United States
                  Environmental Protection
                  Agency
                                Industrial Environmental Researc
                                Laboratory
                                Research Triangle Park NC 27711
                  Research and Development
                                EPA-600/S7-81 -076 May 1982
                  Project Summary
   

"
Environmental Assessment:
Source Test and  Evaluation
Report— B&W  Alliance
Atmospheric Fluidized-Bed
Combustor
                   . J. Kindya, R. R. Hall, C. W. Young, and P. Fennelly
                   A comprehensive emission sam-
                 pling and analysis program was con-
                 ducted on a pilot-scale, atmospheric-
                 pressure,  coal-fired fluidized-bed
                 combustor (AFBC). Screening data
                 were obtained on  organic and inor-
                 ganic pollutants to indicate biological
                 activity. Testing was conducted on the
                 Babcock and Wilcox/Electric Power
                 Research Institute pilot  AFBC at
                 Alliance,  OH (B&W-EPRI/Alliance).
                 This AFBC, with a coal firing capacity
                 of 880 kg/hr (1940  Ib/hr),  was
                 designed  to bridge the gap between
                 smaller bench scale units and the
                 larger,  more expensive demonstration
                 plants.
                   To characterize uncontrolled  air
                 emissions, trace element concentra-
                 tions were measured upstream of par-
                 ticulate controls. Assuming a control
                 efficiency of 99.9 percent, emissions
                 of trace element species would not be
                 a significant problem. This level of par-
                 ticulate control is achievable with con-
                 ventional particle control technology,
                 and is  necessary to meet the utility
                 boiler  Federal New Source  Perfor-
                 mance Standard for paniculate emis-
                 sions of 13 ng/J (0.03 lb/106 Btu).
                 Elemental concentrations do not
                 appear to be significantly different
                 from other FBCs or conventional coal
                 combustion systems.
                                 Air  pollutant emissions  of  total
                                organics were less than for compara-
                                ble conventional  combustion  sys-
                                tems.  Limited analyses for specific
                                polynuclear  aromatic compounds
                                indicated that emissions of these com-
                                pounds are probably not of concern.
                                 Analysis of  laboratory-generated
                                leachates from  solid waste samples
                                reveals that trace metal concentra-
                                tions are well below Federal hazard-
                                ous waste criteria.
                                 Positive test results for mutagenic-
                                ity and cytotoxicity  screening tests
                                require further investigation. Similar
                                bioassay results have been reported
                                for conventional combustion systems
                                and all other FBCs which have been
                                tested.
                                 This Project Summary was devel-
                                oped by EPA's Industrial Environmen-
                                tal  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
                                 The  development of fluidized-bed
                                combustion (FBC) has been supported
                                by the  Federal  government, private
                                industry, and utility groups because of
                                its potential advantages over conven-

-------
tional coal combustion methods. Reduc-
tion  of S02  during the combustion
process is  the  primary  advantage of
atmospheric FBC (AFBC)  because  it
eliminates the need for add-on flue gas
desulfurization equipment. AFBC also
reduces oxides of nitrogen (NOX) emis-
sions. Other advantages include a mod-
est reduction in capital costs compared
to a  conventional coal-fired boiler  and
the capability to burn a wide variety of
fuels including low grade fuels such as
anthracite culm, coal cleaning wastes,
and  industrial wastes. Industrial sized
AFBC units are offered commercially by
several vendors.

  In  a fluidized-bed boiler, a mixture of
coal  and limestone  is supported on a
grid at  the bottom of a boiler. Combus-
tion  air is passed through the grid at
high  velocities, typically 2.2 - 2.4 m/s(4
-  8 ft/s). The upward flow of the air
holds the solids  in suspension, creating
a quasi-fluid that possesses many of the
properties of a liquid. The most impor-
tant  liquid-like  property to  the  boiler
designer is the fact that bed material is
exceptionally  well mixed and  flows
throughout the  system without  agita-
tion.  This  well mixed semiliquid state
produces high  heat transfer rates  and
permits combustion at temperatures of
760  - 930°C (1400 - 1700°F).

