United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
 Research and Development
EPA/600/S2-89/006 Aug. 1989
 Project Summary
 N2O Field  Study
 Russ Clayton, Alston Sykes, Rudi Machilek, Ken Krebs, and Jeff Ryan
  Nitrous oxide (N2O) emissions from
coal-fired  utility boilers were  mea-
sured at three electric power gen-
erating stations. Six units were test-
ed, two at each site, including sizes
ranging from 165 to 700 MW. Several
manufacturers and boiler firing types
(circular, triple cell,  and  tangential)
were represented.  Continuous emis-
sion monitor  (CEM) measurements
were made  for nitrogen oxide (NO),
oxygen  (O2), carbon dioxide  (CO2),
and carbon monoxide  (CO). On-line
N20 measurements were made using
a  gas  chromatograph  with  electron
capture  detection.  On-line  sulfur
dioxide  (SO2)  levels  were measured
on one unit using a gas chromato-
graph with flame photometric detec-
tion. Stainless  steel  sample con-
tainers were used to collect flue gas
samples  for the evaluation  of  N2O
formation as a function of time in the
presence of NO, SO2, and water. The
N2O on-line  results were in the range
from "not detectable" to 4.6 ppm. The
results did not exhibit apparent differ-
ences related to different boiler types
and load  conditions.  The stainless
steel  container results  showed  a
trend  of immediate N2O formation
when SO2 and water are present. The
N2O concentration  in all wet samples
increased dramatically within the first
hour to concentrations  ranging from
9 to 120 ppm. The dry sample results
showed a much smaller increase  in
N2O, ranging from 5 to 25 ppm during
the 2 weeks  after sampling.
  This  Project Summary was  devel-
oped by EPA's Air and  Energy Engin-
eering Research Laboratory, Research
Triangle Park.  NC,  to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
  Concern over atmospheric emissions of
nitrous oxide  (N20) has  been  growing
because of increasing ambient concen-
trations and  the suspected connection
with  stratospheric ozone  depletion  and
expected  worldwide climatic  changes.
Combustion sources, such as coal-fired
utility boilers,  have  been  singled out as
possible  major  contributors to overall
N2O  emission levels. Analyses  of stack
grab  samples in stainless steel (SS) con-
tainers known as "bombs"  have shown
emission  levels as high as  hundreds of
parts per million  of  N20 for  these
sources. Concerned about these  emis-
sion  values, the Environmental Protection
Agency (EPA) has initiated a national and
international research program  to char-
acterize N2O emissions. Recent research,
however, has cast doubt on the validity of
previous  measurements  of   N2O
emissions.
  Laboratory  results suggest that a rapid
reaction occurs with nitrogen oxide (NO)
in the presence of water (H2O) and sulfur
dioxide (SO2)  in the stainless steel con-
tainers to form N20. Preliminary results
from  pilot-scale testing indicate that N2O
levels in  SS  container samples  taken
from  a coal-burning combustor  increase
from   <10 to  >100 ppm  within  1 h.
These tests suggest that  previously re-
ported N2O emissions may be incorrect.
Furthermore,  they indicate that exposing
the flue gas to desiccant  such  as P205
prior  to filling  the SS sample containers
helped to moderate  the increase in N2O
concentration,  yielding a more  represen-
tative result.
  To test  these observations on full-scale
combustors, EPA proposed a field survey
of N2O emissions  from  several  utility
boilers. Several  units representing  two
boiler types  (i.e., tangentially  fired  and
wall  fired) and  various  siaes  were
selected  on   the  basis of availability,

