United States
Environmental Protection
Agency
Atmospheric Sciences
Research Laboratory
Research Triangle Park NC 2771 1
/
Research and Development
EPA/600/S3-86/003 Mar. 1986
v>EPA Project Summary
Impact of Primary Sulfate and
Nitrate Emissions from Selected
Major Sources: Phase II: Sulfuric
Acid Plant and Pulp and Paper
Mill
J.D. McCain, W.G. Kistler, and D.H. Cannes
The impact of local point sources
emissions on the near-source atmos-
phere has been investigated in a two-
phase study. The phase I study was
conducted to investigate the impact of
a coal-fired power plant on the near
source atmosphere. Phase II was con-
ducted to study the impact of a sulf uric
acid plant and a pulp and paper mill. The
results of phase I are presented in a
separate report.
Sampling for phase II took place
during August of 1981 at the sulfuric
acid plant operated by American Cy-
anamid and the pulp and paper mill op-
erated by the Union Camp Corporation.
Both plants are located in Savannah,
Georgia. Source measurements were
made for emission rates of H2SO4, par-
ticulate sulf ate, SO2, NOX, and total
particulate matter; and elemental com-
position by particle size was deter-
mined. Ambient measurements were
made for all the preceding pollutants
plus HNO3, particulate nitrate, and
pertinent meteorological parameters
required for dispersion modeling.
The average concentration of TSP in
the ambient was about 120 fjg/m3,
while the average sulfate and nitrate
concentrations were, respectively, 6
fjg/m3 and 2 jug/m3. Factor analysis
based on elemental analyses of ambient
air particulate samples, the plants'
particulate emissions, and local soil
samples indicated that wind-blown soil
was the dominate source of the ambi-
ent particulate matter. Factor analysis.
dispersion modeling, and comparative
upwind-downwind concentrations in-
dicated that primary sulfate and nitrate
emissions from the target sources were
responsible for only a small fraction of
the observed ambient concentrations
of these materials. The major source of
the observed sulfates and nitrates ap-
peared to be secondary aerosols, per-
haps dominated by long-range trans-
port.
This Project Summary was devel-
oped by EPA's Atmospheric Sciences
Research Laboratory. Research Tri-
angle Park, NC, to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
Most current studies on the mecha-
nisms of sulfate and nitrate formation
and transport in the atmosphere have
addressed the problems of long-range
transport. However, a significant portion
of the problems associated with these
compounds may arise in the near-source
vicinity. The program described here was
intended to provide information on the
partitioning of sulfates and nitrates in
ambient air between primary (local) and
secondary (long-range-transport) sources.
This report presents the results of the
second phase of a two-phase source-
ambient study. Phase I results of the
program are reported elsewhere. Phase II
concerned the impacts on the ambient air
-------
of a sulfuric acid plant and a pulp and
paper mill. Sampling for Phase II took
place during August 1981 in the environs
of Savannah, Georgia.
Sulfur and nitrogen compounds are of
particular concern because they are
preponderant in flue gas and in decom-
position products of other process
streams. High levels of SOz, NO, NOz, and
various sulfates and nitrates are found in
ambient air associated with industrial
activity and areas with high automotive
traffic densities. Serious deterioration of
the environment has been linked to these
compounds. In 1964, regulation of SOz
emissions began in this country. The use
of high-sulfur fuels was inhibited, and
taller stacks were employed to improve
local air quality. This reduced the mea-
sured levels of SOz; however, ambient
sulfate levels have not been substantially
improved. Thus, it is important that the
specific contributions of these aerosols to
the local- and long range concentrations
transport be determined.
Three approaches were taken to
quantify the impact of the plants on the
study area. First, dispersion modeling
was used to predict average downwind
concentrations. Second, principal com-
ponent factor analysis (PFA) was applied
to an ambient data base comprising the
concentrations of SOz, NO,, TSP, SO^2,
and NOs~; wind direction relative to the
source and sampler locations; and the
concentrations of a number of elements
found in the particulate matter collected
by hi-vol samplers. Finally, Target Trans-
formation Factor Analysis (TTFA) was
used in analyzing the data to isolate and
quantify the impacts of specific sources
at each sampling location.
Test Site
The study was carried out in and
around the city of Savannah, Georgia.