  The  comprehensive  sampling  and
analysis efforts discussed in this  report
were based on  a phased approach to
environmental  assessment  as  devel-
oped  by the  Process Measurements
Branch of EPA's Industrial Environmen-
tal Research Laboratory at Research Tri-
angle  Park, NC  (IERL-RTP). The first
phase.  Level 1, involves screening with
sampling and analytical techniques that
sacrifice accuracy and compound speci-
ficity to identify problem areas in a cost-
effective manner.' Level  1 should yield
final  analytical results accurate within a
factor of ±3. These results can be used
to: (1) provide preliminary environmen-
tal assessment  data; (2) identify prob-
lem  areas; and  (3) formulate the data
needed to rank energy  and industrial
processes,  streams  within a process,
and  components within  a stream, for
further consideration in the  overall
assessment. The second sampling  and
analysis effort,  Level 2,  is directed by
Level 1 results and is designed to  pro-
vide  additional,  more specific, accurate
and  quantitative information that  will
confirm and expand the data gathered in
Level 1. The primary focus of the sam-
pling and analytical efforts discussed in
this report was at Level 1. Some Level 2
analytical work was conducted.

Facility Description
and Streams Sampled
  Figure 1 is a simplified drawing of the
B&W/EPRI  AFBC. Crushed presized
coal and limestone are fed from separ-
ate storage bunkers into the common
transport system. The coal/limestone
feed mixture is transported  pneumati-
cally to a splitter where the feed stream
is split into four separate lines. The com-
bustion bed consists of four equal 0.9 x
0.9 m (3 ft x 3 ft)quadrants, and one feed
line discharges into the center of each.
The resultant feedpoint/area ratio pro-
vides even distribution of fuel and.sor-
bent across the entire bed.
  The furnace enclosureof the unitcon-
sists of atmospheric pressure  water
walls with a fireside  refractory lining.
Total height of  the facility is  14.6 m (48
ft). Combustion air enters the plenum
and  is  forced  through the  distributor
plate and into the bed.  During operation,
bed depth is maintained between 1.4
and 2 m (4.6 and 6.6 ft) and is  cooled by a
serpentine arrangement of  11 tubes.
This in-bed  tube  bank keeps  the  bed
temperature between 780 and 870°C
(1440 and 1660°F). A convective tube
bank in the freeboard cools the combus-
tor flue gasto below 480°C(900°F). This
high freeboard improves combustion
efficiency while reducing the dust load-
ing to the primary dust collector
  Paniculate laden flue gases exit the
combustor at the top and split into four
parallel streams in the primary multicy-
clone dust collector. Collected material
is  dumped  from  a single hopper into
208-liter  (55-gal.) drums on  a continu-
ous basis and subsequently disposed of.
Although this material could have been
recycled to the combustor, it was not.
  The flue gas leaving  the cyclones
passes through a venturi f lowmeter and
is  then cooled by a heat  exchanger
before entering the baghouse  for final
particulate removal. The baghouse is a
four-cell  unit with a design filtration
velocity of 0.02 m/s (4 ft/min). Ash col-
lected in the baghouse is discharged to a
collection bin for disposal.
  Spent bed solids are removed from
the combustor via five drain pipes that
extend down from the bed through the
distribution plate and plenum  and dis-
charge into a water-cooled screw con-
veyor. Bed height and, therefore, rate of
spent bed material removal are con-
trolled by measurement of the pressure
drop across the bed.
  Emissions  testing  at  the B&W-
EPRI/Alliance AFBC facility was con-
ducted from December 10 to 14, 1979.
The system had operated at steady state
for several days before testing to ensure
stable bed conditions. Operating condi-
tions are presented in Table 1 as daily
averages for the 2 days when combus-
tor flue gas was sampled.
  As shown in Figure  1,  six  streams
were sampled for chemical and biologi-
cal analysis. These streams were 1 -
combustor  flue  gas,  gaseous
components, and entrained  solids,
sampled by GCA; 2 - solids collected by
final particulate  control device  (bag-
house), sampled by B&W; 3 - solids col-
lected  by the primary  dust collector
(cyclones), sampled by B&W; 4 - spent
combustion  bed  solids,  sampled  by
B&W; and 5 and 6, coal and sorbent feed
streams, sampled by B&W.  Because the
baghouse  had  only  recently  been
installed, ports  to sample flue gases
after the baghouse were not available.
The sampling and handling procedures
were as specified in the IERL-RTP Level
1 procedures manual.1