-------
location, scheduling, and costs. Schedule
constraints  limited  sampling locations to
available  ports  at  a  point between  the
particulate  emission  control device  and
the stack.
  A two-fold approach  to sampling  was
used N20 levels were determined on-line
and compared with samples collected in
the SS containers.  Both on-line and SS
container samples were analyzed by gas
chromatography (GC) with electron cap-
ture detection  (ECD). The SS container
samples were taken before and after moi-
sture removal (by refrigeration and P205
desiccation) to  determine the effect of
moisture  on N20  formation.   Each SS
container sample was taken and analyzed
for N20 and S02 at selected intervals to
determine and plot  the N20 reaction over
time. On-line measurements  of  carbon
monoxide  (CO), carbon dioxide (C02),
NO, oxygen (02),  and  S02  were also
taken  to provide  additional  operating
characteristics of each boiler unit. Initially,
on-line  nitrogen dioxide (N02) and SO2
were also to be measured but were not
because of instrument  malfunction  On-
line S02 was measured  at Unit F, using a
GC with flame photometric  detection
(FPD).  Continuous  emissions monitors
(CEMs), GCs, and  associated  calibration
gases were housed in two self-contained
vehicles and transported to  the selected
sites. Multiple  GCs were used  to allow
simultaneous on-line and SS sample con-
tainer testing. The  test  series  field work
was conducted October 2  through  13,
1988.

Project Description
  Samples  were taken  from  six  utility
boilers  on three separate sites. All boilers
were burning medium sulfur coal but the
exact sulfur content of  coal at each site
was not available. The six sampled units
are shown in Table  1.
  The first three units were similar in  size
and represented two manufacturers  and
two firing types. Units  D and  F allowed
data comparison on  two identical units;
Unit  E was a  similar  type but was four
times larger. This selection of boilers pro-
vided sufficient variability to  determine
whether  N2O  emissions  from  utility
boilers  are significant and are affected by
boiler design criteria.
  A  GC/ECD was  used  for the  on-line
N20 measurements,  while  CEMs  were
used for  measuring  the fixed  gases
Samples were also collected m SS con-
tainers  to evaluate the stability of N20 in
grab samples  The grab samples  were
analyzed in the field at 1 and 4 h intervals
after collection  and shipped back to  the
EPA/RTP   laboratory  for  subsequent
analyses at 40, 168,  and 336 h  periods
after collection
Sampling Procedures
  Specific  boiler operation data  (i.e.,
firing  rate,  02) were taken from  trans-
ducers already in place.  Figure  1  illus-
trates the CEM system. Sample gas was
drawn through a heated filter and  passed
through  a  heated sample line  to  a tee.
One  stream was directed  to  the S02
monitor  The second flow stream  contin-
ued  to  a  low-contact-time  refrigerated
moisture removal system.  The sample
passed through a sample pump, silica gel
desiccator, and was directed  into  a
distribution manifold. A slipstream  of the
sample  passed through a second pump
and  into  the on-line GC. Individual
sample  flows  were directed to  CEMs as
required.
  Stainless steel container samples were
collected in both a wet and  dry gaseous
environment for locations shown in Figure
1. The wet sample was collected from the
heated  portion  of the sample system,
upstream of any water removal  device.
The  dry sample was  collected from  a
point downstream of the  refrigerated
condenser but  ahead  of the  silica gel
desiccant  dryer  and  passed through  a
container filled with PaOs before entering
the container. A pump,  operating at  a
flowrate of > 1 L/min, was used  to draw a
sample  into the  container.  This flowrate
was   maintained through  the  sample
container for at least 5 min before clos
the valves on both ends of the contaim

Quality Assurance (QA)
  During each test a quality control ((
check  sample was analyzed  to  ve
instrument calibration. The QC  chi
sample was not a calibration point. 1
levels of QC checks were used,  15.1 <
104  ppm   All  QC check results  w
within the ± 20%  acceptable limits of
QA plan, including the sample line che
The highest deviation was 10.5%.
  An  independent audit was conduc
by Research Triangle  Institute. RTI s
plied  a cylinder  of  N20  in  N2-  1
concentration of N20 in this cylinder v
not known to the sampling team.  1
cylinder was analyzed by direct inject
into the GC and also through the sam
loop  connected  to the sample conditii
ing system. The  sampling location
given in Figure 1  The measured cono
tration  for  both injection methods  vi
38 7  ppm, indicating no sampling systi
bias.