The targeted sources were a pulp and
paper mill operated by the Union Camp
Corporation and a sulfuric acid plant op-
erated by American Cyanamid, Inc.,
located on an east-west line along the
Savannah River. American Cyanamid's
facility is located at the eastern extreme
of the industrial development, and Union
Camp is located inland; the two facilities
are equidistant from the Savannah cen-
tral business district. The industrialized
area extends another 8 km (5 mi) west-
ward from Union Camp. The industrial
and urban developments are confined
wholly to the south side of the Savannah
River.
Two additional sulfuric acid plants
were in operation in this industrialized
area; one of these has a rated capacity of
about 10 percent of that of Cyanamid and
the other a capacity of about half that of
Cyanamid. In addition to Union Camp,
several pulp mills operate in the area,
only one of which uses the Kraft process;
this plant, with about one quarter the
capacity of Union Camp is located at the
western end of the industrial develop-
ment.
The Savannah plant, of the Union
Camp Corporation is the world's largest
pulp and paper plant with a pulp capacity
of 2850 tons per day, of which 2550 tons
per day are produced by the Kraft pro-
cess. The major emission sources within
the complex are three each recovery
boilers, lime kilns, and power boilers.
Measurement Methods
The primary variables of interest in
this study were particulate and gaseous
sulfates and nitrates in the ambient air
that were directly or indirectly the results
of emissions from the acid and pulp and
paper plants. Further, it was desirable to
differentiate between sulfate and nitrate
salts and sulfuric and nitric acid. Thus
data were needed on the ambient and
source concentrations of these variables
and their precursors.
Source measurements were made of
the emission rates of sulfuric acid, sul-
fate salts, sulfur dioxide, nitrogen oxide,
nitrogen dioxide, total particulate; and
elemental composition by particle size
was determined. Measurements in the
ambient included all of the preceding
plus nitrate salts, nitric acid, and all
pertinent meteorological parameters
required for dispersion model predictions
of downwind concentrations.
A grid of acceptable monitoring loca-
tions was developed in the vicinities of
the plants, and based on anticipated wind
flow, sites were selected for monitoring
at the beginning of each test day. Be-
cause of the relatively large separation
between the acid plant and the pulp and
paper plant, sampling was concentrated
in the vicinity of each in turn for about
one half of the tests. Data were taken at
three mobile stations, one of which was
deployed at an upwind location while the
remaining two were located downwind of
the target source(s). Wind flow was
monitored throughout each test day, and
the mobile stations were moved when
winds shifted in order to maximize the
amount of data obtained in the plumes.
The monitoring sites were selected as
much as possible in open areas, where
the prevailing wind flow was not signifi-
cantly disturbed by local topography or
vegetation. Sites near other local sources
of pollutants were avoided.
Data to identify dispersion character-
istics were taken from a portable 10-m
meteorological tower provided by TRC.
Upper air measurements were made
using pilot balloons and temperature
sonde equipment.
Measurement Results
Source
An overview of the source emission
parameters is given in Table 1. This table
summarizes for each unit at the two
facilities all emission rate measurements
for the principal components of interest
particulates, SOz, HzSO* and NO*.
Samples were taken for determining
the composition of the plant's emissions
in six particle-size fractions with South-
ern Research Institute series cyclone
systems and with a special purpose im-
pactor designed and constructed by SoRI
under a previous EPA contract. Speci-
mens from each size fraction from each
of these samplers were analyzed by X-ray
fluorescence (XRF) and those from the
cyclones were analyzed by neutron acti-
vation analysis (NAA) as well. These
analyses provided the desired "emission
signature" for the plants.
Ambient
A summary of the ambient data is
given in Table 2. All values given in this
table insofar as possible represent
averages over the time periods during
which the hi-vol samplers for sulfate and
nitrate determinations were in operation.
Two methods of measuring ambient
SOz concentrations were employed, con-
tinuous monitors and peroxide bubblers.
The primary integrated samples were to
be taken coincident with each sulfuric
acid sample. Problems were encountered
with each method that in the end resulted
in significant loss of data.
High ambient NOx concentrations
showed essentially no correspondence
with occasions when the sources of
interest were located upwind of the
sampling stations. This result was not
unexpected, as none of the sources ex-
cept perhaps the power boilers at the
pulp mill were significant NOx emitters.
