conditions:   a high rate of emissions  (top  filling with
no  collection of vapors and no routine maintenance of  fittings which are not
inside a  hood),  a moderate rate of emissions  (top filling with 95% collection
of  vapors),  and  a low rate of emissions (bottom filling  with 95%  collection
of  vapors).   The range of estimated emission  rates was 220 to 2,924 g/h for
the light liquid and 53 to 94 g/h  for the heavy liquid.

          Although Bomberger discusses methods for modeling  workplace concen-
trations, this topic is covered more thoroughly by Clement Associates (1981).
The models  available for modeling  contaminant concentrations in an industrial
environment may  be grouped into three types based on their theoretical basis:
mass balance,  Gaussian diffusion,  and aerosol  mechanics.

          Mass balance models are  based on mathematical  accounting schemes
whereby all  fractions of the material  passing through the process under study
are quantified,  and the total outputs are always  equal to the total inputs.
These  approaches are most applicable for gaseous  emissions because vapors fol-
low air currents freely and are not influenced by gravity.   Particulate emis-
sions  consisting of fine particles—smaller than  10 to 20 urn in diameter ac-
cording to  Clement Associates (1981) and generally smaller than 5 urn in diam-
eter according to Bomberger et al.  (1983)--may also be modeled effectively
using  mass  balance equations.

          Gaussian dispersion models are not  commonly employed in an indoor
setting.  However,  Clement Associates  (1981)  offer a notable  example, Schroy
(1979), demonstrating the feasibility of such  usage.  Such models are based
on  the  Gaussian  diffusion of particles  as they move away from a source,  and
they predict  the maximum downwind  concentration of the contaminant as a func-
tion of distance from the source.

          The third model  type,  based  on aerosol  mechanics,  is exemplified by
Davis  (1971).  Here,  the modeling  approach is  based on the properties and
mechanics of  airborne particles, such  as particle  generation, coagulation,
and  settling.  Because the use of  this  method  involves solutions of compli-
cated  equations,  use  of a computer is  recommended.

          The discussion section at  the  conclusion of the Clement Associates
report  notes  that mass balance equations are both  the simplest and most com-
monly  used approach for estimating exposure to  a  substance in an enclosed
space.   However,  there is  no widely  accepted methodology for this type of
analysis.

          Berman  (1982)  provides methodology for  estimating workplace exposure
to  PMN  substances.   His  approach is  based on a  collection of typical  and worst
case approximations  combined with  easily understood equations and nomographs.
For bagging operations,  a  typical  concentration of 2 mg/m3 of material  (30%
of which is respirable)  is  assumed  in  the air  surrounding the bag packer and,
if warranted, a worst  case  value of  30 mg/m3 (also 30% respirable) is assumed
for especially dusty  conditions.   For drum filling of liquids, a saturated
plug of contaminated  air equivalent  to the barrel  volume is assumed to be
generated for every container  filled.  The concentration and total volume  of
                                     A-4

-------
 contaminated air  so  released  is calculated using an equation which accounts
 for vapor pressure,  temperature, and gram-molecular weight of the liquid, as
 well as the filling  rate and  filling mode (top filling versus bottom filling).
 Contribution from spillage  is ignored.  A typical ventilation rate for bag-
 ging and drumming areas is  assumed to be 85 m3/min (3,000 ftVmin) and a worst
 case value is 14.2 mVmin (500 ftVmin).  For outdoor operations with only
 minimal structural protection, the average wind velocity of North America,
 14.5 km/h (9 mph), is assumed in place of a building ventilation rate.

          The validity of this methodology remains unknown, because insuffi-
 cient data are available for verification.  However, the Berman report repre-
 sents the most complete set of guidelines presently available for estimating
 workplace exposure to PMN chemicals during bagging and drumming activities.

          2.   NIOSH  Studies

          A visit to the NIOSH offices in Cincinnati revealed the existence
 of two NIOSH studies which were highly relevant to the current study.   These
 were entitled "Health Hazard Control Technology Assessment of the Silica Flour
 Milling Industry" and "Control Technology Assessment of Dry Chemical Bagging
 and Filling Operations."  A set of three in-depth survey reports was obtained
 for each study (Caplan et al.  1981a; Caplan et al.  1981b; Caplan et al.  1981c;
 Cooper 1984;  Cooper  1983a; and Cooper 1983b).   The silica flour industry study
 also included a summary report.

          The purpose of the silica flour processing plant study (Caplan et al.
 1981a,  1981b,  1981c) was to evaluate the effectiveness of health hazard con-
 trol procedures that are presently being used or being developed for bagging,
 handling, and shipping operations.   The specific evaluation of bag filling
 operations comprised only a small  portion of the overall  study.

          The authors noted that bag filling machines used in silica flour
 plants  vary in design and operation.   They found the most common type  to be
 a pneumatically operated,  forced-flow filling system with one to four  spouts.
 For very fine product flour (~ 5 urn), a single-spout,  auger-feed type  machine
 is normally used.   Each bagging machine is usually equipped with a local  ex-
 haust ventilation arrangement that is connected to a main exhaust system.
 The bagging machines are either manual  or semiautomatic systems, and fill
 multi-ply,  valve-type paper bags.

          Air measurements of respirable dust contamination were made  near
 potential  dust sources and at general work areas using MSA Gravimetric Dust
 Samplers with  cyclone separators.   These samples were analyzed quantitatively
 for respirable dust and respirable silica dust.   At the operation where  the
main product  flour was bagged (at  all three  plants  surveyed),  the average
 respirable  silica dust concentration was 84  ug/m3  (56% silica content).   At
 two locations  where fine product flour (< 38 urn) was being packed intermit-
tently,  the respirable silica dust concentrations  averaged 451 ug/m3  (76%
 silica  content).   Table A-2 describes and summarizes the  ventilation control
systems  at  the various bag filling operations.
                                     A-5

-------
                                            Table A-2.  Ventilation Control Systems Used in Silica Flour Packaging
        Location/operation
                                 Description of system
                               Velocity measurement
                                    m/min (fpm)
                                     Respirable
                                dust concentration
                                               Silica
                            Total    Silica    content
                           ((ig/m3)   (Mg/mJ)     (X)
                                        Remarks
    1.  4-spout main product
        packer, packing 100-Ib
        bags
        a.  behind packer
            machine
        b.  directly behind
            packer operator
                               1.5 x 1.2 m (5 x 4 ft) hood
                               directly behind 4 spouts of
                               machine
                               1.2 x 10 cm (4 ft x 4 in.)
                               slot pulls air from con-
                               veyor belt
                               at hoodface - 14 (45) av.;   103
                               30 cm (1 ft) from face -
                               6.1 (20) av.; at slot -
                               305 (1,000) av.
                                                          Fair control - air velocity
                                                          fairly uniform around spouts.
                                                          Slot behind operator pulls in
                                       96        93       opposition to packer machine
                                                          hood.   It could be more stra-
                                                          tegically located.   Host dust
                                                          emanates from packer.
cn
2.  4-spout main product
    packer, packing 100-Ib
    bags behind packer
    machine
1.3 x 8.4 m (53 x 33 in.)
lateral hood, plus area fan
above head of operator
at hoodface (av.):
  top = 40.8 (134)
  middle = 35.3 (116)
  bottom =  8.5 (28)
                             Good control; air moves from
145       63        43       above and back of operator
                             downward and to rear of packer
                             machine.   Host dust emanates
                             from general environment.
        4-spout main product
        packer, packing 100-Ib
        bags
        a.  behind packer
            machine
        b.  below and behind
            packer operator
                               1.5 x 1.1 m (60 x 45 in.)
                               lateral  hood,  directly be-
                               hind packer machine, moves
                               air laterally  away from
                               operator

                               76 x 10  cm (30 x 4 in.)
                               lateral  slot moves air
                               down and away  from operator
                               at hoodface (av.):
                                 top = 105 (343)
                                 middle = 109 (356)
                                 bottom - 88 (288)
                               rt end   = 366 (1,200)
                               control   = 655 (2,150)
                               left end = 335 (1,100)
                                                                                               212
                                      112
                    53
                                                          Good control;  air moves from
                                                          above and back of operator
                                                          downward and to rear of packer
                                                          machine.   Most dust emanates
                                                          from general environment.
                                                                                                                                   (continued)

-------
                                                                     Table A-2.  (Concluded)
          Location/operation
  Description of  system
                                                                   Velocity measurement
                                                                        m/min (fpm)
                                                                    Respirable
                                                               dust concentration
                                                                              Silica
                                                           Total    Silica    content
                                                          (ug/m1)   (pg/m»)     (X)
                                                                     Remarks
     4.  2-spout fine product
         packer, packing and
         palletizing
         a.  left side of No. 1
             fill spout
         b.  behind No. 2 fill
             spout
(56 x 36 cm) 22 x 14 in.)
side hood, adjacent to
No. 1 spout
18 x 11 cm (7 x 4-1/2 in.)
side
at hoodface (av.):
  top = 55 (180)
  middle = 30 (100)
  bottom = 12 (40)

at slot  = 1,460 (4,800)
at spout =   274 (900)
                                                                                                686
         578
                                                                                                                    84
Uneven airflow, good control
control at No. 1 packer; poor
at No. 2 packer, cross drafts
and palletizing permit dust
dispersion.  Most dust ema-
nates from packer spouts and
bag handling.
 i
-j
     5.  1-spout fine product
         packer, packing and
         palletizing
47 x 29 cm (18-1/2 x
11-1/2 in.) hood behind and
below spout
at hoodface (av.):
  top = 61 (200)
  middle = 75 (247)
  bottom = 32 (105)
                             Good air velocity, but open
344      119        58       sides and remoteness of hood
                             from spout, spills on floor
                             and cross drafts reduce ef-
                             ficiency.   Dust from packer
                             and general environment.
     Source:   Caplan et al.  (1982).

-------
           Local  exhaust ventilation systems which  were  observed  in  these fa-
cilities  included exhaust vents behind the bag packer fill  nozzles  (a  system
of  this  nature is described and shown in figure form in the discussion of
Bureau of Mines  studies) and beneath the operator's seat.   The use  of  an over-
head  fan  blowing down from above and behind an operator was also observed in
one location  in  combination with local exhaust dust vents.  Figure  A-l shows
a typical  relationship between the operator's location,  the open side  of the
bag packer hood, and the dispensing nozzles.   The  ventilation systems  observed
in  these  facilities were capable of maintaining silica  dust concentrations
close to  or below the MSHA TLV of 100 (jg/m3,  when  main  product silica  flour
was being packed.   However, they were not capable  of reducing dust  levels to
the NIOSH recommended level of 50 ug/m3.

           The  three in-depth surveys of dry chemical bagging operations were
conducted at plants which produced calcined diatomite,  an asbestos  product,
and a herbicide  (Cooper 1983b;  Cooper 1984; and Cooper  1983a).

           An in-depth survey of three high volume, valve-type bagging opera-
tions of  calcined (product with particle diameters predominantly <  40 urn) and
fluxed calcined  (product with particle diameters predominantly < 125 pm) diat-
omite was  conducted at Manville Products Corporation in  Lompoc, California.
Two of the systems  were high volume,  manual bagging operations which involved
filling hand-tucked-valve bags.   The third system  consisted of a completely
automatic,  high  volume system for filling pasted-valve  bags.  Factors which
served to  minimize  exposure included automation, ventilated capture hoods at
the packers and  in  the palletizing areas,  and a combination of ventilated and
non-ventilated hoppers and hoods beneath the  packers and conveyor belts.

           Area,  source,  and personal  respirable dust samples were collected
and analyzed.  The  control  technology systems at this plant were capable of
maintaining average respirable  "free" crystalline  silica dust (cristobalite)
concentrations below 0.03 mg/m3  at the two manual  packaging stations being
evaluated.  At the  automated packaging station, these concentrations were
maintained below 0.02 mg/m3.

           Air  velocity,  volumetric flow rates,  and flow patterns were also
obtained  and evaluated for the  bagging and palletizing areas.   Air velocities
entering  the emission control  hoods  on the bagging machines ranged from 110
to  910 m/min (370 to 3,000 fpm).

           The  NIOSH work on asbestos  bagging  is not directly relevant to the
current study, because the fibers per cm3  unit  used to describe air concen-
trations  cannot  be  readily compared  with  the  mg/m3 unit more commonly used
to  describe air  concentrations  of other non-fibrous particulates which are
discussed  in the literature and  modeled in the  current laboratory study.

           The  in-depth survey of a continuous,  open mouth bagging operation
was conducted  at Monsanto in Muscatine,  Iowa.   The herbicide being bagged dur-
ing the survey was  AVADEX BW  (particle size  -  24 to 40 mesh).   The bagging
system consisted of a fully automatic open mouth bagger, the Sackmatic Model
207 manufactured by Quachita Machine  Works and  ancillary automated equipment.
                                     A-8

-------
 Bag Packer
 Mood with
 Dust Exhaust
 to the Rear
Operator's Seat
                                \|    ' Indicates direction of exhaust ventilation air flow

Source-  Adapted from Caolan et al  (1982).
 Figure A-l.   Operator  location and exhause  ventilation
                   patterns  for  a silica sand bag packer.
                                 A-9

-------
 Controls  included an exhaust hood over the bagger,  local  ventilation,  general
 ventilation,  and system automation.

           Area and personal  samples  were taken for  total  and  respirable  dust.
 Total  respirable dust in the bagging room ranged from 0.1 to  1.5  mg/m3 during
 testing.   Samples were also  taken inside the hood.   Ventilation velocity and
 flow  rate were measured for  the bagging system.   Air velocities entering the
 emission  control hoods on the bagging machines ranged from 12 to  18 m/min  (40
 to 60  fpm).

           3.   Bureau of Mines Studies

           The Bureau of Mines (BOM)  has performed research on dust control
 during  the bagging of silica sand and silica flour.   One  of the BOM studies
 (Volkwein,  1978) includes emissions  and bagging  area concentrations, as  well
 as a dust control  hood design suitable for bagging  equipment  common to the
 silica  sand industry.   Figure A-2 shows a typical hood design.  Bagging  emis-
 sions were of 17.5 to 18.3 mg of respirable dust per bag  loaded.  These  emis-
 sion rates were measured in  a room housing a four-nozzle  fluidized valve pack-
 ing machine.   Substantial  background dust drifting  into the bagging area from
 silica  flour  grinding mills  had to be accounted  for by the sampling strategy.
 The production rate during the sampling period varied from 1,170  bags (a mix-
 ture of 50- and 100-lb bags) to 2,100 bags per work shift.  Respirable dust
 levels  at the worker station ranged  from 0.118 to 0.607 mg/m3  during the
 sampling.

           Another  BOM study  (Vinson  et al.  1981)  reports  on the effectiveness
 of bagging machine hood enclosures at controlling emissions in three different
 facilities  approximately 3 years after they were  installed.   In each of  these
 facilities, sulfur hexafluoride was  released inside  the hoods  at  a point ad-
 jacent to the filling nozzle.   Air samples were  then  taken  to  see if the
 tracer gas  was removed by the hood exhaust system.

           In  the first of the three  facilities,  the  average velocity of the
 air entering  the hoods on  a  fluidized-type four  nozzle bagging machine was
 37 m/min  (120 fpm).   Jets  of ventilation makeup  air  entered the bagging area
 (a plywood  enclosure  with  two openings  for a bag conveyor)  through grillwork
 in the roof and produced turbulence  on  the open  side  of the hoods (each of
 the four  nozzles had  its own hood).   This  turbulence  caused some dust to es-
 cape the  hoods,  and the tracer gas also exhibited this pattern.

           In  the second facility a three nozzle machine was enclosed in a room
with its  own  ventilation system.   Average  air  velocity entering the hoods was
 29 m/min  (95  fpm).   Tracer gas  studies  were  more encouraging,  as very little
 gas could  be  detected  outside the  hoods.   BOM  personnel attributed the improved
performance,  compared  to the first facility,  to the  lack of jets of air from
 the ventilation  makeup.

           The  third  facility contained  a bagging machine in an open area at
one end of a  large  building  used for  loading box cars and trucks.   The bag-
ging hoods were  particularly well  constructed  and maintained and the perimeter
                                     A-10

-------
                               ,Ventitaflori hood
     Intokt or
    Sag clamp
                                     Note: '/g-incft clearance
                                         between plate
                                          and hood
                                           Deflector
                                            plate
                      x Spill funnel
Figure  A-2.   BOM dust ventilation hood design.
                   (Source:   Volkwein 1978.)
                           A-ll

-------
of the hood configuration  was  flared  to  reduce  the pressure loss of air enter-
ing the hoods.   The  tracer gas  studies  indicated that the hood system was
working very well.   The  intake  air  velocity of  these hoods was relatively high,
65 m/min (213  fpm).

          Other  BOM  studies, which  are  not as directly relevant for this pro-
gram, have also  been performed  on silica bagging related to dust suppression
techniques, e.g.,  use of water  sprays and foams (Volkwein et al.  1982; Volkwein
et al. 1983).  It  is interesting to note that average respirable dust concen-
tration beside bagging machines without  water spray dust suppression averaged
1.21 mg/m3 and 1.76  mg/m3  at two locations (Volkwein et al.  1983).

          BOM  findings suggest  that over half the worker exposure to respir-
able alpha quartz  dust in  bagging areas  may consist of background dust.   This
situation demonstrates the importance of isolating bagging areas from other
industry operations  and  providing clean  incoming ventilation air.

     C.   Industry  Characterization  Summary

          The bagging and  drumming  of PMN chemicals takes place under a wide
range of conditions  which  influence worker exposure during packaging.*  These
variables include  degree of automation;  employee work practices;  level of
occupational health  awareness  (managers  and employees); filling equipment de-
sign, age, and frequence of maintenance;  container and closure design; plant
layout and ventilation;  effectiveness of emission control devices,  if present;
employee use of  personal protective gear; unit value of product (lower product
loss rates, including packaging emissions, will be permitted if the product
value is high);  and  product flammability.  Some volatile liquids and dust-prone
solids are carefully controlled because  of the fire hazard they represent in
the absence of such  controls.   In an effort to assemble a broadly applicable
data base and methodology  for estimating worker exposure in the bagging and
drumming of PMN  chemicals,  only the most relevant of these variables may be
accounted for.

          1.  Bagging of Solids

          In the bagging of PMN solids it is unlikely that production volumes
will justify fully automated packaging lines in many instances (Bomberger
et al. 1983).  However,  the use of  automated or semiautomated equipment re-
mains a key variable in  predicting  worker exposure.  The product will typi-
cally be placed  into bags  (open top or valve) by gravity feed, auger feed,
or vibrator feed equipment.  Emission controls may or may not be present.
Frequently, an operator  may be  required  to manually staple or otherwise seal
the bag tops.
 The literature is inadequate to permit an exhaustive review of bagging and
 drumming practices for all types of chemical products.   However, as previously
 noted, Bomberger et al. (1983) is a good, up-to-date source of information
 describing typical filling technology.  No attempt is made in this document
 to improve on that compilation.
                                     A-12

-------
           The matrix  shown  in Figure A-3  is based on these primary equipment
 variables  combined with a gross segregation of solids into coarse (particle
 diameters  predominantly > 100 urn) and fine (particle diameters predominantly
 <  100 urn).  The  resulting 24 categories represent different packaging environ-
 ments which the  MRI sampling team would like to characterize.  Realistically
 only a  few of these equipment-material combinations will be examined.  However,
 during  the field work, an effort will be  made to sample as many combinations
 as possible.

           2.  Drumming of Liquids

           In the drumming of PMN liquids  it is expected that both fully auto-
 mated and  semiautomated filling would be  commonly practiced.   The product will
 be either  top filled  (dispenser stays near the top of the barrel, liquid
 splashes freely  during filling) or bottom filled (dispenser remains submerged
 inside  the barrel, as much  as possible, to minimize volatile emissions).
 Emission controls may or may not be present.   An operator may be required to
 manually screw a cap on each barrel or wipe droplets off the tops of the bar-
 rels (Bomberger  et al. 1983).

           The matrix shown  in Figure A-4  is based on these primary equipment
 variables  combined with a general subdivision of liquids into heavier liquids
 with relatively  low vapor pressures, g 0.3 kPa (2.25 mm Hg) at 20°C, and
 lighter liquids  with relatively high vapor pressures, > 0.3 kPa at 20°C.
 This criterion follows a precedent established by Bomberger et al.  (1983) in
 a similar  analysis.  The resulting 16 categories represent different packag-
 ing environments which MRI would like to  characterize.   As with the bagging
 matrix, complete coverage of all categories in the drumming matrix was not
 possible within  the scope of the current  contract.   However,  it served as a
 guideline  for sampling as many equipment-material  categories  as possible.

           For the sampling of liquid drumming lines during the current study,
 emphasis should  be placed on sampling 55-gal  steel  drum filling lines wherever
 possible.   The study will focus on one common container and facilitate the
 comparison of the effects of other variables  on drumming emissions.

           Vapor  pressure is a key parameter for any analysis  of emissions from
 the drumming of  liquids.   However,  it was not possible  for MRI to restrict its
 sampling effort  to materials having a limited range of  vapor  pressures,  because
 PMN chemicals are likely to exhibit the considerable range of vapor pressures
 common to  industrial  chemicals in general.  Table  A-3 lists vapor pressure and
 a number of other useful  properties for a group of organic liquids  representing
 a broad range of vapor pressure.


 III.   ARRANGEMENTS FOR PLANT VISITS

          Substantial  effort was devoted to the task of arranging for on-site
 sampling of bagging and drumming operations.   Initially,  the  response to re-
quests for industry data  and permission to sample  in packaging facilities  was
disappointing.   However,  subsequent to the first laboratory study described
 in Appendix B,  sampling was  conducted at two  plants  with  bagging  and drumming
operations.

