United States
 Environmental Protection
 Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
 Research and Development
EPA-600/S7-81 -003b  Apr. 1981
 Project  Summary
 Emissions  Assessment  of
 Conventional  Stationary
 Combustion  Systems:

 Volume  IV.
 Commercial/Institutional
 Combustion Sources
N. F. Surprenant, P. Hung, R. Li, K. T. McGregor, W. Piispanen, and S. M.
Sandberg
  This report characterizes air emis-
sions from commercial/institutional
external combustion sources and
reciprocating engines and is the fourth
of a series of five project reports
characterizing emissions  from con-
ventional combustion sources. The
emissions characterization of commer-
cial/institutional combustion sources
was based on a critical examination of
existing data, followed by  a modified
Level I sampling and analysis approach
to resolve data gaps. The major devia-
tion from Level I procedures was the
addition of QC/MS analysis for poly-
cyclic organic  matter (POM)- Tests
were conducted at 22 external com-
bustion and' 6 internal combustion
sites.
  The results of the environmental
assessment indicate that air emissions
from these sources represent a poten-
tial environmental hazard.  Emissions
of criteria pollutants, with the excep-
tion of carbon monoxide, from most of
the source categories tested are en-
vironmentally significant. Paniculate
sulfate and SO3 emissions from the
coal- and wood-fires sources are also
significant. In addition, emissions of
several trace elements are of concern:
aluminum, barium, beryllium, calcium,
chlorine, cobalt, chromium, copper,
fluorine, iron, potassium, lithium,
sodium, nickel, phosphorus, lead,
silicon, and vanadium from coal-fired
external combustion sources; nickel
from distillate oil sources; and nickel,
chlorine, chromium, and vanadium
from residual oil sources. Several
potentially hazardous POM compounds
were tentatively identified in the emis-
sions from solid fuel-fired sources,
particularly from the one wood-fired
stoker tested. Flue gas emissions of
POM from solid fuel-fired sources will
require further study to positively
identify the POM compounds emitted.
  This Project Summary was developed
by EPA'sIndustrial Environmental Re-
search Laboratory. Research Triangle
Park, NC,to announce key findings of"
the research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back}.

-------
Introduction
  Emissions from commercial/institu-
tional external combustion sources for
space heating and commercial/institu-
tional internal combustion reciprocating
engines are characterized in this report.
The approach involves a critical review
of existing emission data, followed by
the conduct of a sampling and analysis
program to fill gaps in the data base and
to identify additional data needs. Specif-
ically, the objectives of this program
were:
  • to compile and evaluate available
    air emissions data from commer-
    cial/institutional stationary con-
    ventional combustion processes,
  • to acquire needed new emissions
    data  from  field tests of selected
    sources using  modified  Level  I
    procedures,
  • to characterize air emissions from
    commercial/institutional station-
    ary conventional combustion
    processes,  using both the existing
    data base and field test results, and
  • to determine additional data needs,
    including identification of specific
    areas of data uncertainty.
  Level  I procedures use semiquantita-
tive (plus or minus a factor of 3) tech-
nicjues of sample collection and labora-
tory and field analysis:  (1) to provide
preliminary emissions data for waste
streams and pollutants  not adequately
characterized; (2) to identify potential
problem areas; and (3) to prioritize
waste streams and pollutants in those
streams for further, more quantitative
testing. Using  the information from
Level  I,  available  resources can be di-
rected toward Level II testing which
i nvolves specific quantitative analysis of
components of  those streams that do
contain significant pollutant levels. The
data developed  at Level  II are used to
identify control technology needs andto
further define the environmental haz-
ards associated with emissions.