Sampling and  Analytical
Methodology
  The  Source Assessment Sampling
System (SASS) train was used to collect
size-fractionated particles,  volatile and
nonvolatile organic  compounds,  and
trace  metal components of the  B&W-
EPRI/Alliance  combustor flue  gas
stream (stream 1  in Figure 1)for subse-
quent chemical and biological testing.
The SASS train collects particles  in a
series  of three cyclones with  nominal
cut points of 10, 3, and  1  /urn. A filter
collects particles  smaller  than  1  t*m.
Organic species  were captured  by a
polymeric  sorbent  (XAD-2),  in  a
temperature-controlled trap. A series of
impingers follows the resin trap to cap-
ture volatile trace metals (Hg, As, and
Sb).
  A modification to the standard SASS
methodology  was the use of an air-
cooled probe  to lower sample  gas
temperature from 450°C (840°F) to
approximately 200°C (400°F). The parti-
cles were collected at 200°C (400°F)
while the XAD resin used for organic
collection was maintained below 35°C
(95°F).
  Gaseous components of the B&W-
EPRI/Alliance  flue  gas stream  were
sampled for subsequent  analysis using
a combination of continuous withdraw-
al, grab sampling, and special impinger

-------
train techniques. Analyses were con-
ducted for CO, CO2, O2, N2, S02, H2S,
COS, methylmercaptans, ammonia, and
cyanide.
  Solid streams (streams 2 to 6 in Fig-
ure 1) were  sampled once per hour by
taking full-cut stream samples. Sam-
ples collected during each SASS train
were composited.
  Inorganic constituents  of  B&W-
EPRI/Alliance  stream  samples were
quantified  using  a combination  of
Table 1.    Operating Conditions During the Test Program

           Parameter             Test 1  (12/10/79)
                  Test 2 (12/12/79)
Bed temperature. °C (°F)
Bed height, m (ft)
Fluidizing velocity, m/s (ft/s)
Coal type
Coal feed rate, kg/hr (Ib/hr)
Sorbent type
Sorbent fed rate, kg/hr (Ib/hr)
Coal (sorbent) size
Ca/S mole ration
840 (1550)
1.4 (4)
2.3 (7.5)
Pittsburgh No. 8, 2.4%
872 (1920)
Lowellville limestone
204 (450)
1/4x0(5/16x0)
2.58
840 (1550)
1.4 (4)
2.6 (8.4)
S, 12500Btu/lb
878 (1935)
272 (600)
3.37
Cleaned gases
to atmosphere /T\
instrumental  and  wet  chemical
techniques.
  The primary Level 1 inorganic analy-
sis  technique is  spark  source mass
spectrometry, which provides sensitive
detection limits for about 70 elements.
Interferences, such as variations in the
ion source discharge conditions and the,
photoplate  interpretation  techniques
used  in  Level 1, provide only semi-
quantitative results accurate to within a
factor of 2 or  3.
  B&W-EPRI/Alliance bulk  solid
streams having the potential to be dis-
posed of in a landfill or other area where
leaching  could occur  (this included
spent bed material and cyclone catch)
were subjected to inorganic  analysis of
both the solid material and laboratory-
generated distilled  water  leachates.
                 1
             Baghouse
                                     Primary H=»
                Combustor
                  flue gas
                                       dust     -T
                                     collector \ I
                                     Cyclone ash
                                        outlet
           Baghouse solids
                outlet
                                           Fluidizing air i  —»•  J
Figure 1.   Simplified drawing of the B&W/Alliance AFBC indicating sampling points.