The results of all ambient sulfate and
sulfuric acid sampling are given in Table
2. The values reported on each line of the
table are for samples that were taken
over approximately the same time in-
tervals; however, they seldom represent
truly coincident samples. In a few cases
samples obtained with one method over-
-------
Table 1. Source Emissions Summary
Source
Am. Cyanamid
Acid Plant
West Calciner
East Calciner
Union Camp
12 flee. Boiler
13 Rec. Boiler
14 Rec. Boiler
12 Smelt Tank
13 Smelt Tank
14 Smelt Tank
4 Lime Kiln
5 Lime Kiln
6 Lime Kiln
11-12PO. Blr.
10 Power Blr.
Gas Flow
(actual rrf/s)
28.2
34.9
32.8
115
105
219
10.9
11.5
25.6
24.1
25.8
32.6
360
108
H20
1%)
0
21
24
29
34
34
36
36
36
41
51
36
16
12
Temp.
deg.
C
90.6
68.3
71.1
151.7
154.4
148.9
73.9
73.9
73.9
76.7
81.1
73.9
218.3
190.6
Table 2. Summary of Ambient Sampling Results,
Wind Direction
Within
30" 1% of Time) Wind
39 9
>39 31
NO
ppb
21.60
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
• 12.20
ppm
15
32
25
38
32
42
1.8
0.25
45
46
NA
95
148
NITRATE
fug/m'l
NO, HN03 PART.
ppb -ug/m' ug/m*
74.80
NA
NA
NA
NA
NA
6.33
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
18.30
0.45 O.56
1 73 0.69
2.68 1 95
1.44 0.60
2.SO O.47
1.52 0.38
1.41 0.44
1.47 0 10
1.45 0.26
NA NA
1.74 050
NA NA
2.92 NA
0.78 078
0.81 027
1.36 O.O6
0 58 O.OO
064 0.10
2.22 O.OO
-------
lapped two samples taken with another;
in these cases the entry for the longer-
duration sample is duplicated for each of
the shorter samples. Comparisons of
average HzSO4 concentrations with total
sulfate concentrations from the hi-vol
samples show that typically about 50% of
the ambient sulfate was sulfuric acid.
The peak sulfate and nitrate concen-
trations were measured on days when
pervasive, region-wide hazes occurred.
These were days when the winds were
north and west. Because the upwind
concentrations were as great as the
downwind, even though virtually no local
sources were upwind in those directions,
moderate- and long-range-transport
sources must be considered a real possi-
bility.
Sections of the Teflon membrane hi-
vol filters and the Lundgren impactor
substrates and filters were analyzed by
XRF and NAA techniques to provide in-
formation on the elemental composition
of the ambient participate matter. Thirty-
eight elements were detected on some or
all of the Teflon filters by the combination
of XRF and NAA techniques. Some of
these were found only on a few of the
filters and then only near the detection
limits. However, 36 of the 38 were de-
tected at high enough concentrations on
most or all of the filters to be useful in
paniculate source identification.
Source Impact
Three approaches were taken to
quantify the impact of the target sources
on the study area. First, dispersion model
predictions were made for downwind
concentrations of sulfate, SO& and NOx
at the locations of the ambient sampling
stations using averaging times compara-
ble to the sampling periods used in the
sampling. Second, principal component
factor analysis (PFA) was applied to the
correlation matrix of an ambient data
base comprising data for SOa NOx, TSP,
SO4, NOa, wind direction relative to the
source and sampler locations, and 36
elements for which reasonably good con-
centration values were available from the
Teflon membrane hi-vol filters. This
analysis provided much insight into the
probable number and nature of sources
that had a detectable impact on the
ambient stations, but did not quantify the
impact on individual stations. Finally,
Target Transformation Factor Analysis
(TTFA) was used in analyzing the data in
an attempt to isolate and quantify the
impacts of specific sources having known
signatures at each specific sampling
location.
A four-factor PFA solution resulted in
the following. The first factor showed
heavy factor loadings for elements com-
mon to soil, and fairly clearly represents
windblown soil. The second factor
showed high loadings on Ar, Se, Ce, and
Zn. This factor was traced to samples
from a single location and appeared to
result from wind-transported emissions
from a galvanizing plant which was ENE
of that location. The third factor had sig-
nificant loadings on only sulfates and
nitrates and probably represents a sec-
ondary aerosol, possibly brought into the
area by long-range transport. Most of the
explained variance for these two vari-
ables (sulfates and nitrates) arises from
this factor. The fourth factor is associated
with a TiOa/acid plant complex but does
not arise directly from the sources tested.