Summary
  The commercial/institutional external
combustion sources  evaluated in  this
report are sources used for space heating
of trade establishments, health and
educational institutions, and govern-
ment facilities. These application areas
are identical to those used by the De-
partment of Energy (DOE) in  compiling
energy consumption data for the com-
mercial sector. Commercial combustion
units have also been defined as units
with heat  inputs ranging from 0.42 to
13.2 x 109 Joules (J)* per hour. How-
ever, this definition excludes many
smaller and larger  units used in the
commercial/institutional sector. Insti-
tutional  units  especially tend to be
appreciably larger than 13.2* 109 J/hr
and account for almost 20 percent of the
commercial/institutional sector fuel
consumption.
  Commercial/institutional fuel con-
sumption for space  heating was 5.1 x
1018 J in 1978  based on DOE data for
total fuel consumption and estimates of
the fraction of this fuel used for space
heating. This consumption value is less
than 10 percent of the estimated 1978
national fuel consumption figure of 54 x
1018 J, excluding fuel used in the trans-
portation sector. Commercial/institu-
tional external combustion sources for
space heating primarily use  oil (52 per-
cent) and gas (44 percent). Small amounts
of coal and wood  are also used by the
commercial/institutional sector. Inter-
nal combustion sources in the commer-
cial/institutional sector, primarily gas-
and oil-fired reciprocating engines, are
used for pumping municipal water and
sewage. Small amounts of fuel may also
be used by internal combustion sources
for auxiliary power generation.
  Heating systems for commercial/in-
stitutional sources are concentrated in
areas of high population density such as
the Northeast,  Midwest, and parts of
California. Oil  consumption is most
heavily concentrated in the Northeast
with the States of New York, Massachu-
setts, New Jersey,  and Pennsylvania
consuming about 25 percent of the U.S.
total. Commercial gas consumption for
space heating is more widely distributed
than oil,  but is  still  most heavily con-
centrated in the Midwest and Northeast.
  Commercial/institutional external
combustion sources can be sold as
either  packaged units or boilers to be
constructed onsite. Most units in the
commercial sector are packaged units.
Field-erected  units,  for the  most part,
are restricted to larger institutional
facilities. Estimates of the total number
of commercial  external combustion
sources have been  reported and, ac-
cording to these estimates, there are
approximately  1.5 x 106 commercial
sources.  Most of- the smaller units (<
13.2 x 109 J/hr) are cast iron or firetube
units,  and only  5  percent of these
 * 7055 Joules U) =1 Btu. Although it is EPA policy to
  use the metric system, this report uses certain
  nonmetric units for convenience.
smaller units are of watertube design. .
Watertube units, however, constitute'
100 percent of all units above 50 x 109
J/hr input.
  Air pollution control  equipment  is
generally  not installed on the smaller
commercial external combustion sources,
although new burner designs, atomiza-
tion methods, and furnace constructions
are being studied to reduce emissions of
NOx and particulates. Burner modula-
tion during periods of fluctuating de-
mand, instead of on/off cycling, also
reduces particulate and hydrocarbon
emissions from oil-fired sources.
  Gaseous and particulate emissions
from the flue gas stacks are emphasized
in this study of commercial/institutional
combustion sources. Although some of
the larger  institutional external com-
bustion systems are local sources  of
water pollution and fugitive particulate
emissions  from coal pile storage and
ash disposal, their contribution to the
national water pollution and fugitive
emission burden  is negligible.  It  is
estimated, based on the amount of coal
consumed by the commercial/institu-
tional sector, that they contribute less
than 1  percent of such emissions from
all stationary combustion sources.
  Evaluation of existing emissions data ,
has indicated that the data base for gas-*4
and oil-fired external combustion sources,
although limited, is adequate for nitro-
gen oxides (NOx), total hydrocarbon
(HC), carbon monoxide (CO), particulate,
and sulfur dioxide (502). However, the
existing data base for specific organic
emissions for these sources is inade-
quate and, for the oil-fired sources, the
existing data base for sulfur trioxide
(SOs) and trace elements is inadequate.
Emissions data from solid fuel-fired
sources are generally inadequate for all
pollutants.
  In the case of oil-fired internal com-
bustion sources, data are inadequate for
SOa, trace element, and specific organic
emissions. Data for gas-fired recipro-
cating  engines are adequate; however,
one unit was tested in this program  to
confirm data adequacy.
  Modified Level I sampling and analysis
procedures were used in this emissions
assessment of conventional combustion
systems program. These procedures
were developed as an integral part  of
the program and were published  as a
separate document. Although the pro-
cedures differ to some degree from
official  EPA Level I  procedures,  they
have been used throughout this program A