                                     A-13

-------
                Range of Respirable Dust  Levels in the Air of the
                        Bagging Area During Operation
                                   (mg/m3)






Coarse
Materials
whose
Particle
Diameters are
Predominantly
> 1 00/tfli
Fine
Materials
whose
Particle
Diameters are
Predominantly
< 100/zm
Gravity Feed
Fully
Automated
With
Dust
Control














Without
Dust
Control














Semi-
Automated
With
Dust
Control














Without
Dust
Control














Auger Feed
Fully
Automated
With
Dust
Control














Without
Dust
Control














Semi-
Automated
With
Dust
Control














Without
Dust
Control














Vibrator Feed
Fully
Automated
With
Dust
Control














Without
Dust
Control














Semi-
Automated
With
Dust
Control














Without
Dust
Control














Figure A-3.   Model sampling  matrix for banqinn  solids.
                            A-14

-------
                Range of Vapor Concentrations in the Air of the
                     Drumming Area During Operation
                                (mg/m3)




Liquids
with
Vapor
Pressures
10.3KPa
at20°C
Liquids
with
Vapor
Pressures
>0.3KPa
at20°C
Top Fill
Fully Automated
With
Emission
Controls








Without
Emission
Controls








Semi -Automated
With
Emission
Controls








Without
Emission
Controls








Bottom Fill
Fully Automated
With
Emission
Controls








Without
Emission
Controls








Semi -Automated
With
Emission
Controls








Without
Emission
Controls








Figure A-4.  Model sampling matrix  for  drumming  liauids.
                            A-15

-------
                                     Table  A-3.  Vapor  Pressures and Other Useful Properties  for a Group of Organic Liquids
                                                          Representing a Broad Range of Vapor Pressures
 i
en
Vapor pressure
ftPa at (mm Hg
STP)* at STP)*
Ethyl ethera'b'c
Methlyene chlorideb'd
Acetone3 lb'e
n-Hexanebff
Isopropyl ether3'6
Methyl alcohol9'*1'6
Isopropyl acetate3'6
2-Butanone (MEK)3'b'e
Isopropyl alcohol3'6'*1'6
Methyl cyclohexane3'6'6
n-Propyl acetate3'6
Toluene1 'b'h'e
Propyl alcohola'b'h'e
Isobutyl acetate3'6
Methyl methacrylate3'6
Isobutyl alcohol3'6
Xylenei'b'h'6
(mixed, o-, m-, p-)
Isoamyl acetate-1'6
Styrene monomer1' 'e
Turpentine1'11'1
Vinyl toluene1'1
Oiethylphthalate3
59
52
30
20
16
13
10
9.3
5.9
5.7
4.7
4.0
2.8
2.7
1.7
1.6
1.3

0.08
0.67
0.67
0.13
0.07
440
390
227
150
119
95
73
70
44
43
35
30
21
20
13
12
10

6
5
5
1
0.5
Boiling
point
<°C)
35
40
57
69
69
64
84
79
83
101
102
111
97
118
100
107
140

143
146
149
171
296
Surface OSIIA
tension (mg/m3
(dynes/
cm)
17 1,200
1,800
24 2,400
1.800
2.100
260
950
25 590
980
2.000
840
28 750
500
700
410
300
~ 30 435

525
420
560
480
-
P.E.L.
) (ppm)
400
500
1,000
500
500
200
250
200
400
500
200
200
200
150
100
100
100

100
100
100
100
none
Flash
point
-45
1
-18
-22
-28
11
4
-6
12
-4
14
4
25
18
10
28
29

25
32
35
54
163
Lower explosive
limit (% by vol.)
1.
15.
2.
1.
1
7.
1
1
2.
1.
2.
1.
2.
2.
1.
1 2 (at
1.

1.
1.
0
.
-
9
5
6
1
4
3
8
8
0
2
0
2
1
4
7
212°F)
1

0
1
8


Upper explosive
limit (X by vnl
36
66
12.
7.
7.
36
8
10
12
6.
8
7.
13.
10.
8.
10.
7.

7
6.

-
-


8
5
9




7
0
1
5
5
2
9
0

5
1



Specific
. ) grav i ly
0 71
1.34
0.791
0.660
0.723
0.792
0.877
0 805
0. 785
0.770
0.887
0 866
0 804
0 871
0.940
0.806
0.87

0 876
0.905
0.857
0.890
1.12
        Source:  Fundamentals of Industrial Hygiene, 1979 and Kirk-Othmer, 1983.
        *STP = standard temperature and pressure, i.e., 0°C and 760 mm of Hg.
        tNonflammable in oxygen.
        Hazards
        .Mild irritation - eye, nose, throat
         Narcosis.
        dFire (severe)
         Chronic systemic toxicity
        6F»re
         Polyneuritis
?Acute toxicity (narcosis, eye damage)
.Dermal absorption
 Moderate irritation - eye, nose,  throat
^Moderate irritation - upper respiratory tract
^Cumulative kidney damage
 Fire (moderate)

-------
      A.   Industry  Contacts

           Initial  attempts were  made  to  gain access  to  plants  through  contacts
with  national  chemical manufacturers  associations  and by  contacting  local
firms in  the  Kansas  City  area.

           1.   National Associations

           Contacts were established with the Chemical Manufacturers  Associa-
tion  (CMA), the Synthetic Organic Chemical Manufacturers  Association (SOCMA),
and the Chemical Specialties Manufacturers Association  (CSMA)  in an  effort  to
locate facilities willing to cooperate.   These three associations were  thought
to represent  a significant fraction of the companies involved  in the bagging
and drumming  of PMN  chemicals.

           2.   Local  Manufacturers

           A review of several Kansas  City business reference books and  chamber
of commerce directories was completed to  compile an expanded list of local
industries.   Table A-4 lists the local industries considered for contact to
gain  access to bag loading or drum filling process operations.  Six  telephone
contacts  have  been made from Table A-4 but no substantial leads have been
achieved  in the initial attempt to find  suitable test sites.  However,  two
referrals  to  other facilities were received.

      B.  Obstacles Encountered in Plant Visit Arrangements

           One  of the obstacles confronting MRI in the effort to gain entry to
packaging  lines, is  the inability to offer an incentive to cooperating  plants.
The only rationale available to persuade  industry personnel  to cooperate is a
statement  of the mutual benefit which both industry and EPA enjoy when  EPA
rule  making is based on sound data.   Thus far, this line of reasoning has not
been  effective.

          As  long as there is no compelling reason, industry personnel   are
not inclined to allow outsiders into their facilities to examine their  opera-
tions and gather data that may eventually be  used to support regulations.
Given the present circumstances, plant managers and corporate officers  appar-
ently feel that cooperation offers the prospect of negligible benefits  and
potentially significant risks.   However,  we will  continue to pursue any leads
which offer the possibility of plant visits and/or the acquisition of relevant
data.
                                     A-17

-------
     Table A-4.   Potential  Sampling Opportunities  in  Greater  Kansas City
                        Along Bagging or Drumming Lines
Adhesives
     H. B. Fuller  Co.
     National Starch & Chemical
       Corp.

Agricultural Chemicals
     Chem-Trol,  Inc.
     Farmland Industries
     Mobay Chemical Corp.
     Pfizer, Inc.
Building Materials
     Aylward Products Co.
     GAF Corp.
     Owens-Corning Fiberglass Corp.
     Sawyer Material and  Sand Co.

Chemicals. General
     Chemcentral-Kansas City
     Dow Chemical Co.
     Reichold Chemical, Inc.
     Thompson-Hayward Chemical Co.
     Union Chemicals
Dairy Products
     Fairmont Foods
     Zarda Brothers Dairy

Emulsifiers
     Grindsted Products, Inc.

Feed
     Carnation-Milling Division
     Columbian Hog and Cattle
       Powder Co.
     National Alfalfa
     Ralston Purina Co.

Industrial Chemicals
     Century Laboratories,  Inc.
     Chemtech Industries,  Inc.
     Liquid Products, Inc.
Protective Coatings
     Epic Manufacturing Co.
     Outgo, Inc.
Inc.
                                     A-18

-------
 IV.   REFERENCES

 Berman DW.   1982.   Methods for estimating workplace exposure to PMN substances
 Clement Associates, Inc.,  Arlington,  VA.   Prepared for U.S.  Environmental Pro-
 tection Agency, Economics  and Technology Division, Washington,  D.C.

 Bomberger DC,  Boughton R,  Endlich R,  Fowler D,  Gikis B, Ludwig  F,  Wilhelm D,
 Witham C.   1983.   Industrial  process  profiles to support PMN review:   filling
 of drums and bags,  draft.   SRI International, Menlo Park,  CA.   Prepared for
 U.S.  Environmental  Protection Agency,  Office of Toxic Substances,  Washington
 D.C.   Contract No.  68-01-6016, Technical  Directive 69.

 Bureau of Mines.   1978.  Dust control  hood for  bag-filling machines.   Techno!-
 logy  News No.  54,  Technology  Transfer  Group.  Washington,  DC:   Bureau of Mines
 U.S.  Department of  the Interior.   Revised June  1983.

 Caplan PE,  Reed LD,  Amendola  AA,  Cooper  TC.   1981a.   Report  on  an  in-depth
 survey of silica  flour dust during packing,  transfer and shipping  at
 Pennsylvania Glass  Sand Corporation, Berkeley Springs,  West  Virginia.   Draft
 report.   Cincinnati,  OH:   National  Institute for Occupational Safety  and Health.

 Caplan PE,  Reed LD,  Amendola  AA,  Cooper  TC.   1981b.   Report  on  an  in-depth
 survey of silica  flour dust during packing,  transfer and shipping  at  the
 Central  Silica Company,  Glass Rock, Ohio.   Draft report.   Cincinnati,  OH:
 National  Institute  for Occupational Safety and  Health.

 Caplan PE,  Reed LD,  Amendola  AA,  Cooper  TC.   1981c.   Report  on  an  in-depth
 survey of  silica  flour dust during packing,  transfer  and shipping  at  Ottawa
 Silica Company, Ottawa,  Illinois.   Draft  report.   Cincinnati, OH:  National
 Institute  for  Occupational  Safety and  Health.

 Clement  Associates.   1981.   Mathematical  models  for  estimating workplace
 concentration  levels:   a literature review.   Clement  Associates, Inc.,
 Washington,  D.C.  Prepared  for U.S. Environmental  Protection Agency,  Eco-
 nomics  and Technology  Division, Washington, D.C.

 Cooper TC.   1984.   In-depth survey  report:  control technology for bag  filling
 operations at  Union Carbide Corporation,  King City, California.   Draft  report.
 Cincinnati,  OH:  National Institute for Occupational  Safety and Health.   Report
 No. 112.21b.

 Cooper TC.   1983a.  In-depth  survey report control technology for a dry  chem-
 ical bagging and filling operation at Monsanto (Monsanto Agricultural Products
 Company), Muscatine, Iowa.   Draft  report.  Cincinnati, OH:  National  Institute
 for Occupational Safety and Health.

Cooper TC.   1983b.  In-depth survey report:  control  technology for bag  filling
operations at Manville Products Corporation, Lompoc,  California.  Draft  report.
Cincinnati, OH:  National Institute for Occupational  Safety and Health.  Report
No. 112,23.

Davis  RJ.  1971.  A simple  model for the estimation of aerosol  concentration
in a closed vessel.   Am Ind Hyg Assoc  J 32:603-609.

                                    A-19

-------
 Fundamentals of Industrial Hygiene.   1979.   Olishifski  JB,  ed.   2nd ed.
 National  Safety Council, Chicago, 111.

 Kirk-Othmer.   1978-1984.  Encyclopedia of chemical  technology,  3rd ed.
 New York:   John Wiley and Sons.

 Schroy JM.   1981.   Prediction of workplace  contaminant  levels,  pp.  190-206.
 Symposium Proceedings on Control Technology in the  Plastics and Resins  Indus-
 try,  Atlanta,  GA,  February 27-28.  Cincinnati, OH:   National  Institute  for
 Occupational  Health and Safety.   NIOSH Publication  No.  81-107.

 Vinson RP,  Volkwein JC, Thimons  ED.   1981.   SF6 tracer  gas  tests  of bagging
 machine hood enclosures.  Washington, DC:   U.S.  Department  of the Interior.
 Report of Investigations, Bureau of  Mines,  8527.

 Volkwein  JC,  Thimons ED.  1983.   Ventilation hoods  to reduce  dust during bag
 filling.   Mining Equipment International.

 Volkwein  JC,  Vinson RP, Thimons  ED.   1982.   Effectiveness of  three  water spray
 methods used  to control dust during  bagging.   Washington, DC:   U.S.  Department
 of  the Interior.   Report of Investigations,  Bureau  of Mines,  8614.

 Volkwein  JC.   1979.   Dust control in bagging operations, pp.  51-57.  ACGIH
 Symposium Proceedings on Industrial  Hygiene  for Mining  and  Tunneling, November
 6-7,  1978.   Cincinnati, OH:   American Conference of Governmental  Industrial
 Hygienist.


 V.  ANNOTATED  BIBLIOGRAPHY OF REFERENCES ON  FILE

          Table A-5 summarizes references on file at MRI.   Each document in
 the file  is  summarized below by  document number.

 P101

 Bomberger DC,  Boughton R,  Endlich R,  Fowler  D,  Gikis B, Ludwig  F, Wilhelm D,
Witham C.  1983.   Industrial  process  profiles  to support PMN  review:  filling
 of drums  and bags,  draft.   SRI International,  Menlo  Park, CA.   Prepared for
 U.S.  Environmental  Protection Agency, Office of Toxic Substances, Washington,
 D.C.   Contract No.  68-01-6016, Technical Directive  69.

          Technologies commonly,  used  for bagging>and -drumming solids and for
drumming  liquids are  discussed.'   Methodologies  for modelling workplace con-
centrations are also  discussed.

          Typical  equipment  and  modes of operation are described  in detail
 (i.e.,  bag and drum designs,  container closure methods, conveying and filling
equipment configurations,  and emission control  systems).  Vapor and particu-
 late  emission  rates and the  resulting workplace air concentrations during bag
and drum filling of chemicals are estimated.
                                     A-20

-------
                                                        Table A-5   Summary of References on File at MRI
 I
PO
Doc.
No.
P101
P102
P103
P104
P105
P106
P107
P108

P109
P110
Pill
P112
P113
P114
P115
P116
P117
P118
P119
P120
P123
P124
P125
P126
P127
P128
P129
P130
P131
P132
P133
P134
P135
P136
P137
P138
P139
P140
P141
P142
Reference
Bomberger et al. (1983)
SRI (1980)
Clement Assoc. (1982)
Clement Assoc. (1981)
Caplan et al. (Sep. 1981)
Caplan et al. (Nov. 1981)
Caplan et al. (Aug 1981)
Cooper (Jan. 1984)

Cooper (Oct. 1983)
Wang (1981)
Wolfe et al. (1978)
Winegar (1983)
Clarke (1983)
Buyers Guide & Dir. (1982)
EPA (1983)
NIOSII (July 1978)
Caplan et al. (June 1982)
Becker et al. (1979)
Cooper (Mar. 1983)
Wechsler et al. (1982)
Hertle (1981)
Holmgren (1982)
Cahners (1982)
Checkoway & Williams (1982)
Davies et al. (1982)
Volkwein et al. (1982)
Vinson et al. (1981)
Brest in & Niewicdomski (1982)
Ward & Campbell (1983)
Phillips & Jones (1978)
Thomson (1977)
Midwest Int. (1981)
ACS (1983)
ACS (1982)
Freed et al. (1983
Adkins et al. (1983)
Dixon et al. (1983)
Versar (Apr. 9. 1982)
Versar (Apr. 7, 1982)
Flores (1983)
Specific
chemicals


X
X
X
X
X
asbestos
X
X
X
X




X
X

X

X
X
X
X
X
X
X
X



X
X
X
X
X



X
Physical Workplace
properties filling
of chemicals operations*
X X D,B
X X 0,8
X D.B

X X B
X X B
X X B
X B

X B
X X B
X B
X D.B
X


B


X B

X B
X X
X B

X
B
B
B

X
X
X X








Emissions Workplace
controls
X X
X


X
X
X
X

X
X





X
X

X
X

X


X
X
X X
X
X


X





X

X
Workplace
concentrations
or exposures
X
X
X

X
X
X
X

X

X




X

X
X




X

X
X


X







X

X
Dispersion
mode 1 s
X

X
X































X
X




                                                                                                                                 (continued)

-------
                                                                        Table A-5.  (Concluded)
Doc.
No.

P143
P144
P145
P146
P147
P148
P150
P151
P152
P154
P155
P156
An J--
Reference
r \ -

Dunn et al. (1983)
Tomb & Treattis (1983)
Peach & Myers (1981)
Schroy (1981)
Volkwein (1979)
Bureau of Mines (1983)
Volkwein et al. (1983)
Volkwein & Thimons (1983)
Bureau of Mines (undated)
O'Leary et al. (1983)
Callahan (1978)
Cahners (1983)
Specific
chemicals

X

X
X

X




X
X
Physical
properties
of chemicals


X








X
Workplace
filling
operations*
X

B

B
0

B
B
X
X

Emissions Workplace
controls

X
?
X X
X
X X
x'
X
X
X
X
X X

Workplace
concentrations
or exposures
X

X
X








Dispersion
models










x


(\>
ro

-------
          The discussion on volatile emissions from drumming liquids is based
 largely  on  data  from the petroleum refining industry, because no other data
 were available.  Emission rates for drum filling of a light liquid with a
 relatively  high  vapor pressure (benzene) and for a heavy liquid with a rela-
 tively low  vapor pressure (cyclohexanol) are estimated under three sets of
 filling  conditions:  a high rate of emissions (top filling with no collection
 of vapors and no routine maintenance of fittings which are not inside a hood),
 a moderate  rate  of emissions (top filling with 95% collection of vapors), and
 a low rate  of emissions (bottom filling with 95% collection of vapors).

          The authors conclude that it is difficult to make general statements
 on packaging practices, especially on emission control practices, in the
 chemical manufacturing industry.   The bulk of packaging machinery is designed
 and sold for the packaging of foods, cosmetics, and Pharmaceuticals.  Thus,
 emission controls, if present at all on original equipment, are generally for
 control  of  nuisance dusts.

 P102

 SRI.  1980.  Control technology assessment of the pesticides manufacturing
 and formulating  industry, draft.   SRI International, Menlo Park, CA.  Pre-
 pared for National Institute for Occupational Safety and Health, Cincinnati,
 OH.   Contract No. 210-77-0093.

          The 50-page product packaging section of this 652-page report on
 pesticide production covers methods used to reduce operator exposure during
 13 specific bagging and liquid filling operations.   Most dry pesticide is
 packaged in bags and most bags are open-top rather than valve-top or bags
 formed of heat-sealable films.   Drum filling is generally semi automated.

 P103

 Clement  Associates.  1982.  Methods for estimating workplace exposure to PMN
 substances.  Clement Associates,  Inc., Arlington, VA.   Prepared for U.S.
 Environmental Protection Agency,  Economics and Technology Division, Washington,
 D.C.

          Representative scenarios were developed for operations associated
with processing  chemicals.  Operations considered were drumming of liquids,
 drumming and bagging of solids,  cleaning and maintenance, and sampling and
 analysis.  Rates of generating emissions during filling of drums with liquids
 depend principally on drum size and filling rate (drums/h).   For each drum
 filled,  a plug of chemical-saturated air is assumed to be emitted.   The com-
plex dependence  of dust formation during drumming and bagging on properties
of solids gives  a range of values for dust generation from a variety of re-
 lated processes.   Models are applied to the data developed for  the process
operations to estimate workplace  exposures.   Airborne particulate concentra-
tions associated with handling a  few specific solid chemicals were compiled
 from the literature.   For example, total  particulate concentration in air
while bagging azelaic acid was  0.2 to 6.2 mg/m3  (30% respirable) and in drum-
ming azo dyes was 13.1 mg/m3.
                                     A-23

-------
 P104

 Clement Associates.   1981.  Mathematical models for estimating workplace  con-
 centration levels:   a literature review.  Clement Associates,  Inc.,  Washington,
 D.C.   Prepared for U.S.  Environmental Protection Agency,  Economics  and  Technol-
 ogy Division,  Washington, D.C.

           Mathematical  models for estimating workplace concentrations of  chem-
 icals include  mass  balance models (simple models, simple  equilibrium models,
 and complex models); the Schroy model;  and the Davis model.  Measured work-
 place concentrations for benzene, hexane, and toluene are compared with con-
 centrations predicted by use of the equation for the simple  mass  balance
 equilibrium model:   Cgq  = G/kQ, where C   is the equilibrium concentration,

 G  is  the generation  rate, k is  the mixing factor, and Q is the ventilation
 flow  rate.   The Schroy model [see Schroy (1981)], which uses both emission
 rate  and Gaussian diffusion, predicts maximum downwind concentration as a
 function of distance from the source.   The complex Davis  model  calculates
 aerosol  concentration in a closed vessel.

 P105

 Caplan  PE,  Reed LD,  Amendola AA,  Cooper TC.   1981.   Report on  an  in-depth
 survey  of silica flour dust during packing,  transfer,  and shipping at
 Pennsylvania Glass Sand  Corporation,  Berkeley Springs,  West  Virginia.   Draft
 report.   Cincinnati,  OH:   National  Institute for Occupational  Safety and
 Health.

           Ventilation and other dust  control  measures  during bagging of silica
 flour into  50-  or 100-Ib bags and bulk  loading of railroad cars and  trucks are
 described in detail  with diagrams of  plant layout and  machinery.  The average
 silica  dust concentration was 578 ug/m3  in the vicinities  of the  one-spout
 packer  (bagging)  machines for the 10-15-30 micron product.   High  silica dust
 concentrations  were  probably caused by  bag breakage,  ineffective  ventilation
 at  packer spouts, and poor housekeeping.   The ventilation  was better at the
 four-spout  packer and at the packer for  the  5-micron product.

 P106

 Caplan  PE,  Reed LD,  Amendola AA,  Cooper  TC.   1981.  Report on an  in-depth
 survey  of silica  flour dust during packing,  transfer,  and  shipping at the
 Central  Silica  Company,  Glass Rock, Ohio.   Draft  report.   Cincinnati, OH:
 National  Institute for Occupational Safety and Health.

          Atmospheric dust concentrations  and ventilation  control  systems were
 evaluated  for packing (filling  of 100-Ib  bags  with a four-spout packer)  and
 transfer  operations.   The  bag handling workers  rotate  from packing to loading
 to  stacking  opeations in  each shift.  The  average  silica dust concentrations
were  112  ug/m3  at the packer area and 165  ug/m3 at the bag handling area.
                                     A-24

-------
 The  ventilation was  so  good  that  dust concentrations were practically  the
 same whether bagging and truck  loading were occurring  or not.  The  high  con-
 centrations (264 ug/m3  during loading compared to < 28 ug/m3 during no load-
 ing) at  the conveyor discharge, where bags were off-loaded  from  the conveyor
 and  stacked on pallets, apparently arose from the bags as leakage,  spills,
 or surface contamination.