-------
Atomic  absorption spectrometry was
used to accurately quantify selected ele-
ments in the leachates.
  The Level 1 test protocol attempts to
identify the  major organic  compound
classes  within each  sample  stream
tested. Methylene chloride  extracts of
the samples are analyzed.  Qualitative
and some quantitative data are gener-
ated by the use of gravimetry, gas chro-
matography,  liquid  chromatography,
infrared spectroscopy, and low resolu-
tion mass spectrometry. These Level 1
techniques, m general, do not provide
data on  specific compounds. Level 1
methods were supplemented with high
performance  liquid  chromatography
using fluorescence detection for poly-
nuclear aromatic compounds.
  Volatile organics (boiling points below
100°C)  are separated into  six boiling
point ranges  by  onsite gas chromato-
graphy.   Moderately  volatile  organics
(boiling  points 100 - 300°C) in methy-
lene chloride extracts are analyzed in
the laboratory by gas chromatography.
Flame ionization detection  is used in
both of the above cases. Nonvolatile or-
ganics (boiling point above  300°C)  are
measured by gravimetric methods in
methylene chloride extracts. These or-
ganic concentration data  provide some
qualitative indication  of  the  types of
compounds that  may be present
  The Level  1 liquid chromatographic
(LC) separation technique wasdesigned
to separate samples into seven reason-
ably distinct  classes of organic com-
pounds  and was applied to all samples
that  contained _a minimum of 15  mg
combined  volatile  and  nonvolatile
organics. A sample was placed on a sil-
ica gel liquid chromatographic column,
and a series of eluants of sequentially
increasing polarity was  employed to
separate  into  fractions  for  further
analyses.
  Infrared analysis was used to deter-
mine  the functional  groups  in  an
organic  sample  or liquid chromatog-
raphy fraction of a partitioned sample.
The  interpreted  spectra  provide  in-
formation on  functionality (e.g., car-
bonyl, aromatic  hydrocarbon, alcohol,
amine,  aliphatic hydrocarbon, halo-
genated organic).
  Level  1 bioassays are designed as a
cost-effective initial screening tool that
indicates potential  health or ecological
effects. As such,  the test results should
be used to point out areas requiring
further  investigation.  Health  effects
tests  consisted  of the Ames test  for
mutagenicity, and mammalian cell cyto-
toxicity assays  using  rabbit alveolar
macrophages  (RAM)  and  Chinese
hamster ovary  (CHO)  cells. Fathead
minnows, daphnids, and algae were the
organisms used to test for acute ecolog-
ical effects. Table 2 is the test matrix.

Results

Data Handling
  Criteria or  standards for  air, water,
and solid waste pollutants are needed to
properly determine the implications of
test results. Federal standards exist for
some pollutants (e.g., total particulates,
NO*, and SO2) in boiler flue gases, and
Federal criteria are available to deter-
mine if a solid waste is hazardous. How-
ever, emission standards do not exist for
most of the pollutants measured in this
and other environmental assessment
programs.
  IERL-RTP has developed a set of con-
servative Discharge  Multimedia Envi-
ronmental Goa Is (DMEGs).2 These goals
are derived using models  incorporating
available data such as industrial Thres-
hold Limit Values (TLVs), NIOSH recom-
mendations  for  worker  exposure,
drinking water criteria, results of toxic-
ity  experiments using  animals,  and
EPA/NIOSH   ordering   numbers  or
animal data  on carcmogenicity. These
DMEGs  are  emission  concentrations
that are used in IERL-RTP research pro-
grams to provide perspective on poten-
tial environmental  hazards, to provide
direction for  control technology re-
search programs, and to rank emission
streams for future investigation. These
conservative goals  are a screening tool
to provide focus for further, more de-
tailed investigations
  The  simplest  model  incorporates
TLVs  as air  pollutant emission goals.
Table 2.
                                       Emission concentrations below the TLV
                                       are assumed to be safe  dispersion will
                                       usually  produce  ground-level  con-
                                       centrations lower than stack concentra-
                                       tions by a factor of 1000 or more. This
                                       dilution factor, in general,  should pro-
                                       vide adequate protection for chronic ex-
                                       posure of the general population The
                                       other goals and models are based  on
                                       similar but more complex and, at times,
                                       more tenuous extrapolations.
                                         Measured  trace element concentra-
                                       tions for each stream sample were com-
                                       pared to  their  individual  DMEGs,
                                       specific for the  media of interest (air,
                                       water, solid wastes). Elements in excess
                                       of their DMEGs were "flagged," provid-
                                       ing a mechanism to  estimate potential
                                       hazards associated  with emission  of
                                       that stream.  Such  estimates  may  be
                                       made to reflect both  human health and
                                       the  ecological domain  Elements  in
                                       excess of their DMEGs indicate a need
                                       for further  investigation.