-------
 to ensure continuity of sampling and
 analysis.
   Because of the  deficiencies in the
 existing emissions data base,  the fol-
 lowing 22 external combustion systems
 were tested: five gas-fired, three distil-
 late oil-fired, five residual oil-fired,
 three anthracite stokers, three bitumi-
 nous stokers, two bituminous pulverized
 dry units, and one wood-fired stoker.
 Four oil-fired, one gas-fired, and one
 dual-fired internal combustion  recipro-
 cating engines werealso tested.  Specific
 sites were chosen  based on the repre-
 sentativeness of the sites as determined
 by the important system characteristics
 within  each source category, including
 system design, size, and age. Many of
 the sites tested fall  within the commer-
 cial size classification range, although
 some, particularly the pulverized bitumi-
 nous-fired units, greatly exceed the
 upper commercial  size limit of 13.2 *
 10  J/hr input capacity.
   The  Source Assessment Sampling
 System (SASS) train, developed under
 contract to EPA, was used to collect both
 gaseous and paniculate emissions in
 quantities sufficient for the wide range
 of analyses needed to adequately char-
 acterize emissions from commercial/
 institutional combustion sources. The
 SASS train consists of a conventional
 heated probe, three cyclones and a filter
 mounted in a Heated oven,  a gas condi-
 tioning  system, an XAD-2  polymer
 adsorbent trap, and a series of impingers.
 Particulate matter  is size fractionated
 and collected in the cyclones and on the
 filter; gaseous organics and some inor-
 ganics are  collected  by  the  XAD-2
 adsorbent; and the remaining gaseous
 inorganics and trace elements are
 captured by the impingers. The train is
 run until at least 30 m3 of gas have been
 collected. This criterion was established
 in conjunction with analytical technique
 sensitivities to ensure that any emis-
 sions that would increase the ambient
 loading by more than 1/ug/m3 would be
 detected. The cyclones were deleted for
 the tests at the gas- and oil-fired sites
 because paniculate loadings were too
 low to provide a weighable quantity of
 sample from each cyclone.
   In addition to the SASS train, other
 equipment was used to collect gaseous
 components not captured by the train. A
 gas chromatograph (GC) with a flame
 ionization detector was used in the field
 to analyze Ci-C8 hydrocarbons collected
.in gas sampling bags. Additionally,
fthese samples were analyzed for  CO,
 carbon dioxide (COz), oxygen (02), and
 nitrogen (Nz) by the GC using a thermal
 conductivity detector. Field sampling for
 NOx and SOa was also conducted at
 selected sites using a Method 7 train for
 NO, (40 CFR 60, Appendix A, Method 7)
 and a modified Goksoyr-Ross train for
 80s collection.
   A modified  Level I sampling and
 analytical procedure was used in this
 emissions assessment program. Major
 deviations from Level I Procedures were
 the addition of gas chromatography/
 mass spectroscopy (GC/MS) to the or-
 ganic analyses, the combination of
 certain SASS train fractions before
 analysis, and the deletion of inorganic
 analysis of SASS train samples col-
 lected from gas- and oil-fired sources.
 The combination and deletion guidelines
 were instituted as a result of low levels
 of pollutants found in the emissions of
 previously tested  gas- and oil-fired
 utility boilers and residential heating
 systems.
   The Level  I inorganic analysis was
 designed to identify all elemental species
 collected in  the SASS train fractions
 and to provide semiquantitative data on
 the elemental  distributions and total
 emission factors. The primary tool for
 Level I  inorganic analysis is the Spark
 Source Mass Spectrograph (SSMS).
 SSMS data were supplemented with
 Atomic Absorption Spectrometry (AAS)
 data for mercury (Hg), arsenic (As), and
 antimony (Sb), and with standard method
 determinations for sulfates.
  The following SASS train fractions
 from the solid fuel-fired  sources were
 analyzed for  their  elemental composi-
 tion: (1) the  cyclone catches, (2) the
 particulate filter, (3) the XAD-2 sorbent,
 and (4)  a composite sample containing
 portions of the XAD-2 module conden-
 sate and nitric acid rinse, and the first
 impinger solution.  Fuel was also ana-
 lyzed for the solid  fuel- and oil-fired
 sources.
  Level I organic analysis provides data
 on  volatile (C7-Ci6)  and nonvolatile (>
 Cie) organic compounds to supplement
 data for gaseous organics (Ci-Ce) mea-
 sured  in the field. Organics  in the
 paniculate fractions, the XAD-2 sorbent,
 and XAD-2 module condensate trap
 were recovered by methylene chloride
 extraction. SASS  train  components
 including the tubing were carefully
 rinsed with methylene chloride or meth-
ylene chloride/acetone solvent to re-
cover all organics collected in the SASS
train.  SASS train rinses and extracts
 recovered from the gas- and oiC-fired
 sites were combined for analysis; how-
 ever, samples collected from solid fuel-
 fired sources were analyzed separately:
   Because all samples contain signifi-
 cant quantities of solvents from rinsing
 and are too dilute to detect organic
 compounds  by most instrumental tech-
 niques used by Level I procedures, the
 first step in the analysis  was to con-
 centrate the sample fractions from as
 much as 100 ml to 10 ml in a Kuderna-
 Danish apparatus in which rinse solvent
 was evaporated while the organics of
 interest were retained. Kuderna-Danish
 concentrates were then evaluated by
 GC, gravimetric analysis, infrared spec-
 trometry (IR), and sequential GC/MS.*
 The extent of the organic analysis was
 determined by the stack gas concentra-
 tions found  for total organics (volatile
 and nonvolatile). If the  total organics
 indicated a stack gas concentration
 below 500 fjg/m3, further analysis was
 not conducted. If the concentration was
 above 500 /ug/m3, a class fractionation
 by liquid chromatography was conducted
 followed by GC, gravimetric, and IR
 analyses. Fractions that contained more
 than the equivalent of  500 yug/m3 or
 were of special interest were also analyzed
 by low resolution mass spectroscopy
 (LRMS).
   The results of the field measurement
 program for flue gas emissions from
 commercial/institutional sources, along
 with supplementary values obtained
 from the existing data base for certain
 pollutants, are presented in Table 1.
 Also listed  in this  table are ambient
 severity factors, defined  as the ratio of
 the calculated maximum ground level
 concentration of the pollutant species to
 the level at which a potential environ-
 mental hazard exists. An ambient sever-
 ity factor of greaterthan 0.05 indicates a
 potential problem  requiring further
 attention.
  The emission factors shown in Table 1
 are  uncontrolled emissions factors.
 However, in the case of  the solid fuel-
 fired combustion categories, some
 degree of paniculate control does exist
 in the commercial/institutional sector.
 Overall paniculate control  efficiency is
 estimated to be 40 percent for bitumi-
 nous, pulverized dry bottom boilers and
 20 percent for all stokers. Gas- and oil-
fired units are essentially uncontrolled.
 •The major modification in the Level I sampling and
 analysis procedure was the GC/MS analysis for
 polycyclic organic matter (POM).