 P107

 Caplan PE, Reed LD,  Amendola AA,  Cooper TC.  1981.  Report  on an  in-depth
 survey of silica flour  dust  during packing, transfer and shipping at Ottawa
 Silica Company, Ottawa, Illinois.  Draft report. Cincinnati, OH:  National
 Institute for Occupational Safety and Health.

          Dust concentrations in  the air, ventilation  control systems, and
 other dust control techniques were evaluated in the packing area  (where  100-lb
 paper bags were filled with  silica flour), in the bag  handling area, and in
 the  boxcar loading area.  Bag handling workers normally rotate among the jobs
 of operating the packer machine,  transferring bags, and stacking bags  in box-
 cars or  trucks.  Two packer machines had two fill-spouts and three  had four
 spill-spouts each.  Water injection into the product and water spray onto bag
 surfaces reduced atmospheric dust concentrations by 60 to 80% (80%  in  the bag
 handling areas).   Local exhaust,  the water control methods, and housekeeping
 reduced  packing station dust concentrations to 50 ug/m3 air.

 P108

 Cooper TC.   1984.   In-depth survey report:   control technology for  bag filling
 operations at Union Carbide Corporation, King City, California.   Draft report.
 Cincinnati, OH:  National  Institute for Occupational Safety and Heatlh.  Report
 No.   112.21b.

          Exemplary controls used in packaging chrysotile asbestos  fibers at
 the  title plant maintained bagging area atmospheric concentrations  at < 0.9
 fiber/cc.  Bag filling and bag flattening operations are enclosed,  and the
 enclosures are equipped with exhaust ventilation.   Pellets are shrink wrapped
 and  stretch wrapped; pelletizing  is automated.   Bagger rooms are isolated.
 Bagging equipment used includes three single-spout force-flow packers and one
 single-spout auger packer.   Filled bag weights are 10 to 82 Ib;  packaged den-
 sities range from 5 to 45 lb/ft3  (pellets).

 P109

 Cooper TC.   1983.   In-depth survey report:   control  technology for bag filling
 operations  at Manville Products Corporation,  Lompoc, California.   Draft report.
Cincinnati,  OH:   National  Institute for Occupational Safety and  Health.  Report
No.  112,23.

          The control technology used at the  title plant for three high-volume
operations  bagging calcined and fluxed calcined diatomite maintained average
 respirable  "free"  crystalline silica dust (cristobalite) at concentrations
                                     A-25

-------
<  0.03  mg/m3  air at the two manual packaging stations  and at  <  0.02  mg/m3 at
the  automated packaging station.   The manual bagging operations filled  hand-
tucked-valve  bags;  the completely automatic system filled pasted-valve  bags.
Ventilated  capture  hoods were used at the packers  and  in  the  pelletizing areas,
and  both  ventilated and nonventilated hoppers and  hoods were  used  beneath the
packers and conveyer belts.   The  relations between occupational  atmospheric
dust exposures (as  determined by  analysis of area, source,  and  personal
respirable  dust) and control systems were evaluated.

P110

Wang CCK.   1981.  Walk through survey report of the pesticide/herbicide bag-
ging operation at Monsanto Agricultural  Products Company,  Muscatine, Iowa.
Cincinnati, OH:   National  Institute for  Occupational Safety and Health.  NTIS
No.  PB83-243840.

          The title bagging operation for filling  50-Ib paper bags with gran-
ulated  (40  to 50 mesh) pesticide/herbicide formulations and the  control tech-
nology  are  described in detail.   Several  figures depict the machinery used.
The  strict  industrial  hygiene policies are also discussed.  The  control tech-
nology  may  be successfully adapted to control  dust problems in  other bag fill-
ing  operations.

Pill

Wolfe HR, Staiff DC,  Armstrong JF.   1978.   Exposure of pesticide formulating
plant workers to parathion.   Research Triangle Park, NC:   U.S.  Environmental
Protection  Agency,  Office  of Research and Development.   EPA-600/J-78-042,
NTIS No.  PB-287  134.

          Worker respirator and dermal exposure potential was determined near
certain work  stations  at a parathion formulating plant.  Ordinary work cloth-
ing  was assumed  to  offer minimum  protection.   Mean parathion exposure was
67.3 mg/h of  work activity;  for inhalation exposure, the value was 0.62 mg/h.
Bagging machine  workers were exposed to more parathion than were workers who
operated  the  bag/closing machine  or handled the filled bags for storage or
shipment.    Calculated  inhalation  exposure for  full-time baggers was 0.69 ±
0.77 mg/h and for baggers  who were also mixers,  the inhalation exposure was
0.93 ±  0.51 mg/h.   No  information is given about the bagging or control ma-
chinery,  and  the  plant is  not identified.

P112

Winegar DL.    1983.   Machinery-filling, dry products.  Package Engineering
28(4):229-231.

          Gravimetric  and  volumetric filling equipment for  dry materials are
described with schematics.   The two types  of gravimetric equipment are net
weighing  and  gross  weighing  fillers.   Volumetric filler types include timed
vibrator,  auger,  vacuum-filled container,  gravity-filled pocket, and vacuum-
filled pocket fillers.   (They are listed  in order  from least to most accurate.)
                                     A-26

-------
 Types  of  materials,  containers,  and container  size  suitable  for  use with  each
 filler type  are mentioned.

 P114

 Buyers Guide and  Directory.   1982.  Package  Engineering  27(11).

          This directory  issue of  Package Engineering  lists  the  names of
 manufacturers of  machinery  and containers, contract packagers, and research
 and testing  laboratories.

 P115

 EPA.   1983.  Premanufacture notification requirements  and review procedures.
 Washington,  DC:   U.S. Environmental Protection Agency.   40CFR720.

          Premanufacture  notification (PMN)  requirements and review procedures
 through 48FR41132, September  13, 1983, include specifications for procedures
 for reporting new chemical  substances by manufacturers and importers under
 Section 5 of the  Toxic Substances  Control Act, applicability, notice forms,
 disposition  of notices, confidentiality and  public  access to information,
 commencement of manufacture or import, and compliance  and inspection.

 P116

 NIOSH.  1978.  Criteria for a recommended standard	 occupational exposure
 during the manufacture and formulation of pesticides.   Cincinnati, OH:   National
 Institute for Occupational Safety  and Health.  Publication No.  78-174.

          Packaging operations in  the pesticide industry may be done remotely
 (automatic dispensing, weighing, and sealing) but is more frequently done with
 a local exhaust system or with no  specific controls.   Pesticides may be pack-
 aged in the form  of powders,  liquids,  dust,  granules,  or pressurized aerosols.
 Containers vary from small glass bottles to drums and railroad tank cars.
 Bagging or mixing stations are often work sites with the highest exposure po-
 tential.  At a plant formulating 4 and 50% carbaryl  dusts, workers at such
 stations had a mean respiratory potential exposure of 1.1 mg/h work activity.
 In the formulation of a 50% DDT wettable powder,  the baggers who filled 50-Ib
 and 4- or 5-lb bags had potential exposures  of 160 and 539 mg/h,  respectively.
Workers in liquid filling operations may be  exposed to the vapors before cap-
 ping or to the pesticide by splashing.   Remote control filling is recommended,
 but bottom filling of drums instead of top filling would minimize splashing.
 Properly designed local  exhaust ventilation  is recommended for dust collection
during solids packaging operations.

P117

Caplan PE, Reed LD,  Amendola AA,  Cooper TC.   1982.   Health hazard control
technology assessment of the silica flour milling industry.   Draft report.
Cincinnati,  OH:   National  Institute for Occupational Safety and Health.
                                     A-27

-------
           Three silica flour mills were evaluated in  depth.   The  primary oper-
 ations  investigated were bag packing,  conveyor transfer  of  bags,  palletizing,
 loading of bags in boxcars and trucks, and bulk loading  of  hopper cars and
 trucks.   Several  innovative dust control procedures were found  to be  effective.
 These  included:   (1) the injection of  agglomerating agents  such as water and
 Deter Micro Foam® into products; (2) water spraying of outer  surfaces or
 product-filled bags during conveyance; and (3) bulk loading of  silica flour
 into enclosed hopper trucks and railroad hopper cars  under  local  exhaust ven-
 tilation control.   Plant personnel have estimated that half of  the environ-
 mental  silica dust problems may be effectively controlled by  good work prac-
 tices and effective housekeeping procedures.

 P118

 Becker  D,  Fochtman E,  Gray A,  Jacobius T.   1979.   Methodology for estimating
 direct  exposure to new chemical  substances.   Washington,  DC:  U.S. Environ-
 mental  Protection  Agency,  Office of Toxic  Substances.  EPA-560/13-79-008, NTIS
 No. PB80-102262.

           The work reported was  directed toward the development of a procedure
 for the  orderly and rapid  prediction of direct human  exposure which might re-
 sult from the manufacture  of new chemical  substances.  The  procedure developed
 involves  the  following steps:   (1) prediction of  unavailable  physical and chem-
 ical properties from analogs and general chemical  knowledge,  (2)  prediction of
 production  volume  based upon company size,  current markets  and  total market
 volume,  (3) prediction of  chemical  operator exposure  and  exposures in the
 vicinity  of the plant  based upon fugitive  emissions,  and  (4)  prediction of
 consumer  exposure  based upon active use and passive use of  the chemical.

 P119

 Cooper TC.  1983.   In-depth survey report  control  technology  for  a dry chem-
 ical bagging  and filling operation at  Monsanto (Monsanto  Agricultural Products
 Company), Muscatine,  Iowa.   Draft report.   Cincinnati, OH:  National Institute
 for Occupational Safety and Health.

          Controls  for packaging the granular herbicide AVADEX BW  in 50-lb
 pinch bottom  bags  by a fully automatic  open-mouth bagger  included  an exhaust
 hood over the  bagger,  local  and  general  ventilation,  and  system automation.
 Relations among ventilation velocity,  flow  rate, and  total and respirable
dust from personal,  area,  and  hood  interior samples were  evaluated.  Average
atmospheric concentrations  over  a 2-day period were 2.9 mg/m3  for  total  dust
 inside the hood, 1.9 mg/m3  for total dust on  the hood face,  1.5 mg/m3 for
total  dust in  the  bagging  room,  0.2 mg/m2  for respirable  dust in the bagging
 room,  1.1 mg/m3 for  concentration of total  dust to which  two  operators were
exposed, and  0.2 mg/m2  for  the background  level.
                                     A-28

-------
 P120

 Wechsler AE,  Eschenroeder AQ,  Gilbert D,  Loos  K,  Poston P,  Stevens  JM.   1982.
 Feasibility of developing a comprehensive methodology for source identification
 and environmental  loading (materials  balance).  Athens, GA:   U.S.  Environmental
 Protection Agency,  Environmental  Research Lab.  EPA-600/3-82-047,  NTIS  No.
 PB82-239286.

           Materials balance methodologies are  discussed from extraction of  the
 raw materials  through  manufacturing/processing, use,  and disposal.   The focus
 is  on  releases to  the  outdoor  environment.   Fugitive  processing  emissions are
 mentioned,  but methodologies for  estimating  releases  to indoor air  are  not
 covered.

 P123

 Mertle  B.   1981.   Bulk packaging—from  bauxite to  bag,  Reynolds  makes easy
 work of it.  Material  Handling Engineering 36:114-119.

           Automatic packaging  of  alumina  in  50- and 100-Ib  bags  is  described.
 Ten auger-type bag  packing  machines are used for hydrates and ground calcines,
 and three  air-type  packers  are used for unground calcined alumina.

 P124

 Holmgren RB.   1982.  Electronics, skilled hands keep  bag  line's  output  high.
 Package Engineering 27:69-72.

          Automated packaging  of  powdered milk in  50-Ib  bags is  described.
 Weighing,  filling,  and  sealing are automated, but  five  men  are required, one
 each to attend the  bagger,  to  hand-tie  each bag liner,  to oversee the sealer,
 to  stack the bags,  and  to operate the forklift truck.    The  first four fre-
 quently alternate their jobs.   Vacuum tubes at the point of  filling, air
 evacuation, and a large hood above the  point of bag tying control dust.

 P125

 Cahners.  1982.  Bag's  valve,  unitizing cut hazardous leakage.  Cahners
 Publishing Company,  Denver, CO.   Package Engineering 27:78-9.

          Bag  leakage in packaging silica flour was solved by switching to a
 bag with a polyethylene valve  and by shrink wrapping pallets before shipping.
 Formerly, use  of standard kraft bags (50- or 100-lb) with tuck-in sleeve and
 pasted  valve allowed a  small amount of  silica flour to force its way back out
 through the bag's valve.  Once  the new bags are filled, the valve overlaps
 itself  to form a seal.

 P126

 Checkoway H, Williams TM.   1982.  A hematology survey of workers at a styrene-
 butadiene synthetic rubber manufacturing plant.   American Industrial Hygiene
Association Journal 43(3):164-169.
                                     A-29

-------
           Of the  workers  at  the  title  plant,  workers  in  shipping  and  receiving,
 who transport (packaging  not explicity mentioned)  the finished  latex  products,
 were exposed to the  lowest concentrations  of  butadiene (0.08  ppm),  styrene
 (0.13 ppm),  and benzene  (benzene is  an impurity  of styrene  and  toluene)  (0
 ppm) and  the second  lowest concentration of toluene (a styrene  impurity,  also
 used as a tank cleaning  agent)  (0.05 ppm).

 P127

 Davies JM, Strunin L,  Craig  DB.   1982.  Leakage  of volatile anaesthetics  from
 agent-specific keyed vapourizer  filling devices.   Canadian  Anaesthetics  Society
 Journal 29(5):473-476.

           Agent-specific  keyed vaporizer filling devices  (designed  to ensure
 that an anaesthetic  vaporizer is filled with  the correct  agent) should be left
 on  the bottle of  volatile  agent  between fillings.   Operating  room pollution
 is  greater when the  device is replaced by  the screw-on cap  between  use.

 P128

 Volkwein  JC,  Vinson  RP, Thimons  ED.  1982.  Effectiveness of  three  water spray
 methods used to control dust during  bagging.  Washington, DC:  U.S. Department
 of  the Interior.   Report of  Investigations, Bureau  of Mines,  8614.

          Three methods of water application  to suppress  dust during the bag-
 ging of industrial sand were evaluated:  injecting  water  into the bag while it
 filled, atomized  spraying  the outside  of the bag during filling and conveying,
 and  wetting  the outside of the bag with a  plain water mist  during conveying.
 Water injection and  external  wetting plus  injection reduced average dust con-
 centrations  beside the bagging machine  from 1.21 mg/m3 and  dry conditions to
 0.61 and  0.87  mg/m3, respectively.

 P129

 Vinson RP, Volkwein  JC, Thimons  ED.  1981.   SF6 tracer gas  tests of bagging
 machine hood  enclosures.    Washington,  DC:   U.S.  Department  of the Interior.
 Report of Investigations,  Bureau  of Mines, 8527.

          SF6  tracer gas tests in specially designed hoods  of machines for
 bagging silica indicated that makeup air must be evenly dispersed,  hood enclo-
 sures and duct systems must  be airtight and must be properly maintained, and
 the  average  intake air velocity  of the  hood must be £ 200 fpm.  The effective-
 ness  of the  hoods during bagging  at three  facilities varied from no tracer
 gas  found outside the hood at one plant to significant SF6 concentrations in
 the  breathing  zone of the  operator at another plant.

 P130

Breslin JA, Niewiadomski  GE.   1982.  Improving dust control  technology for
U.S. mines:  the  bureau of mines  respirable dust research program, 1969-82.
Washington, DC:  Bureau of Mines, U.S.  Department of the Interior.
                                     A-30

-------
           The  title  program  has  studied  new  and  improved  techniques  and  equip-
 ment  to  control  dust during  mining  operations  such  as  cutting.   Control  of
 coal  mine  dust by  ventilation, conveyer  screens, water sprays,  dust  collectors
 or  scrubbers is  described.   Silica  dust  emissions during  silica bagging  opera-
 tions are  controlled by  equipping bagging machine with exhaust  ventilation
 hoods.   Sampling and analysis methods  are discussed.

 P131

 Ward  P,  Campbell D.   1983.   Requirements definition for process control  systems,
 pp. 585-600.   Proceedings of the ISA International  Conference and  Exhibit on
 Advances in Instrumentation, Vol. 38,  Part 1,  Houston,  TX, October 10-13, 1983.
 Instrumental Society of  America.

           "Structured analysis" provides a well-developed framework  for  de-
 scribing requirements for typical data processing systems.  This paper de-
 scribes  an extension of  structured  analysis  to describe process control  system
 requirements.  The use of the technique  is illustrated on a bottling system.
 [Adapted from  the  author abstract.]

 P132

 Phillips CF, Jones RK.   1978.  Gasoline vapor exposure  during bulk handling
 operations.  American Industrial Hygiene Association Journal 39(2):118-128.

           Gasoline vapor concentrations were determined in the  breathing zone
 of  employees during  bulk loading of gasoline tank trucks with the  use of var-
 ious  types of  loading and control systems.   Top and bottom loading operations
 with  and without vapor recovery are described.  Employee exposure  to gasoline
 vapor was  lowest at  the  plant with top loading and  vapor recovery  (98% of the
 samples contained £  25 ppm gasoline).  Full-time loaders are probably exposed
 to  ^ 85 ppm gasoline  on  98%  of the days during top  loading without vapor
 recovery.

 P133

 Thomson FM.  1977.    Storage, feeding, and metering of  fine powders, pp.  123-
 129.  Proceedings of  the Workshop on Particle Technology-Research  Needs,
 Opportunities,  and Priorities, Philadelphia, August 1975.   Washington, DC:
 National  Science Foundation.

          State-of-the-art design for powder storage bins and hoppers is dis-
 cussed.  Several areas needing additional research are pointed out.  For
 example,  poor weight  control  and operator exposure to dust during  package
 filling are ascribed  to  a lack of knowledge about the effect of the gaseous
 phase (air) on dry powder flow behavior.

 P134

Midwest International.   1981.  Dust eliminated during filling operation.
Midwest International, Inc.,  Charlevoix,  MI.   Chemical  Engineering 88(13):
55-56.
                                     A-31

-------
           A new  system  for  filling  semibulk  flexible  containers  improves  oper-
 ating  rates,  reduces manpower  needs,  and  controls dust without separate pollu-
 tion equipment.   The system was  designed  for Pfizer,  Inc.'s, Minerals, Pigments,
 and Metals Division for filling  containers with  approximately 1,400  Ib precipi-
 tated  calcium carbonate powder of bulk  density 17 to  25  lb/ft3.  The  system
 comprises  a cyclone separator  and aerated storage bin, a dust suppression  unit,
 a  hydraulic container support  and lifting mechanism,  and a  scale with digital
 readout.

 P135

 ACS.   1983.   Production by  the U.S. Chemical  Industry.  American Chemical
 Society, Washington, DC.  Chemical  and  Engineering News 61(24):28-34.

           Statistics are given for  U.S. production in 1982  and previous
 years  of the  top  50 chemicals, fertilizer raw materials, and synthetics
 (rubber, plastics, and  fibers).

 P136

 Plant  AF,  ed.  1982.  ACS.  American Chemical Society, Washington, DC.
 Chemcyclopedia 1982-83:   Volume  1.

           Thousands of  commercially available chemicals are tabulated along
 with information  including  physical forms in which the chemicals are  available,
 shipping quantities, and shipping containers.  This version is more useful for
 quick  identification and enumeration of chemicals that are commonly drummed
 or bagged  than the 1983 edition, which  is formatted in paragraphs for each
 chemical.

 P137

 Freed JR,  Chambers T,  Christie WN, Carpenter CE.   1983.   Methods for assessing
 exposure to chemical substances, volume 2:  methods for assessing exposure to
 chemical substances in  the ambient environment.   Washington, DC:   U.S. Environ-
 mental  Protection Agency, Office of Toxic Substances.   EPA 560/5-83-015.

          One of nine  volumes,  this volume describes environmental  pathways,
 sources of monitoring  data,  and the calculation of exposures by different
 routes.  (Models used  to calculate inhalation exposures  are discussed.)   The
 bulk of the volume comprises three appendices:  (a)  a compilation of pertinent
computerized data bases  and  printed sources;  (b)  lists of chemicals regulated
by EPA under the Resource, Conservation, and Recycling Act (RCRA),  the Clean
Air Act, and the Clean Water Act and by OSHA, and lists  of chemicals studied
by these agencies and  NIOSH; and (c) emission characteristics  of the organic
chemicals  industry, the  plastics industry, and the lubrication and hydraulic
fluids  industry.
                                     A-32

-------
 P138

 Adkins  LC,  Doria  JJ,  Christopher  MT.   1983.   Methods  for  assessing  exposure
 to  chemical  substances,  volume  3:   methods  for  assessing  exposure from  dis-
 posal of  chemical  substances.   Washington,  DC:   U.S.  Environmental  Protection
 Agency, Office  of Toxic  Substances.   EPA  560/5-83-016.

          Environmental  releases  from chemical  disposal are  estimated in  a
 five-step process:   (1)  estimate  releases to  disposal;  (2) characterize waste
 stream  releases and  concentrations;  (3) allocate waste  streams to disposal
 practices;  (4)  allocate  waste streams to  individual disposal  sites;  and (5)
 estimate  environmental releases from disposal sites.  Disposal practices  con-
 sidered are  landfilling,  land treatment,  holding in surface  impoundments,
 sewage  treatment,  incineration, and  deep-well injection.  Besides appendices
 E through J  containing auxiliary  information  on  the disposal  practices, other
 appendices cover:  (a) useful models  and  data bases;  (b)  information collected
 by  state  solid  waste  agencies on  waste generation, disposal  practices,  and
 disposal  facilities;  (c) waste  disposal practices of  the  petroleum  refining
 and  organic  chemicals industries; and (d) various kinds of information  needed
 for  step  (3) above.

 P139

 Dixon DA, Hendrickson GL, Jennings P, Chambers T, Borenstein A, Doria J,
 Faha T.    1983.  Methods  for assessing exposure to chemical substances,  volume
 9:   methods  for enumerating and characterizing populations exposed to chemical
 substances.   Washington, DC:  U.S. Environmental Protection Agency,   Office of
 Toxic Substances.  EPA 560/5-83-022.