                                       Inorganic Analytical
                                       Results
                                         Particulate emissions from the B&W-
                                       EPRI/Alliance  AFBC  were  sampled
                                       before the application  of  paniculate
                                       control  equipment.  The particle size
                                       results show that 80 percent by weight
                                       were larger  than  10 //m,  15 percent
                                       were in the 3 to 10 fjm range, 4.4 per-
                                       cent were  in the 1 to 3 //m range, and
                                       0.85 percent were smaller  than 1 /urn.
                                       Total paniculate concentration was 35 -
                                       40 g/m3 which, at the measured excess
                                       air, is equivalent to about 13,000 ng/J
                                       (30lb/106Btu). Elemental analysis indi-
                                       cates that much of the larger than 10/um
                                       material was limestone elutriated from
                                       the combustor.
                                         Many elements in the combustor flue
                                       gas exceeded their air DMEGs as might
          Bioassay Testing of B&W-EPRI/Alliance Samples

                              Health effects testing
                                       Cytotoxicity       Ecological effects testing
Sample description
                           Mutagenicity
                             Ames
                                       RAM
                                               CHO
                                                       Fish   Daphnids   Algae
Combustor Flue Gasa
  Fine f<3 fjm) particles
  Coarse f> 3 fjm) particles
XAD extract (organics J
Bulk Solids
Cyclone catch
Spent bed material
y

y
y
c

y
y
y

y
y
—

yd
yd
— —

yd yd
/" /d
V V
a Collected by SASS train.
to(j) - this assay performed on the sample indicated.
c(—) - this assay not performed.
"Ecological effects testing was performed on laboratory-genfated distilled water
 leachates of these two bulk solid streams.

-------
be expected  from  the extremely high
particle concentrations. Table 3 shows
emission concentrations for the 25 ele-
ments that exceeded their DMEGs. Also
shown in Table 3 are the discharge se-
verities (the ratio of the emission con-
centration  to the  DMEG)  for  each
element. The highest discharge severi-
ties are for vanadium 2700, chromium
2050, nickel  230, and  phosphorous
160. Note that the DMEGs are based on
the assumption that the emissions are
in their most  toxic form. For example,
the DMEG of 1/jg/m3 is based on hexa-
valent chromium; but if the emissions
were trivalent chromium, a goal of 50
/ug/m3 would be appropriate.
   A full-scale utility boiler would be
required to reduce particulate emis-
sions by 99.9 percent from those mea-
sured at the B&W-EPRI/Alliance AFBC.
This collection efficiency is within the
Table 3.     Trace Element Emissions in B&W-EPRI/Alliance Flue Gas
             Compared to Emission Goals