The area around the Ti02/acid plant
complex was liberally coated with Ti02
dust, and this factor may represent re-
suspension of some of this material,
together with fugitive emissions from the
complex. The four factors are sufficient to
explain 83% of the variance in the input
data matrix. Inclusion of a greater
number of factors resulted in additional
factors whose loadings were less than
their standard errors.
The TTFA input data matrix consisted
of various combinations of 20 to 47
elements, gases and radicals for each of
39 filters. The data on the elemental con-
centrations were augmented with values
for TSP from the other parallel hi-vol run;
sulfate, nitrate, and nitric acid concentra-
tions determined from the same hi-vol
Table 2. (Continued}
Wind Direction
Within
3O° * 1% of Time/ Wind
.
-------
runs which provided the data on the ele-
ments; covered by the hi-vol samples.
There were too many missing values in
the NOx data for their inclusion to be
meaningful.
Source vectors used in attempting to
regenerate the data were the measured
emission strengths of the various units in
the plants and combinations thereof; the
compositions of soil samples from the
area; and vectors obtained from the
literature and the phase I test for auto-
motive sources, power plants, sea-salt,
etc. The predictive capabilities of TTFA
were used to refine the vectors obtained
from the literature for this specific ap-
plication. No satisfactory regeneration
could be obtained from any combination
of these source vectors as long as sulfur
(or sulfate) and nitrates were included in
the data set. Uniqueness tests for these
variables confirmed the conclusions
reached from PFA: they do indeed exhibit
unique behavior and cannot be directly
related to any source which might be
potentially identified by its signature in
the remaining elements.
Potential source factors were also
obtained by using the predictive capabili-
ties of TTFA to independently generate a
set of possible source vectors. Six possi-
ble source vectors were generated in this
fashion. One of these very closely
matches a composite of the soil samples,
another shows high concentrations of Ti
and P, and a third is very high in sulfur
and nitrates and has a high association
with vanadium. The Ti-P vector is, as
previously mentioned, associated with
the area in which the acid plant is lo-
cated; but it is not directly related to the
sources sampled there, nor does it have
much sulfur associated with it. The sul-
fate-nit rate vector has no resemblance to
either the acid plant sources or the pulp
mill and cannot be accounted for by any
range transport for its origin.
Source emission data and meteoro-
logical data for each test day were used
with two versions of a dispersion model
to predict average concentrations for
each test day at all downwind sampling
sites. Predictions were made only for the
plant's contributions to SO4'2 and in the
case of the pulpmill, NOX. One of the
models used the Briggs plume rise al-
gorithms whereas the other used the
older Holland equations. Observations of
plume behavior at the plants in question
lead the authors to believe that the
Holland equations are more realistic in
the current application. The plume rise
predicted by the Briggs equations seemed
too large in all cases. The effect of the
overprediction of plume rise is an under-
estimation of near-source impact thus
the result of using the Holland plume rise
algorithm is to make conservative (high)
predictions of downwind concentrations
as compared to those from the Briggs
model. Even using the conservative
model, the plants' contributions to the
measured downwind sulfate concentra-
tions were predicted to be no more than
five to twenty-five percent of the totals in
the case of Cyanamid, and one to ten per-
cent in the case of Union Camp. Only ten
percent of Cyanamid's predicted impact
would be from the acid plant since ninety
percent of the Cyanamid sulfate emis-
sions were from the calciners. Predicted
NOx and SOz concentrations were also
much lower than the measured down-
wind impact in terms of these gases.
Conclusions
Neither the sulfuric acid plant or the
pulp and paper mill was found to be a
significant contributor to the local ambi-
ent burdens of sulfates and nitrates. The
bulk of the sulfates and nitrates appeared
to result from secondary aerosols and/or
long range transport, especially during
periods of high sulfate concentration.
J. D. McCain and W. G. Kistlerare with Southern Research Institute, Birmingham,
AL 35255-5305; and D. H. Carnes is with TRC. East Hartford. CT06108.
Kenneth T. Knapp is the EPA Project Officer (see below).
The complete report, entitled "Impact of Primary Sulfate and Nitrate Emissions
from Selected Major Sources, Phase II: Sulfuric Acid Plant and Pulp and Paper
Mill." (Order No. PB 86-145 489/AS; Cost: $16.95, subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
S. GOVERNMENT PRINTING OFFICE:1986/646-l 16/20782
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-86/003
0000329 PS
U S ENVIS PROTECTION AGENCY
REGION 5 LIBRARY
230 S 0£ARS0R« STREET
CHICAGO IL 60604
------- |