          Methods  for identifying, enumerating,  and characterizing populations
 exposed to chemicals are described.    Populations considered are those exposed
 in the ambient  environment, in the occupational  environment, via food, via
 consumer  products, and via drinking water.  Appendix A applies the methods to
 example problems for each exposure pathway.   Appendix B gives examples of data
 bases used in the  occupational populations methods section.   (No information
 is presented on enumerating subgroups of a particular occupational population
 such as packagers.)

 P140

Versar.    1982.   Appendix A, processes and exposure potential, methodology for
assessing occupational exposure.  Versar,  Inc.,  Springfield,  VA.   Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances.   EPA Contract
No.  68-01-6271,  Task 10.

          Multimedia emission characteristics by unit process and by unit
operations are  identified.   Twenty-three large-volume SOCMI  (Synthetic Organic
Chemicals Manufacturing Industry)  unit process components  are considered.
(These are basic synthesis steps such as halogenation and  alkylation.)  Each
unit process is  described with available release data.  Data are presented on
controlled and uncontrolled releases from  process,  storage and handling, and
fugitive sources to air,  land,  and water.


                                     A-33

-------
 P141

 Versar.   1982.   Appendix  B,  information  resource matrix,  methodology for
 assessing occupational  exposure  to  chemical  substances.   Versar,  Inc.,
 Springfield,  VA.   Prepared  for U.S.  Environmental  Protection  Agency, Office
 of Toxic  Substances,  Washington,  DC.   EPA  Contract No.  68-01-6271,  Task 10.

           This  matrix is  similar to the  one  in  Appendix B of  Freed  et al.
 (September 1983).   Information resources included  are  numeric data  bases,
 bibliographic retrieval data banks,  standard reference manuals, encyclopedias,
 journals,  and books.   Photocopies of pages of source documents and  of litera-
 ture from the data base producers comprise much of the contents of  this
 document.

 P142

 Flores GH.  1983.   Controlling exposure  to alkylene oxides.   Chemical  Engi-
 neering Progress 79(3):39-43.

           Peak  exposure levels are  given and control measures are described
 for reducing  ethylene oxide  and  propylene oxide exposure  during the  following
 activities:   process  sampling, tank  car  sampling,  and  tank car disconnect  in
 process areas;  laboratory analysis  and cylinder transfer; and maintenance.

 P143

 Dunn DW,  Johnson ML,  Hedley  WH,  Pate JB, Barrett GJ, McKinnery WN, Jr.   1983.
 Control practices  at  formaldehyde production  plants.   Chemical Engineering
 Progress  79(3):35-38.

           Preliminary industrial hygiene/control technology surveys  were done
 at  11 formaldehyde  production  plants.  The plants  represent eight companies
 and  a cross-section of the industry on the basis of size, geography,  and pro-
 cess  (catalyst  type).  Walkthrough surveys began with  the methanol receiving
 area, continued through production, and  concluded with the formaldehyde  stor-
 age  and shipment areas.  Personal sampling revealed exposures  from 0.01  to
 2.4  ppm (|jg/m3) formaldehyde.  Personal  protective  equipment  is generally used
 during the  loading  of formaldehyde--a top-fill  operation  using a dip  tube into
 the  tank truck.   Other plants  use ventilation and  scrubbing to protect  the
 workers during  loading.  Automatic loading devices  are common.

 P144

 Tomb TF, Treattis HN.  1983.   Review of  published experimental calibrations
 performed on the 10-millimeter nylon cyclone.   Pittsburgh, PA:  Mining  En-
 forcement and Safety Administration, Pittsburgh Technical  Support Center.
MESA-IR-1040.

          Research  is reviewed on calibration of the 10-mm Dorr-Oliver  nylon
cyclone to recommend an operational  flow rate that will give  size-selection
characteristics  most closely approximating the Atomic Energy Commission (AEC)
                                     A-34

-------
 criterion  for  respirable  dust.   The  experimental  scatter  precludes  recommend-
 ing  any  flow rate  within  a  tolerance >  0.2  L/min.   A  flow rate  of 2.0  L/min
 during operation of  the cyclone  will  closely  approximate  the  AEC criterion.

 P145

 Peach MJ,  Myers WR.   1981.   Field  research  report on  the  performance of  MSA
 powered  air  purifying respirators  on silica flour baggers at  Illinois  Mineral
 Company, Elco, Illinois.  Morgantown, WV:   National Institute for Occupational
 Safety and Health.   NTIS  No.  PB83-258012.

           The  Mine Safety Appliances  (MSA)  Powered Air-Purifying Respirator
 (PAPR) was evaluated in the  laboratory  and  in the field when  worn by two bag-
 gers at  the  two-spout bagger  and when placed  in a fixed location minus the
 facepiece.   Size distributions of  suspended particulates  from the two-spout
 bagger were  determined in an  area  4  to  8 ft from  the  baggers.   Silica  concen-
 trations in  air outside the  facepiece of the  respirators  worn by the two bag-
 gers ranged  from 4.365 to 41.68/mg/m3 over  a  3-day period.  The maximum  con-
 centration observed  in air  inside  the facepiece was 1.479 mg/m3; 5  of  11 air
 samples  from inside  the mask  exceeded the calculated  threshold  limit value  of
 0.291 mg/m3.   The  Illinois Mineral Company  was advised to discontinue  use of
 the  MSA  PAPR units until  the  design  problems allowing particulate penetration
 inside the mask are  corrected.

 P146

 Schroy JM.   1981.  Prediction of workplace  contaminant levels, pp.   190-206.
 Symposium  Proceedings on  Control Technology in the  Plastics and Resins Indus-
 try, Atlanta, GA,  February 27-28,  1979.   Cincinnati, OH:    National   Institute
 for  Occupational Health and Safety.  NIOSH  Publication No. 81-107.

           Vapor emission  rate data are  used to estimate volatile losses  from
 equipment, piping, and open pools  of pure substances when  monitoring data are
 unavailable.   The maximum concentration of a volatile chemical at some dis-
 tance downwind from  the leaking  source  is related to the  emission rate,  the
 ventilation  rate, diffusion coefficients, and the molecular weight  of the
 chemical  of  concern.   The relations of  fugitive emission  rates of acryloni-
 trile and  vinyl chloride  to workplace air concentrations  are  graphically
 presented.

 P147

Volkwein JC.   1979.  Dust control  in bagging operations,  pp.   51-57.   ACGIH
Symposium  Proceedings on  Industrial Hygiene for Mining and Tunneling,  November
6-7, 1978.   Cincinnati, OH:   American Conference of Governmental Industrial
Hygienists.

          The Bureau of Mines identified emission sources  from fluidized-type
bag  filling machines at silica sand processing mills and recommended control
measures.  Dust was emitted while filling the bag.  The Bureau of Mines devel-
oped an improved ventilation hood that encloses the bag and nozzle  area and
                                     A-35

-------
 exhausts the dusty air to a dust collector.   Dust from spillage of material
 from nozzle and bag sleeve as  the bag leaves  the machine was  controlled by
 delaying release of the bag and then forcing  material  remaining in the nozzle
 into the bag by a low-volume blast of compressed air.   Other  improvements  were
 needed.   Dust adhering to the  bag surface (which can  become airborne during
 conveying and loading) was not satisfactorily controlled by air brushing.
 Dust emissions due to  poor bag quality (inadequate closure, broken bags, and
 leaky seams) could be  improved by administrative control  over bag  quality.
 Background dust levels from other dust operations were also important sources
 of  respirable dust exposure.   The bagging operation should be isolated,  lo-
 cated far from other dusty operations,  or fresh  makeup air should  be supplied
 to  the baggers.

 P148

 Bureau of Mines.   1978.   Dust  control  hood for bag-filling machines.   Technol-
 ogy News  No.  54,  Technology Transfer Group.   Washington,  DC:   Bureau of Mines,
 U.S.  Department of the Interior.   Revised June 1983.

           The Bureau of Mines  hood for  controlling silica sand  dust  during
 bagging  is  described.   A typical  hood  has 4 ft2  of open  surface  area and re-
 quires an air velocity of 800  ftVmin.  Preliminary tests showed dust concen-
 trations  by up to  90%;  tests with  SF6  tracer  gas  showed  that  the hoods can be
 operated  with 100% efficiency  in  reducing air contamination.

 P150

 Volkwein  JC,  Cecala  AB,  Thimons  ED.   1983.  Use  of foam  for dust control in
 minerals  processing.   Washington,  DC:   U.S. Department of the  Interior.  Re-
 port  of  Investigations,  Bureau of  Mines,  8808.

          Total  respirable  dust  reductions of  80  to 90% were achieved by mix-
 ing  foam  (water  plus surfactant) generated by  compressed air with whole-grain
 silica sand.   Samples  were  taken under  the screw  conveyor, at the top of the
 elevator, at  the chute  to  the conveyer, and at the  feeder.

 P151

 Volkwein  JC,  Thimons ED.   1983.  Ventilation  hoods  to reduce dust during bag-
 filling.  Mining Equipment  International.

          The  Bureau of Mines developed an improved ventilation  hood  for
 fluidized valve-type bagging machines to  capture  silica sand dust close to
 the generation source  and  to minimize the amount of air needed.  The  same
 cut away  diagram of the hood that appears in Bureau of Mines (1983)  is ac-
 companied by  an exploded view of the hood in quarter sections.

 P152

Bureau of Mines.  Undated.  Technical information for dust control  ventilation
hood for bag picking machines.   Draft report.   Washington, DC:  Bureau of Mines,
U.S. Department of the  Interior.
                                     A-36

-------
           The  Bureau  of  Mines  hood  developed  for  silica  sand  dust  control  dur-
 ing  bagging  surrounds the  filling bag  so  that exhaust  ventilation  air  is  drawn
 through  the  open  front of  the  hood,  over  the  bag  surface,  and out  through  duct
 work at  the  rear  of the  hood to  a dust collection system.   This  reference  con-
 tains  much of  the same information  as  in  Volkwein and  Thimons (1983).   Diagrams
 show a typical  dust hood installation  and duct placement for  a split type
 weighing arm.

 P154

 O'Leary  DT,  Richter KM,  Hi 11 is PA,  Wood PH, Campbell SE.   1983.  Methodology
 for  estimating  environmental loadings  from manufacture of  synthetic organic
 chemicals.   Athens, GA:  U.S.  Environmental Protection Agency, Environmental
 Research Laboratory.   EPA-600/3-83-064, NTIS  No.  PB83-241331.

           A  methodology  is presented for  estimating the  multimedia environ-
 mental loadings for a "new" chemical,  in  the  absence of  manufacturing  plant
 emission data.  The methodology  draws  on  an environmental  release data  base
 that contains multimedia environmental  loadings for structurally similar com-
 pounds that  undergo similar process  (physical  and  chemical) unit operations.
 The  data base  is  integrated with other pertinent  available  data on the  manu-
 facturing process  of  the new chemical  such as  (a)  physical  and chemical prop-
 erties of process  feedstock, products, and by-products;  (b) reaction stoi-
 chiometry, thermodynamics, and reaction kinetics;  (c)  process  flow diagram
 and  process  mass  balance;  (d)  location and composition of environmental re-
 leases and method  of  disposal; (e)  process environmental control technology
 including performance; (f) process  storage and handling  requirements; and (g)
 plant  equipment components (in numbers and classes).   In practice, sufficient
 direct data  are rarely available for estimating the environmental loadings of
 the  compound under review; the methodology has been designed with this  reality
 in mind.    In every case,  where data deficiencies are likely to occur, alterna-
 tive means are suggested for filling the  data gaps.  The methodology integrates
 all  pertinent data to enable the user to  estimate multimedia  (controlled and
 uncontrolled) environmental loadings under the classifications of storage and
 handling, process  and fugitive emissions,  respectively.  An example is pro-
 vided  to  demonstrate  the methodology's applicability.   [Author abstract]

 P155

 Callahan  JM.   1978.   Refueling emissions-oil  versus auto co.'s.  Automotive
 Industries 158:14-15.

          EPA was expected to require control  of automotive refueling emis-
 sions by  fume collection  onboard the car or at the gas station.  For each
 gallon of gasoline, 5.3 g vaporized hydrocarbons are released during refuel-
 ing.   The collector in the car would be a  large canister containing charcoal.
Three oil companies have  developed prototype  automobile collectors.   Fumes
collected at the gas station could be done by the method used in California.
Vapor is   returned to the  station's storage tanks through special  pump nozzles
and hoses.  The emissions are recycled by  forcing them from the storage tank
to a tanker truck while it replenishes the station.
                                     A-37

-------
 P156

 Cahners.  1983.  FMC gains accuracy,  speed with  new drum,  bagline.   Cahners
 Publishing Company, Denver, CO.   Packaging Engineering 29:92-94.

          Automated bag and drum  filling of microcrystalline cellulose at  FMC's
 food and pharmaceutical products  division is described.

 P157

 Schroy JM.  1985.  Dynamic conditions in filling drums.  Presented at the
 1985 Summer National Meeting of the American Institute of  Chemical Engineers,
 August 1985, Seattle, WA.

          Drum filling is described as a non-steady state  process involving
 both vapor diffusion and convective gas flow.  The process of vessel filling
 and consequent emission rate is defined in terms of the vaporization and
 emission processes and their time dependent relationships.  Large barge load-
 ing tests are used to check a model for organic vapor losses from clean drum
 filling.

 P158

 EPA.  1984.   A manual for the preparation of engineering assessment.
Washington,  DC:  U.S. Environmental Protection Agency, Office of Toxic
 Substances.

          This manual is an aid for the chemical engineer  in the Chemical
 Engineering Branch who is responsible for evaluating occupational exposures
and environmental releases under the Toxic Substances Control Act.

P159

Versar, Inc.   1982.  Methodology for assessing occupational exposure to chem-
 ical substances.   Draft final.   Versar, Inc., Springfield, VA.   Prepared for
U.S. Environmental  Protection Agency, Office of Toxic Substances.  Contract
No. 68-01-6271, Task 10.

          This report describes methods of estimating,.workplace.concentrations
and identifies exposed populations.

P160

JRB Associates.  1984.   Draft data element list for the comprehensive assess-
ment-information rule.   JRB Associates, McLean, VA.   Prepared for U.S.  Environ-
mental  Protection Agency, Washington, DC.   JRB Contract 2-813-975-13, EPA
Contract 68-01-67C9.

          This report identified federal  information collection instruments
used in conducting chemical assessments.   The purpose was to develop a com-
prehensive list of the data elements required to conduct chemical assessments
at EPA and other federal  agencies.


                                     A-38

-------
 P161

 Norwood  SK.   1984.   Estimating  employee  exposures  from  continuous  air  moni-
 toring data.   Dissertation  submitted  to  the  University  of Oklahoma.

           This study investigates  the  relationship between exposure  estimates
 based on personal  samples and exposure estimates based  on area  samples taken
 by  a continuous  air  monitor (CAM).


 VI.  PACKAGING ASSOCIATION  PHONE SURVEY  APPROACH AND RESULTS

           Various  trade associations  related to the packaging industry were
 contacted for  the  packaging characterization study.  The following type of
 information was  requested.

           •  Types of containers and  sizes commonly used in chemical packaging
             industry (packaging of semi bulk chemicals).

           •  Data  on the types  and sizes of containers  produced in this coun-
             try,  and the percentage used in the chemical industry.

           •  Product loss during packaging operations using various types of
             containers (this includes steel drums, fiber drums, plastic
             pails, bags, etc.).

           •  Any information on emission factors from packaging processes
             (same question as  above).

           •  Other sources  of data.

           From this telephone survey, it was concluded  that the information
 sought was not readily available.   The information which is available  is
 either strictly for association members,  or it is not specific enough to
 characterize packaging operations.   In general, 55 gal   steel,  plastic,  and
 fiber drums; 5 gal plastic and  steel  pails;  and 50 Ib multilayer kraft and
 plastic bags comprise the majority of the packaging in the chemical industry.
 No specific breakdown was found of what percentage of these packages are used
 in the chemical industry or any other breakdown relating to the types of
 chemicals  (e.g., volatile,  liquid,  granular, fine powder, etc.).  Product
 loss in packaging operations varies from 0.1% to 1% of the production depend-
 ing upon product value and other characteristics.   If the product is expen-
 sive, then more effort is put in to control  the product loss  during packaging
 and in other operations.   Table A-6 provides a summary of contacts  and  infor-
mation received.
                                     A-39

-------
                             Table A-b.  Summary of Contacts and Information Received
Name of association
  contact person
   phone number
                    Summary of conversation
*Fiber Drum Technical  Council
Mr.  Lawrence W.  Bierlein
(202) 342-5250
Continental  Fiber Drum
Mr.  P.  White,  Sales
(203) 964-6717
Stanford,  Connecticut
'"Packaging Institute
Ms.  Carol  Boyle
(212) 687-8874
Mr.  Bob Temper
Mallinckrodt,  St.  Louis, Missouri
(314) 895-2057
He is a one man operation.  They only handled technical aspects of
fiber drum industry.  They do not keep any statistics.  He referred
me to two companies which make up the largest segment of the  fiber
drum industry.  These companies are Continental Fiber Drum and
Grief Brothers.

Mr. White, Head of Sales, was not available.   In talking with
staff members it was found that the information inquired is
very sensitive.  They cannot disclose, especially considering a
very limited number of companies make up the whole industry.
The packages prevalently used in chemical industry are 55, 40,
and 30 gal.  steel drums, 5, and 15 gal.  steel and plastic pails,
100, 80, and 50 Ib multilayer craft and plastic bags, 18 kg
pails, 25 kg bags, 200 and 50 kg drums.

Ms. Carol Boyle was not available.   I talked to a staff member.
They do not have the inquired information.   He referred me to
Mr. Bob Temper of Mallinckrodt, who is the Chairman of the Chem-
ical Packaging Committee for the Packaging Institute.

He does not have the information and referred me to the Chairman
of the Bag Committee/Plastic Drum Institute and the Steel Shipping
Container Institute.  According to him,  product loss information
during filling of bagging operations would be hard to find.   This
type of information is kept within the plant and only line super-
visors might keep records of this.   He thinks maybe the Chemical
Manufacturer Association can help us.
                                                                                         (continued)

-------
                                              Table A-6.  (continued)
Name of association
  contact person
   phone number
                    Summary of conversation
Packaging Institute
Mr. Warner 0.  Fleming (Du Pont)
Chairman - Bag Committee
(302) 774-0188
*Paper Shipping Sack Manufacturing Assoc.
Mr. Brent C.  Dixon
(914) 723-6440
Packaging Institute
Mr. Dan Barber (Container Corporation
  of America)
Chairman
Plastic Drums Committee
(302) 573-2605

*Steel Shipping Containers Institute
Mr. Mai Anderson
(201) 688-8750
 His committee did not have the information.  He refered me  to  the
•Paper Shipping Sack Manufacturing Assoc.  According to him  product
 loss or "giveaway" during filling and bagging operations can run
 from 0.1 to 1% depending upon the product.  It will be very hard
 to find any numbers on product loss and its relationship to sizes
 and types of containers.  Also, the more expensive the product,
 the less product loss.

 Their association keeps the types of statistics we are interested
 in but they do not distribute this information outside of the
 association, it is for their members use only.  In general, he
 said, valve bags are used more than open mouth bags and typical
 sizes are 50 and 25 Ib bags.  Again, weight depends upon the
 density of the products.

 He sent us his company's product literature.  This covers a
 majority of the type of containers used in the chemical industry.
 His product manager will call me to discuss other information.
According to Mr. Anderson, 55 gal. drums are used  largely  in  the
chemical industry.  Other sizes in use are 30 and  16 gal.  drums,
and 5 gal. pails.  Department of Commerce, Bureau  of Census,  com-
piles various statistics on steel shipping containers using pro-
duction  information from them into a M 34K report.  The  1981
report shows about 17 million steel drums out of 41 million pro-
duced in this country were used in the chemical industry.  This
number includes all sizes and varieties.  For more information
he refered me to Mr. Malcolm Burnhart of Bureau of Census.
                                                                                          (continued)

-------
                                              Table  A-6.  (concluded)
Name of association
  contact person
   phone number
                    Summary of conversation
U.S. Dept.  of Commerce
Bureau of the Census
Mr. Malcolm Burnhart
(301) 763-2518
   "Plastic Shipping  Container  Institute
     (Plastic  Pail*)
   Mr.  Robert  Hultquist
   (312)  368-4500

j^  Society of  Plastic Industry
M  Mr.  Robert  Sherman
   (312)  297-6150

   "Center for Packaging  Education
   Mr.  Bob Goldburg
   (212)  624-6157

   "National Barrel and Drum Association
   Ms.  Pam Terry
   (202)  296-8028
"Packaging Machinery Manufacturers Institute
Mr.  Claude Breeden
(202) 347-3838
Bureau of the Census does not keep the  specific  information which
we are seeking.  They put out. reports on total shipments  of steel
pails and steel drums.  Ho will send us the  1982  and  1983 reports.
These reports do not have any useful information  for  our  purposes.
In his view, no governmental agency keeps the type  of information
In which we are interested.

They do not keep the type of information we  are  looking for.   He
refered me to Mr.  Robert Sherman of the Society of  Plastic
Industry.
                                                 The type of information which we  seek  is  not  kept  by them.
They do not have the information but they can collect  it  for  us
on a fee basis.  He thinks the study may take about 6  months  to  1  year.


Their members are recyclers of used steel drums.  They usually
recondition about 45 to 50 million 55 gal. drums (about 2.5 times
the new steel drum production).  According to her, based  on Census
Bureau's 1980 statistics about 39% of steel drums are  used in
the chemical industry.

Their members build machinery for the Packaging Industry.  They
do not have any information on packaging containers emission
factors related to individual types of machinery.
"These names were obtained  from the "Encyclopedia of Associations."

-------
            APPENDIX B





REPORTS ON ADDITIONAL PLANT SURVEYS

-------
 INTRODUCTION

           Efforts  to  gain  plant access  for characterization of industrial
 drumming  and bagging  operations were  successful  in  gaining approvals  to visit
 five  plants  in  the Kansas  City  area.  At  these  facilities, the following
 operations were observed by  MRI personnel:

           1.  Manual  top filling of 55-gal  drums with  industrial  solvents.

           2.  Manual  top filling of 5-  to 55-gal drums  with automotive
 chemicals.

           3.  Bagging of detergent powders  into  sacks,  cartons and  drums.

           4.  Bagging of pharmaceutical feed mixes  into 50-lb  bags.

           5.  Manual  top filling of open  55-gal drums with adhesives.

           6.  Bagging of Portland cement  into 94-Ib bags.

           Details  of  the five plant visits  are presented below.