                      Discharge      Measured       Discharge severity0
                        goal        concentration
Element               (ug/m3)*        (ug/m3)'3      Uncontrolled  Controlled
capability  of  conventional particulate
control devices. The effect of this emis-
sion reduction on trace elements would
depend on the distribution of trace ele-
ments  by particle size, and on both the
type and fractional efficiency of the con-
trol device.
  Discharge severities, assuming 99.9
percent control for all trace elements,
are also presented in Table 3. If this
assumption is valid,  then only chro-
mium and vanadium would exceed their
goals.  These  estimates are consistent
with measurements at the Exxon Mini-
plant, a pressurized fluidized-bed com-
bustor.3
  Emissions of NO? and S02 were conti-
nously  monitored by  B&W. Emission
rates for N02 were 172 ng/J (0.4 lb/106
Btu) and 224 ng/J (0.52 lb/106 Btu),
respectively,  on  the  first and second
days of testing. The AFBC reduced
Lithium
Beryllium
Boron
Fluorine
Sodium6
Magnesium
Aluminum
Silicon
Phosphorus
Calcium
Titanium
Vanadium
Chromium
Manganese
Iron
Cobalt
Nickel
Copper
Selenium
Strontium
Yttrium"
Cadmium
Barium
Lead
Uranium
22
2
3,100
2.500
53,000
6,000
5,200
10,000
100
16,000
6,000
1
1
5,000
1.000
50
15
200
200
3,100
1,000
10
500
150
9
970
82
28,000
36.000
100,000
44,000
233,000
>454,000
16,000
> 74,500
73,000
2,700
2,050
10,300
> 78.000
690
3,450
1.360
445
8.450
1.100
17
7.650
830
74
44
41
9
14
2
7
4
45
160
5
12
2,700
2,050
2
>78
14
230
7
2
3
1
2
15
6
8
0.044
0.041
0.009
0.014
0.002
0.007
0.004
0.045
0.160
0.005
0.012
2.70
2.05
0.002
>0.078
0.014
0.23
0.007
0.002
0.003
0.001
0.002
0.015
0.006
0.008
 *AII goals are air health DMEGs, except vanadium which is an ecological goal.
 b Average of two test runs.
 cDischarge  Severity = measured concentration/DMEG.  Uncontrolled means as
 sampled at the cyclone inlet approximately 35 to 40 g/m3.  Controlled assumes
 particulate control of 99.9 percent which is sufficient to meet NSPS of 13 ng/J (0.03
 lb/106 Btu) for utility boilers.
 "Exceeds health DMEG in test day 1 samples only.
 eExceeds health DMEG in test day 2 samples only.
emission by 79.6 percent to 395 ng/J
(0.92 lb/106 Btu) on the first day and by
67.2 percent to 980 ng/J (2.28 lb/106
Btu) on the second day. These emission
rates reflect only the operating condi-
tions on the test days and not the overall
capabilities  of  the  B&W-EPRI/AFBC
system.  Much  lower emission  rates
have been achieved under other operat-
ing conditions.
  Leachates were generated from the
two solid waste streams that were avail-
able for analysis, cyclone  catch  and
spent  bed  material.  A distilled water
method  in accordance  with Level 1
procedures was used Analytical results
show that  most elements were below
detection limits  of spark source mass
spectrometry, and were at least one to
two orders of  magnitude  below the
appropriate DMEG.
  Calcium, nickel,  and  selenium
exceeded DMEGs in both leachate sam-
ples as shown in Table 4. Vanadium
exceeded the DMEG in the cyclone ash
leachate  samples.   The phosphorus
result which was higher than the ecol-
ogy DMEG of 0.5 /ug/l can be dropped
from further consideration because the
ecology  goal is  based  on  elemental
phosphorous and is thus not applicable
to AFBC. Calcium was  far above the
DMEG and can  be  attributed to the
limestone in both waste streams. These
elements may require further study.
  The leachate samples were also ana-
lyzed for arsenic, cadmium,  chromium,
lead, mercury, and selenium by graphite
furnace atomic absorption spectrome-
try. This method was selected because
of the very low detection limits,  0.1 to 1
ppb (0.1  to 1 ug/\), and the accuracy.
Comparison of these results to  hazard-
ous waste criteria shows that concen-
trations are well below the criteria (see
Table 5) used to determine if a waste is
hazardous.