-------
Table 1.     Summary of Emissions Characterization of Commercial/Institutional Combustion Sources
                                                         Combustion source category
            Gas-fired
            boilers
Distillate
oil-fired
toilers
Residual
oil-tired
boilers
Bituminous,
pulverized
dry bottom
Bituminous
 stokers'"
Anthracite
 stokers
                                                                                       Wood stokers
 Gas-fired
reciprocating
 engines
  Oil-tired
reciprocating
  engines
 Pollutant'
          Emis-  Ambient Emis-  Ambient  Emis-  Ambient  Emis-  Ambient  Emis-  Ambient Emis-  Ambient  Emis-  Ambient  Emis-  Ambient  Emis-  Ambient
          sion   sever-   sion   sever-    sion   sever-   sion   sever-   sion   sever-   sion   sever-   sion   sever-   sion    sever-    sion    sever-
          factor    ity   factor    ity    factor    ity    factor    ity    factor    ity   factor    ity   factor    ity    factor    ity    factor    ity
          (ng/JI   factor'  Ing/JI   factor   Ing/JI  Factor   Ing/JI   factor   Ing/JI   factor  (ng/J)   factor  (ng/JI   factor  (ng/JI   factor   (ng/JI   factor
Partkulates
HO,
SO,
CO
HC
Paniculate
sutlete'
SO,
Trace
elements
Al
Be
Be
Ca
Co
Cr
Co
f
fe
K
Li
Na
Ni
P
Si
V
Total POM
2
50
0.26
a
3

—
—


—
—
—
_
_
_
—
—
_
—
_
—
—
—
—
_
0.010
0.0007 6 0.0022 37
0.08 68 0.11 172

-------
levels were similar to those found in the
existing data base for oil-fired engines,
and ambient severity factors did not
exceed 0.05 for .any of the compounds
detected. The  high POM emissions
measured for the dual-fired engine
were somewhat surprising because the
quantity of oil used represented only 5
percent of the total thermal  input. No
POM emissions were detected from the
engine fired solely by gas.