 Date  of Survey:  November  5, 1984

 Packaging  Operation:   Drumming  and filling  of automotive chemicals, fuel and
 oil additives,  cooling system cleaners, conditioners and sealers, and cleaning
 compounds.

 Filling Apparatus:  Manual top  filling of 5-, 30- and 55-gal drums  from  inter-
 ior 250-gal  tanks  and exterior  6,000-gal  tanks.   Gravity filling and pump
 filling through a  1-1/2 in.  hose.  Rate estimated at 20 gal/min.  Detergent
 powders filled  in  1-lb sacks, 50-lb cartons and 400-Ib  drums from overhead
 auger into hopper  for sacks  or  down chute for cartons and  drums.

Worker Locations/Movement:    Liquid filling  of 6 cans at a  time  (up  to 1  gal).
 Lids put on  by  hand or moved to  machine for sealing.  Worker uses finger to
 test when  55-gal drum filled.  Worker monitors scale for filling of powder
 into boxes or drums.  Automatic  operation of 1-lb bagger allows worker to
move away  from  bagging area.

Cycle Times:   Liquids into 55-gal drums at *• 20 gal/min.   Small batches of
< 20 drums.  Powders boxed or drummed at ^ 50 Ib/min for large containers.

Ventilation/Control Devices:   None for liquid filling.   Problems exist mostly
because of no open doors or windows and because of the tank cleaning opera-
tion.   Humidity is controlled in solid filling area.  Filters used  in air
circulation of powder filling room.

General  Comments:  Little ventilation  of liquid filling area.  Cleaning of
previously used tank (for 1 pt filling)  caused heavy fumes.
                                     B-l

-------
 Date  of  Survey:   November  5,  1984

 Packaging Operation:   Fifty-five gallon  filling  (also  truck  filling)  of  min-
 eral  spirits,  toluene, xylene,  isopropanol, acetone, methylethyl  ketone  in
 well-ventilated open  building (to  be partly enclosed for winter).

 Filling  Apparatus:  A 2  in. hose is connected to one of many faucets  coming
 from  large outside  tanks and  is first  flushed of previous chemical.   Barrel
 is moved into  position,  ground  cable attached, scale zeroed,  pump  on,  valve
 on, barrel filled,  previous barrel sealed, pump off, valve off.

 Worker Locations/Movement:  Single worker moves filling hose from  one  barrel
 to next  after  shoving filled  barrel down rollers.  When pump is turned on and
 hose  valve is  opened,  worker  seals previously filled barrel.  Ground  cable
 attached to barrel  rim during filling.

 Cycle Times:   One barrel filled every 45-47 s.  (Toluene filled at one
 barre 1/min.)

 Ventilation/Control Devices:  A 5  in.  x 10 in. inlet draws air from the  bung
 area.  The bottom edge of  inlet is about 1-1/2 in.  above the  bung.  Vapors
 exhausted to atmosphere above building.

 General  Comments:   When chemical is changed, hose is flushed  into chute with
 open  lid.  Hose end is placed here to drain between filling  operations.  In
 winter,  hot air draft  is projected toward worker area  from behind right  rear
 shoulder of worker; air inlet is on heater bottom.


 Dates of Survey:   November 6,  1984 and November 9,  1984

 Packaging Operation:   Bagging of pharmaceutical  feed mixes into 50-lb  bags.
 Four or  five bagging machines  (one had no hood).   Worming medication bagged;
 also bagging of antibiotic products in rice hull  or other carriers.

 Filling Apparatus:  Hopper discharges  solid material down chute into bag
which drops slightly while filling, emitting puff of particulate.   Bag moved
 on conveyor to sealing operation.

Worker Locations/Movement:   Worker sets filling time,  trickle time for correct
weight, checks first bag, takes samples for analysis.   Glass shields worker's
 face from chute area.

 Cycle Times:   7 to 8 bag/min  (50-lb bags).

Ventilation/Control Devices:   Horizontal  inlet (rectangularly shaped) posi-
 tioned to rear of drop chute.

General Comments:   Observed bagging of 40% (by wt)  antibiotic (dusty).  Dump-
 ing of material for blending is a dusty operation,  done in three-sided enclo-
sure with overhead hood.   Oil  is blended  into  mixtures  to control  dust.  Plant
has personnel  dust sample data taken at least  twice per year.


                                     B-2

-------
 Date of Survey:   November 13,  1984

 Packaging Operation:   Manual  top filling of creamy white liquid adhesive into
 open 55-gal  drums.   Viscosity  of 1,500 cp.

 Filling Apparatus:   Large 65-barrel  batch is cooked,  then gravity fed through
 3 in.  hose to pump  system with two long bag filters,  then through 2 in.  hose
 to open 55-gal  drum lined with poly  bag.   Approximately 15 drums arranged in
 semicircle around pump.   No splattering or barrel  bubbling seen.

 Worker Locations/Movement:  Worker moves hose from barrel  to  barrel.   New
 drum is tipped  under  rim of drum currently being  filled before  nozzle is
 shifted.   No spillage seen even though pump not shut  off between barrels.
 No shutoff valve  on nozzle end.

 Cycle  Times:  About 85 s to fill  one 55-gal  drum.   While drum is being filled,
 previous  drum is  covered with  plastic,  lidded and  sealed using  a rim  ring
 which  is  tightened  with  its own lever (60-drum batches).

 Ventilation/Control Devices:   No  windows  or door  seen.   Fans  (exhaust?)  are
 located over each large  heated mixing tank.

 General Comments:   When  changing  adhesives,  large  tank  and pump  are cleaned
 with hot  water  flush.  When pump  flow decreases, worker unbolts  pump  filter,
 removes long filter bags and discards in  open cardboard barrel.   Another
 worker removes  filled barrels  with a hand truck and replaces  with empty  bar-
 rels.   Only  slight  smell  discernable.


 Date of Survey:   November 13,  1984

 Packaging  Operation:   Four-spout  St.  Regis  packer  for 94-Ib bags  of Portland
 cement.

 Filling Apparatus:  Cement  is  gravity-fed from  a hopper into  paper-plastic
 bag  after  bag inserted onto outlet.   When bag  is filled, worker  releases bag
 which  falls  back  on large mesh  conveyor for  delivery to  automatic  palletizer.

 Worker  Locations/Movement:  Worker sits on  horizontally  movable bench and
 places bags  in succession on four hopper  spouts.  Button above each outlet
 controls start of cement  flow.   Foot  control  releases bag  to  fall  back on
 conveyor.  Valved bag  is  self  sealing.

 Cycle Times:  Estimated at about 10-15 bags/min.

 Ventilation/Control  Devices:  Hood on packer to bag house controls  cement
 dust released during bag filling, but does not control conveyor emissions.
 Filling operator wears space helment connected to air supply.

General Comments:   Conveyor is meshed, allowing spilled cement to be collected
 underneath by screw conveyor.   Major points of emissions are conveyor drop
points and where conveyor moves around corners.  Lots of doors open in summer.
One  large door open  during survey.

                                     B-3

-------
      APPENDIX C






LABORATORY PILOT STUDY

-------
                                  APPENDIX  C

                              TABLE  OF  CONTENTS


                                                                          Page

 I.         Introduction	      C-l

 II.        Experimental Facilities and Apparatus  	      C-l

               A.  Flow Tunnel	      C-l
               B.  Plume Generation  Apparatus  	      C-3
               C.  Sample Collection Apparatus	      C-3
               D.  Wind Speed Monitors	     C-ll
               E.  Photography	     C-ll

 III.       Experimental Methods	     C-ll

               A.  Sampling Procedures	     C-ll
               B.  Analysis Procedures	     C-14
               C.  Data Reduction and Analysis Procedure	     C-15

 IV.        Results	     C-16

               A.  Methane Experiments	     C-16
               B.  Ammonium Chloride Experiments	     C-16

V.        Model Development 	     C-16

               A.  Dispersion Coefficients Calculated from Methane
                     Experiments	     C-19
               B.  Particulate Dispersion Results 	     C-30

VI.       Procedures for Calculation of Maximum Concentrations. .  .   .     C-32

VII.      References	     C-38
                                     C-i

-------
                                 APPENDIX C

                               LIST OF FIGURES


Figure                              Title                                Page

II-l      Overhead view of pull-through flow tunnel	     C-2
II-2      Overhead view of release configurations 	     C-4
II-3      Bung releases from 55-gal drums (Configurations 2 and 3).  .     C-5
II-4      Ammonium chloride plume generation apparatus	     C-6
II-5      Sampling arrays for methane dispersion experiments	     C-8
II-6      Sampling arrays for ammonium chloride dispersion experi-
            ments—tunnel centerline releases 	     C-9
I1-7      Sampling arrays for ammonium chloride dispersion experi-
            ments—tunnel side releases	    C-10
II-8      30-Second time exposure of ammonium chloride dispersion
            at 0.5 m/s (1.1 mph)	    C-12
II-9      30-Second time exposure of ammonium chloride dispersion
            at 1.2 m/s (2.6 mph)	    C-13
V-l       Comparison of measured versus circular normal  modeled
            concentrations for Experiment 7 	    C-21
V-2       Comparison of measured versus circular normal  modeled
            concentrations for Experiment 8 	    C-22
V-3       Comparison of measured versus circular normal  modeled
            concentrations for Experiment 9 	    C-23
V-4       Comparison of measured versus single-point circular-
            normal concentrations for Test 22-A (4.8 mph wind
            speed and 2 m downwind from the source)	    C-24
V-5       Comparison of measured versus circular normal  modeled
            concentrations for Test 22-B (4.7 mph wind speed and
            4 m downwind from the source)	    C-25
V-6       Comparison of measured versus circular normal  modeled
            concentrations for Test 22-C (4.7 mph wind speed and
            4 m downwind from the source)	    C-26
V-7       Comparison of measured versus modeled concentrations at
            1.05 m above tunnel  floor for Experiment 22  (4.8 mph) .  .    C-27
V-8       Comparison of measured maximum concentrations  (hooded
            drum tests) with two models for the maximum  values.  .  .  .    C-31
V-9       Comparison of photographically determined vertical  spread
            of ammonium chloride plume with Turner dispersion
            coefficients	    C-33
VI-1      Methane concentrations above background level  and  with
            blank criterion applied 	    C-37
                                    C-iii

-------
                                 APPENDIX C

                               LIST OF TABLES


Table                               Title                                Page

II-l      Ammonium Chloride Particle Size Fractions 	      C-7
IV-1      Experimental Conditions for Laboratory Pilot Study of
            Methane Dispersion	     C-17
IV-2      Experimental Conditions for Laboratory Pilot Study of
            Ammonium Chloride 	     C-18
V-l       Empirically Determined Dispersion Coefficients Determined
            from Methane Experiments	     C-20
V-2       Comparison of Measured and Modeled Concentrations of
            Methane	     C-29
V-3       Comparison of Measured and Modeled Maximum Concentrations
            of Ammonium Chloride	     C-32
VI-1      Diffusion Function Reported by Schroy (1981)	     C-36
                                    C-v

-------
I.  INTRODUCTION

          This section describes the initial laboratory pilot study performed
to test dispersion theories and to develop efficient sampling strategies for
use in the plant scale occupational exposure study.  The laboratory pilot
study utilized a walk-in, low-speed flow tunnel which was constructed at MRI's
Deramus Field Station.

          Dispersion patterns of gases and suspended solids were studied for
single and multipoint sources released into free flow and obstructed flow en-
vironments.  Twenty-five experiments of gaseous dispersion and four experi-
ments of particulate dispersion were conducted.  The observed dispersion pat-
terns were compared with existing theories and then used as the basis for a
modified dispersion algorithm.

          The presentation is organized by subject area as follows:

          •  Section II describes the experimental facilities and apparatus
             used to characterize dispersion of packaging emission simulants
             at low flow speeds.

          •  Section III presents the experimental procedures for emission
             generation and for sample collection and analysis.   Section III
             also describes techniques for data reduction and analysis to
             quantify tunnel flow patterns, simulant concentration distribu-
             tions, and empirical dispersion coefficients.

          •  Section IV summarizes the dispersion experiments and presents
             the experimental results.

          •  Section V describes the preliminary dispersion model developed
             in this study and compares measured concentrations with the
             predictions of this model  and of other available models.

          •  Section VI presents calculation procedures for the simple mass
             balance model, the Schroy (1981) model, and the circular normal
             model.
II.   EXPERIMENTAL FACILITIES AND APPARATUS

          This section describes the experimental facilities and apparatus
used to study dispersion of gases and suspended solids.

     A.  Flow Tunnel (see Figure II-l)

          A pull-through flow tunnel (2.1 m in height x 2.7 m in width x
9.8 m in length) was constructed in MRI Field Station Building No.  1 to
gather basic data on dispersion patterns as a function of release point con-
figuration and flow velocity approaching the release point(s).   Building
No.  1 is 8.5 m x 15.2 m in length with a 3 m wall height.   A 2.1 m x 2.7 m
garage door opening centered on the west wall of the building was used for
the wind tunnel entrance.   The building exhaust fan drew air through the tun-
nel  at mean flow velocities from 0.5 to 2.2 m/s (0.9 to 4.9 mph).

                                     C-l

-------
 I
ro
                                                                                    Reducing section at tunnel exit was-

                                                                                    added after methane experiments
                                •1 m
                                                                                                                1.3m
                                     Figure  II-l.   Overhead  view of  pull-through flow tunnel.

-------
           One side wall  and the tunnel  ceiling were constructed of 6 mil
 plastic sheeting mounted on a 2 in.  x 4 in.  wood framework.   The other wall
 of plyboard was  painted  flat black for  both  visual  and photographic monitor-
 ing of particulate plume dispersion.  In addition,  a plexiglass window was
 framed in  the tunnel  wall  facing the black plyboard to allow clear viewing  of
 the tunnel  interior during an experiment.  For photographic  work,  No.  8 white
 thread was  vertically suspended down the tunnel  centerline at 10 cm dis-
 tances.

           In order to reduce the effects of  external  wind on the flow through
 the tunnel,  a special  tunnel  entrance and exit were built.   A 2.1  m x 2.7 m x
 0.6 m  flow  straightener  constrained  ambient  winds entering the tunnel.  In
 addition, a lattice work was  placed  in  front of the straightener to improve
 the uniformity of low wind flows by  reducing the propagation of external at-
 mospheric turbulence  into  the tunnel.   Also  a reducing section was attached
 to the exit of the tunnel  so  that the flow was drawn  through a centrally lo-
 cated  exhaust area of size 1.1 m x 1.4  m.

     B.  Plume Generation  Apparatus

          For gaseous  dispersion experiments,  natural  gas was  released  at a
 monitored rate (Matheson 602  rotameter)  from a Fisher burner to enable  thor-
 ough mixing  of the gas with  air.   The single point  release configurations in-
 cluded:  tunnel  centerline point release into  unobstructed flow, point  re-
 lease  near  a wall,  point release past a  sidedraft hood, and  bung release from
 a  55 gal drum with and without a close-fitting hood.   Figure II-2  shows an
 overhead view of the  eight release configurations.  Figure II-3 illustrates
 the two bung releases  (Configurations 2  and  3  in Figure II-2).   The  double
 point  releases were used to check on  the additivity of plumes.

          Ammonium chloride was  the  simulant for particulate dispersion
 modeling.  This  compound was  formed  by bubbling  air through  two Smith-
 Greenburg impingers containing  concentrated  HC1  and NH4OH solutions,  and by
 combining the  resultant  vapor  flow streams.  The two  nozzles which discharged
 vapors from  the  chemical solutions into  the  tunnel were positioned 3/8  in.
 apart  and faced  each other for  good mixing of  the vapors and consequent pro-
 duction of NH4C1  (see Figure  II-4).   The  airflow rates through  the  impingers
 were monitored with Matheson  (602 and 604) rotameters.  The  white  ammonium
 chloride particles could easily  be seen  to disperse downwind in the  tunnel.
 Particles of  ammonium chloride  formed in  this  manner were mostly smaller than
 10  urn  in aerodynamic diameter, considered the  upper limit for  respirable par-
 ticles.  Table II-l presents the  size distribution of  ammonium  chloride par-
 ticles as measured by a  Climet particle  analyzer.  Approximately 63% of the
 mass consisted of  particles less  than 0.5 pm diameter.  Particles of this
 size range disperse in much the  same manner as gases.

     C.  Sample  Collection Apparatus

          1.  Natural Gas

          For the  sampling of natural gas in  air, polypropylene sample lines
were located at  several  fixed positions  on a  movable two-dimensional sampling


                                     C-3

-------
Tunnel'
Wall   '

          O

                          'Side Drafti
                          "Booth  ;
                          .-** _«v-. . . • ^,-*i
                                              ¥
 Figure  II-2.   Overhead view of release configurations.
                            C-4

-------
                        Gas Source
                                  Free Standing Drum
                        Gas Source
                                  Drum with "Local Hood"
Figure II-3.  Bung releases from 55-gal drums
          (Configurations 2 and 3).
                    C-5

-------
Figure I1-4.   Ammonium chloride plume generation apparatus.
                           C-6

-------
                     Table  II-l.   Ammonium  Chloride  Particle
                                    Size  Fractions
                                   Average  no.        Percent  .
                   Particle         of particles      total mass
                 diameter  (urn)       samples
0.3
0.5
1.0
3.0
5.0
10.0
- 0.5
- 1.0
- 3.0
- 5.0
- 10.0
- > 10.0
272,000
10,700
3,080
374
96
31
62.6
6.8
7.9
8.6
6.1
7.9
                Determined by Climet particle analyzer at
                .7.1 L/min.
                 Calculated based on assumed spherical shape
                 of particles.


array (Figure II-5) and routed to a common manifold connected to the analyti-
cal instrument.  During sampling, each line was opened in sequence by a valv-
ing system controlled by a 24 pole rotary switch connected to a 30 s timer.
This sampling method allowed for rapid switching from sample point to sample
point every 30 s.

          A Beckman 402 total hydrocarbon analyzer was used to determine nat-
ural gas concentrations in air.  This unit utilizes a flame ionization detec-
tor (FID) to quantify total hydrocarbon (THC) levels down to 1 ppm.  The sam-
pling rate was approximately 3 L/min.   An EPSON HX-20 microcomputer and an
interface built around a Wintek microprocessor produced printouts of inte-
grated 10 s methane concentrations.

          2.   Ammonium Chloride

          Respirable particles of ammonium chloride were collected using a
network of standard respirable dust samplers (37 mm filter cassettes with
Dorr-Oliver 10 mm nylon cyclone preseparators).   The cyclones trapped parti-
cles larger than about 10 urn aerodynamic diameter.   Each particulate sample
was collected on a Millipore® filter with 0.2 urn pores.   A standard flow rate
of 2.0 L/min through the filter cartridge was ensured throughout substrate
buildup by maintaining sonic flow through a critical orifice.   A pressure drop
exceeding 0.53 x 760 mm Hg through a 20 gauge needle established the required
flow.   Separate sampling arrays were used for center and side releases,  as
shown in Figures I1-6 and I1-7.
                                     C-7

-------
o
00
                                                                                          •  Source

                                                                                          O  Sampler

                                                                                          ^> Wind Flow

                                                                                          -f-  Anemomaler
                            Figure  II-5.   Sampling  arrays  for  methane  dispersion experiments.

-------
                                                                 %-V Wind Flow


                                                                  I   Anemometer
Figure II-6.  Sampling  arrays for ammonium chloride  dispersion experiments--
                tunnel  centerline releases.

-------
o

o
                                                                                           ler
                                                                                       Wind Flow
                                                                                    ~]~ Anemometer
                    Figure II-7.  Sampling  arrays for ammonium chloride dispersion  experiments—
                                    tunnel  side releases.

-------
      D.   Wind Speed Monitors

           Kurz warm wire anemometers  were  used to  monitor the tunnel  flow.
 Tunnel  centerline  wind speeds  were  monitored at 1  m downwind from the en-
 trance  of the tunnel  and 0.46  m upwind of  the sampling plane for both gaseous
 and  particulate dispersion  tests.   Previously established wind speed  profiles
 were used to  construct flow distributions  from the mean centerline velocity
 measured  during a  given test.   Data from the anemometers were presented on
 both the  Epson HX-20  and the Kurz digital  wind speed display panel.   The
 Epson computer also calculated and  listed  mean wind speeds for user-selected
 time intervals.

      E.   Photography

           Ammonium chloride plume dispersion patterns  were photographically
 recorded  with a single lens reflex  camera  at different wind speeds and  for
 different release  arrays.   Time exposures  of 35 mm black and white film (ASA
 32)  effectively documented  the time-averaged plume shape for a given  set of
 experimental  conditions  (Figures II-8  and  II-9).


 III.  EXPERIMENTAL METHODS

      A.   Sampling  Procedures

           1.   Methane

           Prior  to  each  day  of natural gas dispersion  experiments,  the
 Beckman 402 was  heated  for  1 h, then calibrated with a  reference gas.    In the
 absence of a  0  ppm  THC gas,  the zero end of  scale  was  determined by blocking
 all  input  to  the Beckman.   The  Fisher  burner was adjusted  for  maximum mixing
 of natural gas with air.  A  rotameter  between  the  gas  source and the Fisher
 burner was used  to  adjust and  monitor  the gas  release  rate.

           Before each  test  the tunnel   flow was  set  for  the  desired centerline
wind  speed.  Wind  speeds from  the warm wire anemometers were monitored on
 Kurz  digital displays and simultaneously on the Epson HX-20.  Once the sam-
pling array was  in  the proper  position, testing began with  activation of the
rotary switching device and starting of the Epson HX-20  "TEST" program.
Three separate tests, each with the sampling plane at different distances
downwind of the source (5, 7, and 9 m  from the tunnel entrance), constituted
one dispersion experiment.

          During testing, each  sampling tube on the array supplied a sampled
air/methane mixture to the Beckman 402 analyzer for a period of 30 s.   The
THC concentration from the first 10 s, as the sampling tube was flushed and
the Beckman adjusted to the new THC value,  was discarded.  Two 10 s readings
(in ppm v/v) were then obtained before the  unit automatically switched to a
new sampling position.
                                     C-ll

-------
o
I
ro
                        Figure  II-8.  30-Second time exposure of ammonium chloride dispersion  at

                                       0.5 m/s (1.1 mph).

-------
o

OJ
                       Figure  II-9.  30-Second time exposure of ammonium  chloride  dispersion at
                                       1.2 m/s (2.6 mph).