Results and  Conclusions

On-Une Results
  The N20 on-line results, summarized
Table 2, were  in  the  range from "i
detectable" to 4.6  ppm. Table 3 summ
izes  the  CEM  average concentratioi
The results showed no trends with  va
ing boiler type, or operating and lo
conditions. The  detection limits for N
varied depending  on the ECD used: t
ECDs were required for each unit test*
Each  ECD exhibited  different  detecti
limits that varied throughout the test. T
variability was due to sample effects a
signal/noise  ratio  changes. These d<
agree with recent EPA in-house studi
that show N2O levels ranging from 1.2
5.4 ppm  from the combustion of vario
coal types.
  Because very little  on-line SO2 de
were  obtained,  no conclusions can
made. Data were collected only on Unil
Table 1.
Unit
A
B
C
0
E
F
Boilers tested
Date
1 0/05/88
f 0/07 '88
70/08/88
70/70/88
10/1 r/88
10/12/88
for emissions
Size (MW)
250
250
250
765
700
765
Class
Pre-NSPS
Pre-NSPS
Pre-NSPS
Pre-NSPS
Pre-NSPS
Pre-NSPS
Manufacturer
B&W
B& W
Riley
CE
CE
CE
Firing Type
Circular
Triple Cell
Circular
Tangential
Tangential
Tangential

-------
       Stack

         . Sample &
         Bias Cal
         Valve
                                GC Wet
                                Samplf
                                Port
Probe
          •»
               t-
                               Pump l^T]
                 Heated
                 Filter
                      Heated
                      Sample
                      I ine
                                              Rcfri(/er,ttt'
-------
Table 3
Unit
A
B
C
D
E
F
OEM Summary -
CO2(%)
14.8
13.5
14.3
11.7
13.1
119
Average Concentrations
NO (ppm)
386
513
559
354
374
319
CO (ppm)
-
13.3
8.6
2.2
30.7
3.1
02 (%)
4.6
7.1
6.1
8.3
6.0
8.1
"Missing data.
 between 11:54 a.m. and 2:46 p.m. Initial
 concentration was measured at 778 ppm
 and gradually  increased to  1004  ppm. A
 total of 25 injections were made,  with an
 average concentration of 924 ppm

 SS Container Results
   SS container N20 wet sample  results
 demonstrate that N20 formation with S02
 present is immediate.  All  wet  samples
          increased dramatically  within the  first
          hour to concentrations ranging from  9 to
          120 ppm. The N20 dry sample results
          show a much smaller increase to  5-25
          ppm over the 336 h time period.  These
          results verify the same  phenomena ob-
          served in  the laboratory studies for  both
          the wet and dry samples for all  fuels
          (coal, fuel oil, natural gas).
 Russ Clayton, Alston  Sykes, Rudi Machilek,  Ken Krebs, and Jeff wen are  with
  Acurex Corp., Research Triangle Park, NC 27709.
 William P. Linak is the EPA Protect Officer (see below).
 The complete report,  entitled "N20  Field Study," (Order No. PB 89-166 623/AS;
  Cost: $15.95, subject to change) will be available only from:
        National Technical Information Service
        5285 Port Royal Road
        Springfield, VA22161
        Telephone: 703-487-4650
 The EPA Protect Officer can be contacted at:
        Air and Energy Engineering Research Laboratory
        U.S. Environmental Protection Agency
        Research Triangle Park, NC 27711
 United States
 Environmental Protection
 Agency
Center for Environmental Research
Information
Cincinnati OH 45268
 Official Business
 Penalty for Private Use $300

 EPA/600/S2-89/006
       8°S°Si!lH PROTECTION  AGEMCY
       CHICAGO

-------