Conclusions

  Several conclusions, listed below,
can  be  drawn  from this   emissions
assessment of commercial/institutional
combustion sources:
  • Emissions of paniculate, NO* SOz,
    CO, and HC from commercial/insti-
    tutional sources represent approxi-
    mately  1.7 percent, 4.9 percent,
    3.0 percent, 0.5 percent, and 0."
    percent, respectively, of total emis-
    sions from stationary combustion
    sources.  Despite  this  relatively
    minor contribution to national emis-
    sions, criteria pollutant emissions
    from  individual combustion sources
    can have a significant local impact
    as noted below.
  • Flue gas emissions of N0> from all
    of the commercial/institutional
    source categories studied in this
    program, with the exception of a
    wood-fired stoker tested at low
    load  conditions, are of concern.
    Ambient severity factors exceed
    0.05  and, thus, individual sources
    can have a significant local impact.
  • Flue gas emissions of S02from the
    coal-fired and residual oil-fired
    combustion sources are associated
    with ambient severity factors
    greater than 0.05 and, thus, are of
    environmental significance. Ambi-
    ent severity factors associated with
    paniculate sulfate and 80s emis-
    sions from the solid  fuel-fired
    sources tested also exceed 0.05
    and are of concern.
  • Flue gas emissions of particulates
    from  uncontrolled solid fuel-fired
    sources are associated  with high
    ambient severity factors. Moderate-
    to-high efficiency control devices
    are required in many cases to re-
    duce severity factors to 0.05. Be-
    cause the application of paniculate
    control devices to solid fuel-fired
    commercial/institutional combus-
    tion sources is not extensive, these
    sources are of practical concern.
• F|ue gas emissions of total hydro-
   carbon are significant for commer-
   cial/institutional  bituminous
   stokers, wood stokers, and recip-
   rocating engines. Ambient severity
   factors exceed 0.05 for these com-
   bustion source categories.
• Flue gas emissions of CO do not
   appear to be a problem. Ambient
   severity factors for all source cate-
   gories are 0.001 or less.
• Particle size distribution data for
   particulate emissions from solid
   fuel-fired sources are inadequate.
   Data collected in this study exhibit
   large variability and contribute
   little to the limited information
   contained in the existing particle
   size distribution data base.
• Trace element emissions from un-
   controlled  coal-fired combustion
   sources are of concern.  Elements
   with ambient severity factors in
   excess of 0.05 include aluminum,
   barium, beryllium, calcium, chlo-
   rine, cobalt, chromium, copper,
   fluorine, iron,  potassium, lithium,
   sodium, nickel, phosphorus, lead,
   silicon, and vanadium. Emissions of
   other elements  also could be of
   significance given the variability of
   the elemental content of coals.
• Nickel emissions form distillate oil
   combustion sources, and nickel,
   chlorine, chromium, and vanadium
  emissions from residual oil com-
  bustion sources are of concern.
  Ambient severity factors for these
  elements exceed 0.05.

• Flue gas emission data for POM
  compounds from gas- and oil-fired
  commercial/institutional sites ap-
  pear  to  be  adequate. Emission
  levels are generally low and the
  compounds that were detected
  have  relatively high MATE values.
  Ambient severity factors  for the
  compounds detected are all below
  0.05.

• Flue gas emission data for POM
  compounds from solid fuel-fired
  combustion sources are still inade-
  quate. Level II techniques should
  be  used  to study emissions from
  small coal- and wood-fired com-
  bustion sources. The effect of heat
  input levels, on/off operating
  modes, excess air levels, and other
  operating parameters on POM
  emissions should be studied in
  more detail. If these factors are
  found to contribute significantly to
  POM  emissions, studies to deter-
  mine  the prevalence of contribu-
  tory source operating parameters
  in the commercial/institutional
  sector should be undertaken to
  establish the magnitude of the
  problem.
 N. F. Surprenant,  P. Hung. R. Li, K. T. McGregor,  W. Piispanen, and S.  M.
  Sandberg are with GCA Corporation, Bedford, MA 01730.
 Michael C.  Osborne is the EPA Project Officer (see below).
 The complete report, entitled "Emissions Assessment of Conventional Station-
  ary Combustions Systems: Volume IV. Commercial/Institutional Combus-
  tion Sources," (Order No. PB81-145187; Cost: $17.00, 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:
        Industrial Environmental Research Laboratory
        U.S. Environmental Protection Agency
        Research Triangle Park. A/C 27711
                                                                                     « U.S. aOVEBNMENT PRINTINC OFFICE 1»1 757-012/7071

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300

-------