-------
           2.   Ammonium Chloride

           Before each test the sampler holders were fitted with new filter
 cartridges.   With the vacuum pump turned on momentarily,  each cartridge was
 checked for  a 2.0 L/min flow rate using a GCA RDM-101 rotameter.   If a car-
 tridge failed to meet a ±10% flow rate tolerance,  that filter (usually in a
 cracked cartridge) was designated a blank,  and a new cartridge assembly in-
 stalled.   (Other blanks were also selected, as necessary,  for test quality
 assurance.)   Then a leak check was performed on the entire system using a
 mercury manometer.

           Before each test of approximately 1 to 1-1/2 h  duration,  fresh
 solutions  of NH4OH and HC1  were cooled in an ice bath surrounding the two
 Smith-Greenburg  impingers.   The production  rate of  NH4C1  was  adjusted using a
 bypass valve on  the air pump supplying the  impingers  and  by separately ad-
 justing the  airflow to each impinger.   After the tunnel flow  was  set to the
 desired centerline wind speed,  testing was  initiated  with  the activation of
 the  air pump.  Three separate tests, with the sampling planes at  5,  6.5,  and
 8.0  m  from the tunnel  entrance,  constituted one dispersion experiment.

           During a test,  the generation rate of ammonium  chloride tended to
 decrease with time.   Consequently,  the bypass valve on the air pump  supplying
 the  impingers was  adjusted  to allow for a relatively  constant mass  rate of
 particulate  formed.   The  production of heat from ammonium  chloride  formation
 was  also diminished at lower generation rates.   As  a  test  progressed  and  the
 chemicals  in  the impingers  became  less  concentrated,  the pump bypass  valve
 was  opened to increase ammonium  chloride  production.

           During a  test,  the ammonium  chloride  which  agglomerated on  the  noz-
 zles was removed at regular intervals  using a water spray.  The downwind
 anemometer, which was  exposed to NH4C1  particulate, was also  washed  at  regu-
 lar  intervals  with  distilled water.

           A test was  aborted if  tunnel  wind speeds  were subject to signifi-
 cant ambient  wind variability.   Consequently, all particulate  testing was
 done during evening hours when calm ambient winds prevailed.

           Black  and white photographs  of  plume  dispersion were taken with  a
 35 mm  SLR  camera positioned  on the  side wall  at  1.06 m above  the  floor  and
 1 m downwind  from the  release point.  The field  of  view from  this point was
 about  1 m  at  the tunnel centerline.  Later  a  Plexiglas window was constructed,
 and the camera position moved both away from  the tunnel and downwind to allow
 a field of view  of  1.5 m at  the tunnel  centerline.

     B.  Analysis Procedures

           1.   Methane

          Methane concentrations (10 s averages) measured by the Beckman 402
 instrument were printed by the Epson HX-20 during each test.  Methane concen-
 trations were used  directly  for calculation of gaseous dispersion coeffi-
 cients after subtraction of  background values and elimination of values not
meeting a blank criterion.

                                     C-14

-------
           2.   Ammonium Chloride

           At  the end of each participate  dispersion test,  the cartridges with
 the Millipore® filters were removed from  the  sampling array,  stoppers re-
 placed on both the inlets  and outlets,  and the  cartridges  returned  to the
 laboratory in a protective container.   After  every  second  test,  the cyclones
 were washed with distilled water  and air  dried.

           The chloride analysis of  each particulate sample was initiated by
 removing  the  filter from the cartridge  and placing  the filter (exposed side
 up) in a  bottle with a diameter slightly  larger than the filter.  Ten milli-
 liters of H20 was added.   The bottle was  capped and then placed  in  an ultra-
 sonic bath for 5 min.   Two milliliters  from each bottle was transferred to  a
 disposable sample cap for  analysis  using  the  Technicon Autoanalyzer.   A sec-
 ond 2 ml  aliquot was taken from each filter bottle  for replicate  analysis.

           After analysis of every tenth sample,  a distilled H20 blank was
 analyzed.   Every 2 h standard chloride  solutions (100 and  200 ug/mL)  were
 analyzed.   If high concentrations forced  the  analyzer off  scale,  a  diluted
 sample of the filter solution was prepared and  analyzed.   The technique de-
 scribed above is commonly  used for  water  quality analysis  and will  detect
 chloride  ion  down to 4 ug.

      C.   Data Reduction and Analysis  Procedure

           1.   Tunnel  Flow  Patterns

           Ten-second wind  speed averages  from the upwind anemometer were  av-
 eraged to  characterize the  overall  centerline tunnel  flow  for each  natural
 gas  test.   One-minute  wind  speed averages  were  averaged over  an entire  par-
 ticulate  dispersion  test period.  This  single test wind speed  value was ad-
 justed to  sampling plane wind  speeds  using wind  speed  traverse factors.
 These  factors  correlated centerline wind  speeds  measured at a  point 1 or  2 m
 from  the  tunnel  entrance to wind speed  values at each  of nine  points  in each
 sampling  plane.

           2.   THC  Concentrations

           THC  concentrations were input into a BASIC computer program for av-
 eraging of  the 4  to  12 replicate THC values obtained at each  sampling point
 during  a  test.   These  final THC concentrations were then plotted by the com-
 puter  on a  representation of the array grid.  The THC concentration data
 plotted for each  test  included raw THC averages, standard deviations, and
 numbers of values.   Three arrays of THC averages were plotted—raw values,
 values with a  blank  correction applied to them,  and values  from which the
 background had been  removed.  The blank correction consisted of excluding THC
 values of  less than  5 times the standard deviation of blank values.

          3.  Chloride Concentrations

          The Technicon Autoanalyzer was used to determine  NH4C1  (as Cl")
present in the filter solution.  The NH4C1 concentrations were plotted on a
 representation of the sampling array for analysis of the plume shape.


                                     C-15

-------
 IV.   RESULTS

           This  section  identifies  and  describes  the  dispersion  tests  per-
 formed  both with  methane  gas  and ammonium  chloride particulate.

      A.   Methane  Experiments

           Table IV-1  lists  the  testing conditions for  Experiments  1 through
 26.   Methane was  the  dispersion simulant for  all except the  last experiment,
 which was performed with  propane.   Figure  II-2 illustrates  the  release  con-
 figurations described in  Table  IV-1.   Tunnel  centerline wind  speeds measured
 1 m  from  the tunnel entrance  ranged from 0.42 to 2.2 m/s  (0.93  to  4.9 mph).

      B.   Ammonium Chloride  Experiments

           Ammonium chloride particulate was released from a single point
 source  in Experiments 27  through 30.   Each of the three tests comprising a
 single  experiment are listed  in Table  IV-2 together with  respective experi-
 mental  conditions.  Dispersion  was  studied for wind speeds  in the  range of
 0.50  to 1.6 m/s (1.1  to 3.7 mph) and for both tunnel centerline releases and
 tunnel  side releases.
V.  MODEL DEVELOPMENT

          The model proposed for the simplest release patterns in the current
program is based on the bivariate Gaussian distribution of concentration that
is commonly used in modeling of outdoor air pollution sources.  The general
form of this distribution is fairly complicated, but several simplifying as-
sumptions have been made for the present application.  First, it is assumed
that the distribution may be written as the product of a function of the lat-
eral crosswind distance times a function of the vertical crosswind distance.
This is the form taken by the distribution when applied to outdoor sources.

          The second assumption is based upon the configuration of the test
chamber.  Because the chamber has roughly a square cross-section, it was as-
sumed that the lateral and vertical dispersion coefficients are identical.
Under these assumptions, the model takes the form:

          C(y,z) = A exp {- £ [(y-yQ)2 + (z-zQ)2]}                   (1)


where:     C = modeled concentration (ppm v/v)

          A = maximum modeled concentration (ppm v/v)

          a = dispersion coefficient (m)

        y,z = lateral  and vertical distances from chamber centerline (m)

      V0»Z0 = location of maximum modeled concentration (m)
                                     C-16

-------
      Table IV-1.  Experimental Conditions for Laboratory Pilot Study
                     of Methane Dispersion
Date Experiment Release Nominal tunnel
no. configuration centerline velocity

5/30/84
5/29/84
5/29/84
5/30/84
5/30/84
6/4/84
6/5/84
6/5/84
6/5/84
6/5/84

6/5/84

6/5/84

5/30/84
5/30/84
5/30/84
6/5/84
6/4/84
6/4/84
6/4/84
6/4/84
6/4/84
6/6/84

6/6/84

6/6/84

6/27/84
1
2
3
4
5
6
7
8
9
11

12

13

14
15
16
17
18
19
20
21
22
23

24

25

26a

Single point centerline release
Single point centerline release
Single point centerline release
Drum centerline release
Drum centerline release
Drum centerline release
Stacked drum centerline release
Stacked drum centerline release
Stacked drum centerline release
Single point release with in-
serted side-draft hood
Single point release with in-
serted side-draft hood
Single point release with in-
serted side-draft hood
Single point side release
Single point side release
Single point side release
Double point centerline release
Double point centerline release
Double point centerline release
Double point side-by-side release
Double point side-by-side release
Double point side-by-side release
Single point release with flush
side-draft hood
Single point release with flush
side- draft hood
Single point release with flush
side-draft hood
Single point centerline release
(m/s)
0.42
0.99
1.7
0.44
0.77
2.2
0.50
0.87
2.1
0.50

0.87

2.1

0.43
0.79
1.6
0.52
1.0
2.1
0.53
1.1
2.1
0.50

0.90

2.0

0.8
(mph)
0.93
2.2
3.8
0.99
1.7
4.9
1.1
1.9
4.7
1.1

1.9

4.7

0.97
1.8
3.6
1.2
2.3
4.6
1.2
2.4
4.7
1.1

2.0

4.5

1.8
Test aborted; wind speeds ranged from 1.3 to 2.1 mph.
                                   C-17

-------
       Table  IV-2.   Experimental  Conditions  for  Laboratory  Pilot  Study
                      of Ammonium Chloride
Date
Experiment Release
no." configuration

7/18/84
7/18/84
7/24/84
7/19/84
7/23/84
7/19/84
8/2/84
7/31/84
7/31/84
7/30/84
7/30/84
7/30/84

27A
27B
27C
28A
28B
28C
29A
298
29C
30A
30B
30C

Single point center line release
Single point centerline release
Single point centerline release
Single point centerline release
Single point centerline release
Single point centerline release
Single point side release
Single point side release
Single point side release
Single point side release
Single point side release
Single point side release
Nominal tunnel
centerline
(m/s)
. 0.65
0.63
0.50
1.6
1.5
1.4
0.55
0.56
0.52
1.6
1.6
1.6
velocity
(mph)
1.4
1.4
1.1
3.6
3.4
3.2
1.2
1.2
1.2
3.5
3.7
3.7
 Position A was 5 m into tunnel; B was 6.5 m into tunnel; C was 8 m into
 tunnel.
          Other models are also currently under consideration; especially, for
more complicated release configurations (such as the side draft hood).  These
models include a Gaussian distribution similar to the above, but w'ith differ-
ent dispersion characteristics in the lateral and vertical directions (which
has been found useful in the preliminary analysis of the particulate tests),
and a distribution represented by an elliptical paraboloid.

          The nonlinear normal equations resulting from a least-squares cri-
terion applied to Eq. 1 (after taking the logarithm of each side) can be
linearized using the following change of variables:
In A =


I/a2 =
                         4a,
                  a2

                  a3/2a2

                  a4/2a2
                                     C-18

-------
 With this  change  of variables,  the  normal  equations  become:


   Nr    I(y2+z2)    Sy         iz

         I(y2+Z2)2   Z(y3+Z2y)  Z(y2Z+Z3)

                    Iy2        lyz

     "Symmetric                iz2

 where:     N = number of  data points

         y,z = lateral and vertical distances (m) from tunnel centerline,
              respectively

           C = measured concentration (ppm  v/v) at the point y,z


 The  modified normal  equations are easily solved for  each location and each
 experiment using a  Gauss-Jordan elimination algorithm for simultaneous linear
 equations.  Results  are  presented in the following sections.

     A.  Dispersion  Coefficients Calculated from Methane Experiments

           Experimentally determined concentrations were input into a BASIC
 computer program for determination of dispersion coefficients at 2, 4, and
 6 m  from the source(s).

          The circular bivariate normal form was applied to the symmetric re-
 lease configurations (such as the single point source and the drum with a
 hood).   Experimentally determined dispersion coefficients for the methane
 tests are reported  in Table V-l.  Figures V-l through V-3 compare the model
 results  for Experiments 7 through 9 (involving the hooded drum) with the mea-
 sured concentrations.

          Results from the simple release configurations can be used to es-
 timate the concentration profile for more complicated configurations.   For
 example, Figures V-4 through V-6 compare the measured concentrations for Ex-
 periment 22 (double point release) with the superposition of the single point
 model developed on the basis of Experiment 3.   Figure V-7 presents a view of
 concentrations measured in Experiment 22.

          Because the filling of 55 gal drums  with a close fitting hood/fill
mechanism may be fairly common for liquids emitting organic vapors,  the fol-
 lowing discussion will  be limited to the model  results for Experiments 7
through 9.

          In order to estimate the spatial  variation of methane concentration
at downwind stations other than  2,  4,  and 6 m  (see Figures V-l  through V-3),
the regression coefficients  obtained for each  downwind array were themselves
fit to power law models  with downwind  distance  (x)  and wind speed (U)  as the
 independent variables.
                                     C-19

-------
      Table V-l.
Empirically Determined Dispersion Coefficients
  Determined from Methane Experiments
Expt.
1


2


3


4


5


6


7


8


9


17


18


19


Distance
from
source
(m)
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
Maximum concentration,
Cmax (pPm v/v)
Measured
100
44
21
55
19
15
53
20
11
55
24
14
55
22
17
18
29
-
20
7.8
7.2
18
7.6
4.5
15
5.1
5.1
88
62
31
110
33
18
88
27
16
Modeled
—
28
13
36
16
-
47
16
7.5
81
*
*
-
27
*
21
*
-
15
7.7
6.4
13.5
5.7
4.4
18.5
7.1
4.8
40.7
49.1
28.3
100
7.3
13.3
70.1
10.4
13.2
I/O2
(m-2)
.
2.4
0.41
11
1.3
-
17
3.9
1.3
9.1
ik
*
-
7.6
*
6.8
*
-
1.6
1.3
0.19
3.4
2.0
1.2
8.8
4.4
2.3
1.8
4.2
1.6
35
3.8
4.4
14
7.7
7.2
Location of max .
concentration (m)
yo
.
-0.063
-0.18
-0.032
-0.10
-
-0.036
-0.15
-0.15
-0.12
*
*
-
-0.055
*
-0.20
*
-
-0.095
-0.15
-0.48
-0.058
0.0063
0.047
-0.12
-0.21
-0.035
0.056
-0.012
0.037
-0.037
0.083
0.19
-0.001
-0.10
-0.17
Z0
.
0.15
-0.12
-0.040
0.076
-
0.065
0.074
0.13
-0.55
*
*
-
-0.70
*
-0.52
*
-
-0.15
-0.58
-1.6
-0.12
-0.22
-0.33
-0.29
-0.26
-0.38
-0.15
0.19
-0.11
0.020
0.039
-0.038
0.22
-0.013
0.041
  Imaginary dispersion coefficent.
- Not enough valid data to model.
 Using the bivariate Gaussian model  described in Section VI for the
.simple release patterns.
 Maximum concentration locations are offset distances from tunnel
 centerline.
                                  C-20

-------
                     2m
2m
2m
          Drum
          and Hood
2.7m
                                                    10   20
         Plan View
                                  0   10   20
                                     ppm
                  0 Measured Concentrations

                 — Circular Normal Modeled
                    Concentrations
         Figure V-l.  Comparison of measured versus  circular normal  modeled
                         concentrations for Experiment 7.
                                       C-21

-------
                     2 m:
 m
•2 m v-
          Drum
          and Hood
2.7 ni'
et --
                            o
                                                                 , O
         Plan View
                                   'A J!°L
                                      ppm]
                  0  Measured Concentrations

                 — Circular Normal Modeled
                    Concentrations.
         Figure V-2.  Comparison of measured versus circular normal modeled
                         concentrations  for Experiment 8.
                                       C-22

-------
                     2m
2m	\-
2m
           Drum
           and Hood
2.7 m
         Plan View
                             0   5   10
                                ppm
                   Measured Concentrations

                   Circular Normal Modeled
                   Concentrations
         Figure V-3.  Comparison of measured versus circular  normal modeled
                         concentrations  for Experiment 9.
                                       C-23

-------
    Wall:
-Wall'
                           t

              ——— Measured Concentration

                     Single -Point Modeled Concentration
Figure V-4.   Comparison of measured versus single-point circular-normal
         concentrations for Test 22-A (4.8 mph wind speed and
                    2 m downwind from the source).
                                 C-24

-------
                                 
-------
                                 
-------
 Wall-
                                          2 m Downwind
                                              4 m
• Wall
                         I	  _t
                          Release Points
Figure V-7.   Comparison  of measured versus modeled concentrations
   at 1.05 m above  tunnel floor for Experiment 22.  (4.8 mph)
                              C-27

-------
 The  correlation matrices are:

                              Ax         U
                         A    1    -0.95    -0.036
                         x           10
                         U                     1


 for  the maximum modeled concentration A, and:

a
X
U
a
1


X
0.58
1

U
-0.72
0
1
for the dispersion coefficient a.  From these matrices, the downwind distance
appears to have the greater influence on maximum concentration while wind
speed is apparently more important in determining the dispersion.  The power
fits were found to be:

          A = 32 x"1-0

          a = 0.48 x°-65 u"°-55
where the caret indicates the estimated value.  Thus, Eq. (1) as applied to
the hooded drum Experiments 7 through 9 may then be written as:
          C = 32 x"1-0 exp
'—  V *  i')'
(2)
where:    C = modeled concentration (ppm v/v)

          x = downwind distance (m)

          U = wind speed (mph)

        y,z = horizontal and vertical  distances (m) from plume centerline,
              respectively


          Table V-2 compares the maximum concentrations measured during the
simplest methane release tests with:   (a) the maximum values modeled by
Eq.  (1); (b) the maximum values estimated by the Schroy (1981) model; and
(c)  the steady-state concentrations assuming the common mass balance model,
                                     C-28

-------
             Table V-2.  Comparison of Measured and Modeled
                          Concentrations of Methane
Expt.
1


2


3


4


5


6


7


8


9


Distance
from
source
(m)
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
Maximum concentration, C (ppm v/v)
(TlaX
Meas.a
100
44
21
55
19
15
53
20
11
55
24
14
55
22
17
18
29
NA
20
7.8
7.2
18
7.6
4.5
15
5.1
5.1
Modeledb
NA*
28
13
36
16
NA
47
16
7.5
81
NA
NA
NA
27
NA
21
NA
NA
15
7.7
6.4
14
5.7
4.4
18
7.1
4.8
Schroy model
8.4
5.8
4.8
3.6
2.4
2.0
2.1
1.5
1.1
8.0
5.6
4.3
4.6
3.2
2.5
1.6
1.1
0.90
7.2
5.0
4.0
4.0
2.9
2.3
1.7
1.2
0.94
Uniform mixing
model
16
16
16
6.8
6.8
6.8
3.6
3.6
3.6
14
14
14
8.7
8.7
8.7
3.1
3.1
3.1
14
14
14
7.7
7.7
7.7
3.2
3.2
3.2
    = not available.
.Maximum concentrations are not necessarily on tunnel  centerline.
 Using Equation 1 (circular normal  model).
                                   C-29

-------
 as described by Clement Associates  (1981).   The Appendix presents  example
 calculations for these models.

           It is clear from Table V-2  that the  simple  mass balance  (uniform
 mixing)  model,  in addition to  being incapable  of describing  any spatial  vari-
 ation, grossly  underestimates  the measured maximum concentration at  the  2 m
 distance.   Although  the uniform  mixing  model more closely estimates  the  mea-
 sured maximum further downwind,  this  is the result of the small  mixing cross-
 section  of the  flow  tunnel.  Because  full-scale packaging facilities  would be
 located  in much larger room volumes,  the large underestimation  of  concentra-
 tion would persist throughout  the area  occupied by packaging equipment opera-
 tors.

           On the basis of  Table  V-2,  the Schroy (1981) model  also  appears to
 substantially underestimate the  maximum concentrations.   The degree of under-
 estimation lessens as  downwind distance increases,  as  the flow  rate de-
 reases,  or as more obstructions  are introduced into the  flow.   As  an  illus-
 tration  of the  last  point,  note  that  in Table  V-2 the  agreement  between  the
 Schroy model  and measured  maximum values are better for  the  hooded drum  (Ex-
 periments  7 through  9) than for  the simple  point  release  (Experiments 1
 through  3).   Figure  V-8 compares  the  maximum concentration values  measured
 for the  hooded  drum  tests  with estimates of the maximum  from both  the Schroy
 model and  Eq. (2).

           Schroy's model is  based primarily upon  the Gaussian dispersion
 model developed  for  outdoor  diffusion.   In addition, because  the source and
 receptor can  be  very close  to each  other in a  workplace setting, Schroy ap-
 peals to dispersion  coefficients  presented by  Bowne (1974) which again only
 concern outdoor  results  and  are basically extensions back from the minimum
 100 m source-receptor  separation  of the  Pasquill-Gifford curves  (Turner,
 1970).   From  Schroy1s  paper, it  is  not  clear whether rural or urban disper-
 sion coefficients are  included in the workplace model, nor is any  substanti-
 ation offered for his  claim  that  "at present Gaussian models  represented
 [sic] the  most practical method of  analyzing the  downwind impact of an emis-
 sion to an  [indoor] workplace."   Finally, no experimental evidence is pre-
 sented in  the paper to  justify the  approach.

           Thus,   on the  basis of the methane tests conducted to date,  it would
 appear that the  models  commonly employed to assess worker exposure (a) tend
 to substantially underestimate the maximum concentration (especially close to
 the source),  and (b)  are not capable of  describing the spatial variation  in
 concentration.

     B.   Particulate  Dispersion Results

          Table V-3 compares the measured maximum concentrations of ammonium
chloride (single-point center release) with the values calculated using the
circular normal  model  (Eq.  (2)).   The model predictions and the measured
values  compare favorably.
                                     C-30

-------
       30
       25
       20
    o
8  15
o
O
|
i
S  10
5
                    I           I
                       Experiment No.
                 \

             Schroy
             Model
             (1 mph)
            Schroy
            Model
            (2 mph)
                         7
                         8
                         9
• 1_mph
O 2 mph
• 5 mph
                           •Maximum Estimated
                              MRI Model
                               i
                    246

                    Downwind Distance (m)
                                                8
Figure V-8.  Comparison of measured maximum concentrations
              (hooded drum tests) with two models for
              the  maximum values.
                          C-31

-------
     Table  V-3.   Comparison of Measured  and Modeled Maximum Concentrations
                             of Ammonium Chloride


Expt.
27


28


Downwi nd
distance
(m)
2
3.5
5
2
3.5
5
Measured
maximum
(mg/m3)
75
12
9.2
110
48
14
Modeled
maximum
(mg/m3 )
70
11
11
130
77
5.0
Measured
dispersion
coefficients
ay
0.26
0.49
0.40
0.60
0.23
0.53

0
0
0
0
0
0
(m)
CT2
.31
.51
.43
.16
.22
.49
Locations
of
concentrations
>
0.
-0.
-0.
0.
0.
-0.
r
0080
064
093
27
18
0022

0
-0
-0
0
-0
-0
max
(m)
zo
.082
.033
.31

d




.0036
.14
.040


 Maximum concentration locations are offset distances from tunnel centerline
          The vertical plume spread was photographed at 1 m downwind for a
single point tunnel centerline release of ammonium chloride under different
wind speed conditions.  The plume height was visually estimated for several
different exposures at the same wind speed, averaged, and plotted in Figure
V-9.  These visible plume spreads were assumed to represent 4 a (4.3 ay would
encompass ~ 98% of the plume).   Turner (1970) indicates that the 4 oy disper-
sion coefficients for stability classes D, E, and F range from 0.2 to 0.4 m
(if Turner's curves are extended on log-log paper to 1 m from the source).
These figures compare well with results from the photographs (e.g.,  0.18 to
0.54 m).
VI.  PROCEDURES FOR CALCULATION OF MAXIMUM CONCENTRATIONS

          Three models which have been proposed for estimating worker exposure
to gaseous emissions are described in this section.   The first two models ad-
dress only maximum concentrations downwind of the source; the third (based on
the pilot laboratory study test results) provides a description of the spatial
variation in concentration.   The other major differences between all  models
in this section and that presented in Volume I of this report lie in  the omis-
sion of flow obstacles and worker position/orientation in the estimation of
exposure.

          1.   Mass Balance Model

          Assuming plug flow through the chamber with velocity U (in  mph),
the volumetric flow Q is:
          Q = (2.68 m) (2.12 m) (U (mph))


            = 9,140 U

where Q is in m3/h and U is in mph.
                                                                      (D

                                                                      (2)
                                     C-32

-------
o
 I
CO
oo
L1.0I
"^"
•^^
O
i
3
O.
"5
XI


Q.
"5
0)
>
^^
;o.8J
'e

'I
in ' »
' 5 • 0'6
co j •«"-•!
"»!
QCi
E;
* , .
^M I 1 J
^^ (
:E]
^j
i QJ - i
n o
L- - v.f.


^

—


_
Is" = 4l
\
\
X
>v
N|n = 4 ; Extrapolated from Turner's ;
X Workbook for Stability
\Tn = 2' Classes D1 E and F.. '
•^^^^^^ '
* ^ ^! (
^^~^~~-~-JLl • '
i i 1 I 1 I 1
0 1 (mph) 2 3
I | i I i i i i
b| 0.2,! 0.4; 0.6j(m/s) 0.8 .1.01 '1.2; '1.4
                                                       Wind Speed


                           Figure \J-9.  Comparison of photographically determined vertical spread
                                          of ammonium chloride plume with Turner dispersion
                                          coefficients.

-------
           The  steady-state  concentration  C    estimated  by simple  mass  balance
 model  (e.g., Eq.  (7)  on  page  9  ofClement Associates, 1981)  is:
           Css =  G/Q                                                    (3)

where:       G =  mass  emission  rate  (mass/time)

             Q =  ventilation  flow  rate  (volume/time)

           As an  example,  for Experiment  3 with an emission  rate  of  82.5  g/h
and a tunnel velocity of  3.8 mph:

             Q =  34,700 mVh

and, from  (3),

           r   =  82.5 g/h
           Ss    34,700 mVh

              =2.4 mg/m3

For methane, 1 ppm =  656 |jg/m3 and:

           Css =3.6 ppm

           2.  Schroy  Model

           The Schroy  (1981)  model for methane is:
                max
where:         C    = estimated maximum concentration at a downwind
                md*   distance d (ppm)

                  G = mass release rate (g/h)

                  U = velocity of air (fpm)

               -y- = function of distance d given in tabular form
For the same* example given above:

                  G = 82.5 g/h

                  U = 3.8 mph = 334 fpm

and, from Table VI-1, the following dispersion parameters are found:

                                           1
                             d          (27tayoz)


                            2 m         0.1020 m~2
                            4 m         0.07074 m"2
                            6 m         0.05507 m 2

                                     C-34

-------
 and substituted into  Eq.  (4).
 as  follows:
      Thus, the estimated maximum concentrations are
                             2  m
                             4  m
                             6  m

           3.   Circular  Normal  Model
                  max

                2.1 ppm
                1.5 ppm
                1.1 ppm
          As  an  example  of the circular  normal model, consider  the  2 m  down-
wind  station  for Experiment  3.   Figure VI-1  shows the concentration values
used  for modeling purposes.  Values of -9.99 are less than background concen-
tration levels or do  not meet a  blank criterion.  Because the spacing between
samplers on the  y (lateral)  axis  is 0.383 m  and 0.303 m on the  z  (vertical),
the normal equations  (see Section  IV) in this example are:
                                                     10.8
                                                      .793
                                                     -0.354
                                                     0.396
-5
0.477
0
-n
0.477
0.0599
0
n
0
0
0.293
n
o i
0
0
n i on -

a2(
a3j
          Note that zero or negative values of net concentration are not
used.  In this example, only five data points can be used.  The solution
vector is:
          3i = 3.75

          a2 = -16.6
a3 = -1.20

a4 = 2.16
          Thus, with a change of variables discussed in Section V, the fol-
lowing dispersion parameters are found:

             A = 47

          I/a2 = 17

            yQ = -0.036

            ZQ = +0.065

          These are the values given for the 2 m downwind station of Experi
ment 3 on Table V-l.   In practice,  a BASIC language program was used to gen-
erate the dispersion parameters reported here.
                                     C-35

-------
Table VI-1.   Diffusion Function Reported by Schroy (1981)
Distance,
d i
ft m
1.5
1.64 0.5
2.0
3.0
3.28 1.0
4.0
5.0
6.0
6.56 2.0
7.0
8.0
9.0
9.84 3.0
10.0
Value of
U/(2nayoz))
0.1458
0.1447
0.1420
0.1347
0.1326
0.1259
0.1166
0.1072
0.1020
0.09929
0.09288
0.08653
0.08120
0.08069
Distance,
d
ft m
12.0
13.12 4.0
14.0
16.0
16.41 5.0
18.0
19.69 6.0
20.0
25.0
26.25 8.0
29.53 9.0
30.0
32.80 10.0
35.0
Value of
(l/(27toyaz);
0.07431
0.07074
0.06844
0.06324
0.06217
0.05873
0.05507
0.05451
0.04912
0.04528
0.04188
0.04131
0.03789
0.03521
Distance,
) d
ft
65.6
98.4
131.0
164.0
196.8
229.7
262.5
295.3
328.1





m
20
30
40
50
60
70
80
90
100





Value of
(l/(2noyo-z))
0.01989
0.01255
0.00938
0.00793
0.00642
0.00553
0.00463
0.00426
0.00372





                           C-36

-------
 EXRERIMENT 3    LOCATION  A


          MEASUREMENTS < EXCLUDING  BACKGROUND)	

    N
    MEAN                   THC VEL
    STANDARD  DEU.









04 94
•1.99 3.90 -9.99 3.42
•f.9? D.20 -9.99 O.U









0
-9.99
-9.99
0
-9.99
-9.99
2
13.19
1.99
66 6
7.10 3.7* 53.34
4.24 0.17 15.54
3
3.57
3.10
0
-9.99
-9.99
0
-9.99
-9.99
4
3.91
0.43
4
3.70
0.20
1
*
3.83
0.10
4 44 04 0 4
3.44 2.82 3.95 -9.99 3.85 -9.99 3.S7
0.20 1.2? 0.34 -9.99 0.18 -9.99 ft.24
4
3.31
0.23
4
3.84
0.22
4
3.77
0.24
Figure VI-1.   Methane concentrations above background level  and
                with blank criterion applied.
                               C-37

-------
VII.  REFERENCES

Bowne NE.  1974.  Diffusion rates.  Air Pollution Control Association J
24(9): 832-835.

Clement Associates.  1981.  Mathematical models for estimating workplace
concentration levels:  a literature review.  Clement Associates, Inc.,
Washington, D.C.  Prepared for U.S. Environmental Protection Agency, Eco-
nomics and Technology Division, Washington, D.C.

Schroy JM.  1981.  Prediction of workplace contaminant levels, pp.  190-206.
Symposium Proceedings on Control Technology in the Plastics and Resins Indus-
try, Atlanta, Georgia, February 27-28.  Cincinnati, Ohio:  National Institute
for Occupational Health and Safety.  NIOSH Publication No. 81-107.

Turner DB.  1970 (Rev.).   Workbook of atmospheric dispersion estimates.
AP-26.  Environmental Protection Agency, Office of Air Programs, Research
Triangle Park, North Carolina.
                                     C-38

-------
     APPENDIX D





QUALITY CONTROL DATA

-------
           Quality  control  data  are presented  in  this  appendix  for  the  two
 plant  studies  and  for  the  laboratory  drumming study.


 A.   PLANT  A  DRUMMING TESTS

           The  internal  quality  control checks performed  in this test program
 included:  analyses of  field blanks,  spiked blank silica gel tubes, and
 analysis of  duplicate  (colocated) samples by  separate  laboratories.  In addi-
 tion,  a Century continuous hydrocarbon analyzer  was used to verify the very
 low  levels of  air  concentrations of Fluid 1 during Test A-l in comparison to
 the  background hydrocarbon concentration of 12 ppm as  propane.  The concen-
 tration of analyte on the  field blank was found  to be  below the limit of
 detection.   Recovery of analyte from blank tubes spiked with 5 ug of analyte
 was  128% for Fluid 1 and 123% for Fluid 2.  This indicated that good recov-
 eries  could  be obtained from silica gel.  Analysis of  randomly selected back-
 up portions  of two tubes for each analyte indicated levels beneath the limit
 of quantisation of the  method.

          The  results of the analysis of colocated samples by MRI and by the
 plant  laboratory are compared in Table D-l.    Total analyte concentrations from
 duplicate (colocated) samples agreed well except in the cross laboratory com-
 parisons for Test A-l.   However, the higher concentrations determined by the
 plant  laboratory appear to be inconsistent with  the consistently lower con-
 centrations  obtained by MRI at all sampling locations  for Test A-l.


             Table D-l.  Analysis of Duplicate (Colocated) Samples

                                       Analyte concentration (ug/m3)
                                                             Plant
                                            MRI            laboratory
         Test-sampling  location          analysis           analysis
Al-2a,b
Al-8a,b
Al-9a,b
Al-15a,b
A2-2a,b
A2-8a,b
A2-9a,b
A2-15a,b
< 72,
< 80
< 76
< 78,
< 184
< 162
< 216
< 154
< 70


< 80
, < 208


, < 197

150
180


< 210
240

          Two external quality control  performance audit samples were per-
formed using Fluid 2 supplied by the QCC.   The first sample showed that the
actual concentration of Fluid 2 in acetone (40 ug/mL) compared well  (103%
recovery) with the reported concentration (41 ug/mL).   The second sample of
25.4 ug/mL produced a reported value of 15.2 ug/mL,  a 60% recovery.
                                     D-l

-------
 8.   PLANT B BAGGING TESTS

           The internal  quality  control  checks  performed  in  this  test program
 included:   analysis of  field  and laboratory blanks  and analysis  of duplicate
 (colocated) samples by  separate laboratories.

           The analyte masses  found  on  five  field  blanks  were  each  less  than
 the  detection limit of  2  ug.  Lab blanks  also  yielded less  than  2  ug for each
 of five  blanks,  and 5 ug  for  the sixth  lab  blank.

           The results of  the  analysis of  colocated  samples  by MRI  and by the
 plant  laboratory may be compared in Tables  IV-9 through  IV-11 in Volume I  of
 this report.   The location  3b samples (suspended  particulate  and respirable
 particulate fractions)  from each test were  turned over the  the plant labora-
 tory for analysis;  the  colocated samples  collected  at locations  3a,  7a,  and
 7b were  analyzed by MRI.

           The external  quality  control  performance  audit samples consisted of
 blind  performance samples prepared  by the QCC.  Reported phosphate  concentra-
 tions  compare favorably with  the exception  of  the low level sample  (2.5  ug/mL)
 near the limit of detection.  The actual  concentrations  of  the four  blind
 performance audit samples are shown below.

                        Phosphate concentrations (ug/mL)
                                               Recovery
                        Actual       Found          (%)
0.0
2.5
5.0
10.0
< 2
1.6
4.7
9.3
_
64
94
93
C.  LABORATORY DRUMMING TESTS

          Two primary instruments were used to characterize analyte concen-
trations carrier gas (air speeds during the laboratory drumming tests.  The
Beckman 402 Total Hydrocarbon Analyzer detected total hydrocarbons emitted
from the drum.  Hastings air meter was used to measure wind speeds in the
flow tunnel.  The quality control data obtained for these - instruments are
presented below.

          1.  Beckman 402 THC Analyzer

          The Beckman 402 THC analyzer was calibrated each day with freshly
prepared ~ 8 L bag samples of chemical vapors.  This entailed the use of inert
dry gas, appropriately-sized and calibrated syringes, 10-L capacity Tedlar
bags,  and a calibrated dry gas meter with temperature gauge.   The gas standard
for the Beckman calibration was prepared using a known amount of the chemical
of interest which was injected into a known amount of hydrocarbon-free gas


                                     D-2

-------
 (N2).   The  bag  contents were  heated  and  then mixed  by  alternately  squeezing
 each  side of  the  bag  from  25-50  times.   The bag contents were  then sampled in
 order to calibrate  the Beckman 402 at the  initial span reading.  Subsequent
 checks  of this  standard and other prepared bag standards established  a  sensi-
 tivity  decay  function for  each day of testing.  Hydrocarbon  free (zero)  air
 was used to calibrate the  Beckman at the low end of the scale.  Table D-2 pre-
 sents the calibration and  calibration check data used  in the drumming tests.

          2.  Hastings Air Meter

          Calibration of the  Hastings air  meter under  low air  flows U  250 fpm)
 required that the probe be inserted  through a side  orifice into the center of
 a 59.5  cm long  (5.1 cm I.D.)  PVC tube.   The outlet  of  the PVC  tube was  con-
 nected  to the inlet of a Roots positive  displacement meter, allowing  the Roots
 meter to adjust the flow through the PVC tube.  The Roots meter is  known to
 read  low by 1£ at 2.5% of  maximum flow (125 ftVmin).  The following  Table D-3
 presents the  calibration data taken  on 8/13/85 with a  temperature  of  292 K and
 a barometric  pressure of 736.3 mm Hg.

          3.  Comparison of Integrated Mass Flux to  Drumming Emissions

          The concentrations measured in the sampling  plane at 3 m downwind
 from  the drum center  were  combined with  corresponding  wind speeds  to  produce
 point mass  flux values.   These were  integrated across  the sampling plane to
 produce a comparison  of total mass fluxes with drumming emission rates.   The
 results of  these comparisons are presented in Table  D-4.

          The emission rate of perchloroethylene was measured at 326  mg/s and
 is higher than each of the three mass flux values of 211, 199, and 255 mg/s.
 The mass flux values  are believed to be  low because  perc emissions consist of
 a large fraction of mist.   Significant perc mass is  suspected to have settled
 to the  floor  before reaching the 3 m downwind sampling plane.

          The methanol 3 m integrated mass fluxes are  all  higher than the
 measured drumming emission rate of 206 mg/s.   As noted in the dispersion and
 exposure test results, the bubbling of air through a half-filled drum cooled
 the methanol  to several  degrees Centigrade less than the temperature  of the
 ambient air.  Methanol vapors were consequently shown  to settle to the tunnel
 floor downwind of the drum.  The concentration values measured very close to
 the floor at the 3 m position were combined with fairly high wind speeds and
 thus produced significant,  yet probably unrealistic, contributions to the
 total  mass  flux.  In addition, the cooling of the drum contents may have pro-
 duced a lower molar volume for the drum exit gas than was  calculated  using
 the methanol  liquid temperature.   Both of these factors would tend to increase
 the calculated mass flux of methanol.

          Two other factors are known to influence the vapor emission rate -
vapor pressure and molecular weight.   When only these two  parameters are
considered,  vapor emissions of methanol  should exceed that of perc by a  factor
of 1.30.
                                     D-3

-------
Table D-2.   Daily Calibration of Beckman 402
Date
Time

Beckman 402
Attenuation Concentration
reading (ppm)
8/1/85

8/2/85

8/5/85


8/6/85

8/9/85
8/13/85

8/20/85



8/21/85


8/22/85

8/23/85

8/28/85


8/29/85


8/30/85

9/2/85





9/3/85



1109
1423
1632
1658
1046
1146
1635
1223
1331
1316
1448
1634
1011
1156
1437
1713
1735
2159
2226
1809
2048
1811
2345
1927
2147
2341
1606
1918
2244
1643
1950
1800
1810
1828
1950
2300
2331
1906
2223
2223
2341
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X10
X10
X10
X10
X10
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X1000
X100
X100
X5000
X5000
X5000
X500
X500
X5000
X10
X50
X50
X100
22941
21154
18338
18406
21044
19815 •
18451
20885
19449
18521
21411
107
87
84
87
81.5
103
84
104
104
95
101
79.3
889
779
816
1028
954
33423
1208
1208
104755
108870
107788
5238
4246
94728
50
29
29.8
29
Bag concen-
tration (ppm)
22941
22941
18338
18338
21044
21044
21044
20885
20885
18521
21411
107
87
87
87
87
103
103
103
104
104
101
101
889
889
889
1028
1028
33423
1208
1129
104755
104755
104755
5238
5238
94728
50
29
29
29
Corr. factor3
1.00
1.08
1.00
1.00
1.00
1.06
1.14
1.00
1.07
1.033/hr est.
1.033/hr est.
1.033/hr est.
1.00
1.04
1.00
1.07
1.00
1.27
0.99
1.00
1.09
1.00
1.27
1.00
1.14
1.09
1.00
1.08
1.033/hr est.
1.00
1.07
1.00
0.96
0.97
1.00
1.23
1.033/hr est.
1.033/hr est.
1.00
0.96
1.033/hr est.
                    D-4

-------
                          Table D-2. (Concluded)
Date      Time   	Beckman 402	   Bag concen-   Corr.  factor3
                 Attenuation   Concentration  tration (ppm)
                               reading (ppm)
9/11/85


9/11/85
9/13/85




1200
1203
1420
1421
1422
1137
1448
1449
1606
1607
2044
2045
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
57.9
43
1007
999
349
1000
939
321
319
946
885
273
57.9
57.9.
1000°
1000°
350 h
1000°
1000°
350^
350 c.
1000°
1000°
350C
1.00
-
1.01
1.00
1.00
1.00
1.06
1.09
1.10
1.06
1.13
1.28
 Bag concentration -=- Beckman 402 concentration reading; values of 1.00
bsignify initial span calibration was done.
 Empty wet drum "A" of ethylene glycol used to reproduce 9/11/85 calibra-
ction on 9/13/85.
 Empty wet drum "B" of ethylene glycol used to check 9/13/85 calibration
 against 9/11/85 calibration.
                                    D-5

-------
             Table D-3.  Calibration Data of Hastings Air Meter
Roots meter flow rate
(m3/min)
Actual
0.043
0.075
0.140
Corrected (STP)
0.0393
0.686
0.1281
Tube air speed (fpm)
At STP
63.5
111
207
+1% Correction
64.1
112
209
Indicated
air speed (fpm)
Hastings
72
115
200
Indicated/
actual
1.12
1.03
0.96
   Table D-4.  Comparison of Integrated Mass Flux With Drumming Emissions
Nominal
tunnel
centerline
velocity
Date
8/20/85
8/22/85
8/23/85
8/28/85
8/29/85
8/30/85
8/30/85
Times
1716-1822
1942-2047
2130-2252
2155-2338
2054-2209
1834-1950
2317-0018
(fpm)
200
200
100
200
200
100
100
Chemical
Perc
Perc
Perc
Methanol
Methanol
Methanol
Methanol
Fluid
temp.
(°C)
22
20
21
20
21
23
20
Mass

Raw
196
186
213
266
303
329
326
flux (mg/s)

Corrected
211
199 222
255
293
345 0/M
347 341
378
Emission
rate
(mg/s)

326



206

Based on checks of Beckman THC analyzer with bag standards.
                                     D-6

-------
            APPENDIX E






DATA FROM LABORATORY DRUMMING TESTS

-------
                                 APPENDIX E

                              TABLE OF CONTENTS

                                                                    Pages

Dispersion Tests

     Perch!oroethylene	     E-l thru E-16
     Methanol	    E-17 thru E-31

Emissions Tests

     Perchloroethylene	    E-32 thru E-47
     Ethylene glycol	    E-48 thru E-70
                                     E-i

-------
                .Tunnel Celling.
                                                              Chemical:  Pcrchlorcthy-
                                                                        lene
^
x;
^
>*
^
Left ;X
Tunnel^ 4
Wall X^
^
P
f?
^ ^ ' 	 ^ -^ ^ ,
1 meter , c
i

2
3 8 14 9 2
2 10 60 231 87 «
81 108
10 29 260 45 6
2 10 49 275 79 7 2
5 13 60 24 63
6 11 45 5 2
6 10 25 3 2
i 1 	 1 	 L 	 L. . \ . 	 \ „ , I . 1 1 ^ ^ J
>ung Post 1 ion: Side
ianpllng Plane: 1 KM; for
real Date: 8/22/85
Concentration (ppn)
•
^
^X Right
X- Tunnel
>^Wall
0
^
^
-2-, -.--a
                  Tunnel Floor

-------
ro
^
Left ^
Tunnel ^4
Wall 9S • l'Q\ • K08 • •
.97 .76 .56 .97 ^ .97^
.56 .18 .94 0
-1 	 * 	 ,-* 	 1 	 * 	 » 	 * 	 1 	 • > - k > ^ -
                                                                                                                    Chemicali   Perchlorothy-
                                                                                                                               Icne

                                                                                                                         PasItIant   Side

                                                                                                                    Sampling Planet   1  meter

                                                                                                                    Test Datei  8/22/85


                                                                                                                        Hind  speed  (m/a)
                                                                       Tunnel Floor

-------
                                             Chemical:  Pcrchlorethy-
                                                        lenc
Left
Tunnel -^ 4
Wall
1 meter ,
0
1 8
53
2 21 73
393
1 24 101 369
1 19 157 2'J7
1 34 111 223
1 26 128 223
!
<
1
1
10
137
101 1
128
". I I
105
79 6 0
108 9 1
45 5 1
1 i 1 I
                                             Bung Position:  Side

                                             Sampling Plnnci  1 meter

                                             (2 cfra fill rate)

                                             Test Date:  8/23/85


                                                Concentration  (pptn)
                                                  Right
                                              -<>- Tunnel
Tunnel Floor

-------
                                             Chemicali  Perchlorethy-
                                                        lenc
0
^
0
1
Left ^
Tunnel 
^
x-1
Xfj
i
. 1 mater ,
I I
i
	
	
.46 ^ .41^ .41
.18
• -5l • '°*« •**•
.04
.38 .04 .36
1111111144
Sampling Planet 1 meter
(2 cfm fill rate)
feat Datet 8/23/8S
Hind apeed (m/a)
%
^ Right
-^X Tunnel
X^X tJUnll
x^*" wail
0
%
^
^
Tunnel Floor

-------
cn
                                                                                                                   Chemical i  Perchloretliy-
                                                                                                                              lene

                                                                                                                   Bung Position:  Side

                                                                                                                   Sampling Planet  1 meter

                                                                                                                   (3 cfo fill rate)

                                                                                                                   Test Date:  8/23/05


                                                                                                                      Concentration (ppn)
                                                                      Tunnel Floor

-------
CT>
                                                                                                                  Chemicalt  Porchlarethy-
                                                                                                                             lene

                                                                                                                  Bung Positioni  Side

                                                                                                                  Sampling Planet   1 meter

                                                                                                                  (3 eCn fill rate)

                                                                                                                  Test Datei  8/23/85


                                                                                                                      Hind speed (m/s)
                                                                                                                         unnel
                                                                      unnel Floor

-------
 I
•vj
Left
Tunnel
Wall
                                                                     .Tunnel Celling)
, 1 meter ,

i
l
13
1 8 16 15
60
1 6 13 31 48
43 84
5 19 26 52 86
3 32 66 90 111
13 60 59 131 135
1 i i i k 	 U^-
— • — r s , scssrrrff Bu,
San
°.
0 1
2 1
32
22 2 1
92
41 5 0 1^
94
113 41 15 3
114 59 32 22
140 107 97 43
1111
                                                                                       Chemlealt   Perchlorethy-
                                                                                                  lene

                                                                                       Bung Posltlont  Side

                                                                                       Sampling Plane:   3 meters

                                                                                       Test Datei  8/23/85


                                                                                          Concentration (ppn)
                                                                                                                    ^X Right
                                                                                                                    x> Tunnel
                                                                       Tunnel Floor

-------
                                                                                                               Chemicali  Perchlorethy-
                                                                                                                          Icne
00
1
^
2
0
Left ^
Tunnel ^4
Wall ^
Ij.
-Xx
__ — — — — — — — rrrrrrrr
1 meter ,
•
Bung FOB it ton t Side
Sampling Plane i 3 meters
Fest Date: 8/23/85
Wind speed (n/s)
v •••••••••
• ••••••0*
.46 .51 .46

.43 .43 .46 .41 .43 .51
.38 .36 .30 .36 .41
1111111111

^
^ Right
-^ Tunnel
'^^^ Ufa II
^^^^ Wflll
^
^
                                                                  Tunnel Floor

-------
                                             Chemical:   Perchlorethy-
                                                        Icnc

                                             Bung PositIani  Side

                                             Sampling Planet  3 meters

                                             Test  Datei  8/22/8S



                                               Concentration  (ppm)
                                                  Right
                                              -Xx Tunnel
Tunnel Floor

-------
m
 i
 Left
, Tunnel
 Wall
                                                                                         Chemicali  Perchlorcthy-
                                                                                                    lene

                                                                                         Bung Posit loot  Side

                                                                                         Sampling Plane:  3 meters

                                                                                         Test Datei  8/22/85



                                                                                             Wind  speed (D/S)
                                                                                                                     Right
                                                                                                                 /x Tunnel

-------
Lelt    „
Tunnel o 4
Wall
                                            Chemicali  Perchlorethy-
                                                       lenc

                                            Bung Posltloni   Center

                                            Sampling Planet  1  meter

                                            Test Datei   8/20/B5



                                               Concentration (ppm)

                                                        •
                                                  Right
                                             X^ Tunnel
Tunnel Floor

-------
ro
                                                             Tunnel Celling'.
                          Chenlcalt  Pcrchlorethy-

i';^$$$^^
                          Bung Poaitloni  Center

                          Sampling Planet  1 meter

                          Test Datei  8/20/85


                              Wind speed (in/a)
                                                                                                            Right
                                                                                                             unnel
                                                             Tunnel Floor

-------
                                               Chemicali  Perchlorethy-
                                                         lene
-^x
xjS
^
^x
^X
x;6
x^
Left ;X
Tunnel ^4
Wall - Wall
^x
^
^
^
^
^x
Xx
^^xjjx;;
    Tunnel Floor

-------
       emnj, ^
^
^
%
^
Led ;X
Tunnel ^4
Wall '^x
g
^5
^
. . Background 1, 1
1 meter , sai
Te
1.00 l.OS 1.00
• • • • . • • • • •
• ••••••••
1.00 .92 1.05
• • • ••• • • • •
.34 .79
• ••••••••
.67 .67 .94
• ••a*****
• ••••••••
.34 .41 .64
.1 	 i._ 	 i * 	 i . .* . ,. 1 J , A . , * . ., j ., ; .. '
Chenlea1i  Perchlorethy-
           lene

Bung PositIoni  Center

Sampling Planei  2 metera

Teat Datei  8/20/8S


    Wind speed (m/s)
Tunnel Floor

-------

Left
Tunnel
Wall
1 meter
                                          ^^
                                                                                        Chemical:   Pcrchlorcthy-
                                                                                                   lene

                                                                                        Bung Position:  Center

                                                                                        Sampling Plane:  3 meters

                                                                                        Test Date:  8/20/85



                                                                                          Concentration (ppo)

                                                                                                   •
* . 3 . | . ' .
6 21 27 19 6
2 6 26 32 55 44 21 S(
22 27 38 38 19
40
6 21 36 26
6 24 40 32
1 . . A . . , .i * v s± ; ; j > ; ^
I
.
3.
.
4
4
.1 . ,
                                                           <    Right
                                                           Xx Tunnel
                                                ^
              Tunnel Floor

-------
VZ&&&2f2#lf&
Lett
Tunnel<4
Wall
                   1 meter
                           \ZZ%Z2ZZ%
                                                                    Chemical:  Perchlorethy-
                                                                             lenc

                                                                    Bung PositIont  Center

                                                                    Sampling Planet 3 meters

                                                                    Test Datei  8/70/85


                                                                        Wind speed (m/a)
                                                                    Right
                                                                    Tunnel
                             Tunnel Floor
                                1 "v *

-------
i
^j
Left
Tunnel
Wall
XXCXX^''X^XCXx^XXx'x''S-'S»-
1 meter

6 10 12
7 9 11
6 9 109
19 IB 188
6 IB 227
7 21 173
6 72 211
8 128 253
1 i 1 i i

1
11
9
14
1367
1417
1322
1791
601
429
453
1
X S • ^X*- •r'JX/'X^''^XXX'^^.x'
s
Ti
• • • •
11 9
39 11
726 8 8
419 58 14
344 39 7
398 68 6
134
213 14 8
54
103 9 4
1 1 1 1
                                                                                                                 Chemicali  Hcthanol

                                                                                                                 Bung Positiont  Side

                                                                                                                 Sampling Planet  1 meter

                                                                                                                 Toot Date:  8/29/85



                                                                                                                   Concentration (ppo)
                                                                                                                  Xx Right
                                                                                                                  Xx Tunnel
                                                                    Tunnel Floor

-------
CO
                       Left
                       Tunn8\^4
                       Wall

, 1 meter
5
7 4
2 S 11^
29 13
6 2 326
6 3 105 165
2 3 332 837
4 44 743 1564
15 317 1260 1944
1 . . i . . i . . i i.
	 1
3 6
23 68
2273 178g
4951 402
4262 163
2589 655
I860 661
1303
2180 559
1 i
So
Tc
12
4 ^ t
3 6
3 3
182 31
70 13
                                                                                                              Chemicali  Methanol

                                                                                                              Bung Positioni  Side

                                                                                                              Sampling Planei  1 meter

                                                                                                              Teat Datei  8/30/85



                                                                                                                 Concentration (ppn)
                                                                   Tunnel Floor

-------
      ^^^^^e^^^.C.ftnj^.;^^-^^

      ,—"/
Left
Tunnel
Wall
Chemical:   Melhanol


Bung Position]   Side


Sanpllng Planei   1 meter


Test Date:   8/30/85



    Wind speed (m/a)
                                           Tunnel Floor

-------
ro
o
                        Left   ^
                        Tunnel^

                        Wall
1 meter ,
7
15 7 9
18 70 27
14 149 391 127 18
9 497 302 44
11 16 46 225 424 286 177 57
10 39 42 260 423 260 164 84
13 38 108 217 299 244^ 196 56
32 73 210 250 251 140
1 . A 	 A 	 .. i . * ,. , * , . * * .
Sai
Tc
•
.
.
14
•
46
27
                                                                                                             Chemicalt  Mcthanol


                                                          x^x^Tunnel Cejlngx^".%%S&si2222,   B™B Position,  side


                                                                                                             Saiapllng Plane i  3 metera


                                                                                                             Test Datei   8/29/85



                                                                                                                Concentration (ppm)
                                                                   Tunnel Floor

-------
ro
Left
Tunnel 
-------
                                                                                                                    Ctiemlca 11   He thnno 1




                                                                                                                    Bung Positions   Side
ro
^
^
^J
xl
1
XJ/-
Left ;x:
Tunnel^; 4
^
0
^
P
•^x
1 meter ,
2

' ,2
5 55
20 464
4 250 583
835
112 226 -328 701
516
398 573 1006
774 578 1092 1390
\ 11111
^0'<.'"' 4'^x^ 5x^
Sanpllng Planet 3 nctc
Test Dote. 8/30/85
Concentration (ppn)
5 1
35 45 ^ ^X"
66 3 ^ Right
• • • • -^> Tunnel
145 ^<
"• . "4 . ^
275 . 275 85. ^
631 795 883 . ^
1359 1171 582 ^
>^ 6 -^' 8  9 x-^^^x-
                                                                       Tunnel Floor

-------
                                            Chemicali  Hethanol



                                            Sampling Planet  3 meters

                                            Teat  Date:  8/30/8S


                                                 Wind apccd  (in/a)
                                                  Right
                                              x^x- Tunnel
Tunnel Floor

-------
ro
Left
Tunnel-^4
Wall
                                                                         Celn
1 meter ,

. 8 3 4
360
3 8 226 £ %
2893
12 391 3440 9 3
• ••••••
4532
13 35 1345 4974 632 20
• • • • - • • •
3498
7 35 1416 3862 1100 63
2331 161
16 697 1581 2391 1822 496
2071
9 59 200 1915 2913 2249 195
1 i i 	 1 	 *,,*,,!,*,
Sa
Tc
• •
• .
*
7
I .
7
6 6
Tj
19 12
,1,. 1 , ,^
                                                                                      Chemicali  Hcthanol

                                                                                      Bung Position,  Center

                                                                                      Sampling Planet  1 meter

                                                                                      Test Dotei  8/30/85


                                                                                         Concentration (ppm)
                                                                                                                    Right
                                                                                                                -X- Tunnel
                                                   •V^-" x
                                 3^'
                                 ' .S s' .•
                                                                    Tunnel Floor

-------
I
INS
Ul
                         Left
                         Tunnel
                                                                                                                Chenleali  Hcthanol

                                                                                                                Bung Position:   Center

                                                                                                                Sampling Planei   1 meter

                                                                                                                Test Date:   8/30/85



                                                                                                                    Wind speed  (m/s)
    Right
xx Tunnel
X^Wall
                                                                     Tunnel Floor

-------
Left
Tunnel
Wall
                                              rel CeHIng>>''
, 1 meter

8 10
10 31
60 819
8 9 29 661
1034
37 - 1063
5 21 SOS
5 108 162
8 108 242
1— A J 	 1 	 i-
1

5
27
1936
1552^
1830
1960
2138
1087
495
307
1

.
14
IS
263
426
440
531
384
255
113
1
Background 5, 7, 4 s
T<
i
11
7
" •
54 14 11
69 6
25 5
10 4
1 1 1
                                        ''  '
                                           Tunnel Floor
Chemical:   Hethanol

Bung Position,   Center

Sampling Plane,   1 meter

Tent Date:   8/28/85



  Concentration (ppm)
     Right
 x>- Tunnel

-------
m
^
!
/J
Left ;X
Tunnel ^4
Wall -^
^
^
I
XQ
Background 5, 7, 4 S
, 1 meter ,
1 1 T
• ••••••••
.89 .86 .91
• ••••••••

• ••••••••

.94 .51 .46 .51 .94
• ••••••••
1.02
.08
1.02 .97 .08 .61
.61 .46 .41
. 1 1 1 1 A 1 1 i _ 1 	
                                                                                                                   Chemical:   Hethanol




                                                                                                                   Bung Positioni   Center




                                                                                                                   Sampling Planei   1 meter




                                                                                                                   Test Datei  8/28/85






                                                                                                                       Wind speed  (n/s)
                                                                      Tunnel Floor

-------
^Tunnel Celling x^
Chemicali   Hcthanol




Bung Positioni  Center
^
f
0
%
Left x^
71 Tunnel^
co W«» ;xj
x J
x?
X"
^
Sampling Planet 3 met
1 meter j
	
.' -7 2-
. 8-° ? ^
3 49 129 SO J .
7 30 302 290 IK 126 26
• • • ••• • • • *
7 72 268 400 S9Q 309^ 432 224^ 129^
216 524 746 784 559^ ^ 657 682^
626 833 889 1210 1239 1101 1371 1119 909
, i i j 	 1 	 Aii*^A 	 ••
at Date t 8/30/85
Concentration (ppo)

0
^/ Right
.Xx- Tunnel
^
0
^
^
   Tunnel Floor

-------
     i       i       A      A - .  1
                                           Chenleal:   Hethanol


                                       .   Bung Position:  Center


                                           Sampling Plane:  3 meters


                                           Test Datei  8/30/85




                                               Wind speed (a/a)
                                                Wow
                                            Xx- Tunnel
Tunnel Floor

-------
0
^
Xx
>5
,Lefl ^
'T1 1 Tunnel -^, 4
g [Wall ^
x^^x
>3'
0
^
X* rt
j- U
. 1 meter

11 12
10 8 51
68 139
10 28 36 127
7 20 87 167
8 44 125 204
25 36 101 224
13 69 133 252
1 i 1 1 1
i
7
23
52
235
168
338
337
326
3'2«
307
1

6
897
25 23 7
109
81 37 21
199 96 44
268 166 79
296 151 101^
319 226 199
111
r
s
Ti
•
.
.
10
28
34
'•
54
1
                                             Chemicali  Hothanol




                                             Bung Posltloni  Center




                                             Sampling Planet  3 meters




                                             Test Datet  8/28/85






                                               Concentration (ppm)
Tunnel Floor

-------
 I
CO
Left
Tunnel
Wall
                                           .97
                                           .89
                                           .56
                                                          .89
                                                          .86
                                                          .91
                                                          .43
                                             	1 meter	.
                                                                         .99
                                                                         .76
                                                                         .76
                                                                                        .89
                                                                                       .76
                                                                                       .76
                                               .51            .56
                                          1111
                                                                      Tunnel Floor
                                                                                                       -91
                                                                                                       .89
                                                                                                       .56
                                                                                          Chemicali  Methanol

                                                                                          Bung  PasIt ioni  Center

                                                                                          Sampling Planet  3 mete

                                                                                          Teat  Date:  8/28/85


                                                                                             Hind speed (m/a)
^ Right
/Zs Tunnel

-------
CHEMICAL NAME

TEST NO. 	2-

FILL MODE

MAX.
CONCENTRATION

ATTENUATION

TOTAL MASS LOSS
                                    E-32

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                     E-33

-------
CHEMICAL NAME

TEST NO.

FILL MODE

MAX.
CONCENTRATION

ATTENUATION

TOTAL MASS LOSS
                                          E-34

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
                32-100
TOTAL MASS LOSS   /?,/
                                      E-35

-------
CHEMICAL NAME %rcJ,/o

TEST NO.

FILL MODE
                    onJ

MAX.
CONCENTRATION

ATTENUATION
                         11
TOTAL MASS LOSS
                                    E-36

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION
                3/300
ATTENUATION
TOTAL MASS LOSS
                                    E-37

-------
CHEMICAL NAME

TEST NO. 	//_

FILL MODE

MAX.
CONCENTRATION

ATTENUATION

TOTAL MASS LOSS
                                     E-38

-------
CHEMICAL NAME


TEST NO.     /Z


FILL MODE   700
            ~
MAX
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                       c/
                                    E-39

-------
CHEMICAL NAME fflrdi/oree

TEST NO.

FILL MODE

MAX.
CONCENTRATION

ATTENUATION

TOTAL MASS LOSS
                                  E-40

-------
CHEMICAL NAME

TEST NO.      23
FILL MODE
j%>#
               o»t
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
                               E-41

-------
CHEMICAL NAME

TEST NO.
FILL MODE tforfart @ //

MAX.
CONCENTRATION  36>#Q0
^//
ATTENUATION
TOTAL MASS LOSS
                            E-42

-------
CHEMICAL NAME



TEST NO.
FILL MODE
/
CONCENTRATION
ATTENUATION
                   22.
     X/MO
TOTAL MASS LOSS
                                     E-43

-------
CHEMICAL NAME

TEST NO.

FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                         E-44

-------
CHEMICAL NAME

TEST NO.     30

FILL MODE

MAX.
CONCENTRATION   37000

ATTENUATION

TOTAL MASS LOSS
                                       E-45

-------
CHEMICAL NAME

TEST NO.
FILL MODE  8*#
                e»,
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                               E-46

-------
CHEMICAL NAME %rMrt>£

TEST NO.     32
FILL MODE  ffefo* &. /&, 7 a

MAX.
CONCENTRATION  372O0
ATTENUATION
                l/QQQ
TOTAL MASS LOSS
                                E-47

-------
CHEMICAL NAME

TEST NO. 	 /

FILL MODE
MAX.
CONCENTRATION

ATTENUATION
                       TI
TOTAL MASS LOSS
                                  E-48

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
                    27.0
TOTAL MASS LOSS    /. 0
                                      E-49

-------
CHEMICAL NAME

TEST NO.
             ¥•_
FILL MODE  ~7d0 @ 2&,/
MAX
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                   E-50

-------
CHEMICAL NAME

TEST NO.
FILL HODE
MAX.
CONCENTRATION
ATTENUATION

                    37.2
TOTAL MASS LOSS
                                   E-51

-------
CHEMICAL NAME


TEST NO.
FILL MODE  72?
           ~
MAX.
CONCENTRATION


ATTENUATION
TOTAL MASS LOSS   /.2
                                    E-52

-------
CHEMICAL NAME

TEST NO.      "7
FILL MODE   0
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS   /
                                   E-53

-------
CHEMICAL NAME
             ^
TEST NO.     7
FILL MODE  700 @ 2£.3
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                E-54

-------
CHEMICAL NAME

TEST NO.
FILL MODE  700$
MAX.
CONCENTRATION

ATTENUATION

TOTAL MASS LOSS   /. ##
                                 E-55

-------
 CHEMICAL NAME

 TEST NO.     //

     MODE
MAX.
CONCENTRATION
AHENUATION
TOTAL MASS LOSS
                                 E-56

-------
CHEMICAL NAME

TEST NO.
FILL MODE
ATTENUATION
MAX.
CONCENTRATION    //7# W»
TOTAL MASS LOSS
                                  E-57

-------
CHEMICAL NAME

TEST NC.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                    E-58

-------
CHEMICAL NAME

TEST NO.

FILL ,'WDE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                 E-59

-------
CHEMICAL NAME £f6tf/e#e

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                   E-60

-------
CHEMICAL NAME

TEST NO.

FILL MODE
MAX.
CONCENTRATION

AHENUATION
TOTAL MASS LOSS
                                 E-61

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                E-62

-------
CHEMICAL NAME

TEST NO.

FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS    0,22
                                    E-63

-------
CHEMICAL NAME £f6u/e»e
                C7     <


TEST NO.     /?
FILL MODE
MAX.

CONCENTRATION


ATTENUATION
TOTAL MASS LOSS
                                     E-64

-------
CHEMICAL NAME


TEST NO.     20
FILL MODE  7^7
           ~~
MAX.
CONCENTRATION


ATTENUATION
                      / r
TOTAL MASS LOSS   // e,
                                    E-65

-------
CHEMICAL NAME £Mtt/e»f_
                 ~J

TEST NO.      2/
FILL MODE
                   /0.2
MAX.

CONCENTRATION



ATTENUATION
                       / /
TOTAL MASS LOSS
                                   E-66

-------
CHEMICAL NAME Ejj
                  v7

TEST NO.       22.
FILL MODE
                ft //,2

MAX.

CONCENTRATION



ATTENUATION
                       fT
TOTAL MASS LOSS
                                   E-67

-------
CHEMICAL NAME

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL MASS LOSS
                                  E-68

-------
CHEMICAL NAME £#r«/e»
-------
CHEMICAL NAME £*Atf/exe

TEST NO.
FILL MODE
MAX.
CONCENTRATION

ATTENUATION
TOTAL flASS LOSS
                                  E-70

-------