Organic Analytical
Results
  Organic species of potential concern
appear to be limited to the combustion
flue gas because organics in the other
samples are extremely low. Organics in
the  flue gas particulate (fly ash) total 29
ug/m3 compared to 588 to 885/ug/m3 in
the  flue gas (under the sampling condi-
tions used in the  study). The two solid
waste streams, cyclone catch and spent
combustor bed, contain only 1.6/ug/g
and  7.6  ug/g of methylene chloride
extractable organics,  respectively. A
comparison  of  these concentrations
with the goals for solid waste streams

-------
Table 4.    Inorganic Trace Analysis of Laboratory Generated Leachates -
            Elements that Exceed DMEGs
Emission
goal, fjg/l
Element
Phosphorous
Calcium
Nickel
Selenium
Vanadium
Health
15,000
240,000
230
50
2,500
Ecology
0.5
16.000
10
25
150
Concentration in
Cyclone solids
5
610,000
360
21 to 65
320
leachate, fjg/l
Spent bed material
:.s
610,000
220
76
50
Table 5.
Comparison of Trace Element Concentrations in B&W/Alliance
Leachates and EPA-Office of Solid Waste Criteria Identifying
Hazardous Solid Wastes
                    Cyclone catch
                          Spent bed material
Element
Arsenic
Cadmium
Chromium
Lead
Mercury
Selenium
Day 1
44*
0.3
6.0
<1.0
<1.0
66
Day 2
36
0.2
7.5
1.6
<1.0
41
Day 1
26
<0.1
19
<1.0
<1.0
21
Day 2
30
<0.1
20
<1.0
<1.0
41
OSW limits3'
5,000
1,000
5,000
5,000
200
1.000
"See Federal Register Volume 45, No. 98. Book 2, p. 33111, May 19. 1980.
*AII concentrations are in fjg/l.
indicates that these levels are not sig-
nificant and  do not warrant further
study.
  The significance of higher concentra-
tions of organics in the flue gas is diffi-
cult to evaluate because of the lack of
specific  compound  identification.
Results and the method used to evalu-
ate these results are illustrated in Table
6. Candidate organic compounds within
each  boiling  point  range have  been
compiled and categorized by Harris, et
al.4 for use in environmental assess-
ment  programs. In the lowest boiling
point range, -160 to -100°C, a concen-
tration of 392 //g/m3 was measured.
Available data indicate  that the only
compound that could be both present in
this fraction and that has a lower DMEG
is ethylene with a DMEG of 1 /ug/m3.
This result indicates that ethylene may
require further investigation, either test
work or an evaluation of process chem-
istry to determine if it could be present.
The second boiling point range -100 to
-50°C does not contain any compounds
of potential  concern. The  result for
these first two fractions demonstrates
that many compounds are not potential
problems and specific analyses are not
required. This illustrates the purpose of
the Level 1 approach. The list of organic
compounds of potential  concern has
been narrowed from several hundred to
less than 20 as illustrated in Table 6.
                              Some caution,  however, should be
                            observed in using Table 6 because of
                            uncertainty regarding the species to be
                            found  in each fraction and potential
                            overlap between fractions.
                              A comparison  between AFBC  and
                            conventional results for total organics is
                            presented  in  Table  7.  Uncontrolled
                            emissions  from the B&W-EPRI/AFBC
                            Alliance are lower than those measured
                            from one conventional industrial boiler.
                              Limited  Level 2  chemical  analysis
                            was  performed  on  selected  B&W-
                            EPRI/Alliance samples. The initial Level
                            2 analysis  involved screening selected
                            samples using High Performance Liquid
                            Chromatography (HPLC) with floures-
                            cence  detection.  This  procedure  pro-
                            vides  low  detection  limits   and  is
                            especially  useful  for polynuclear  aro-
                            matic  hydrocarbons  (PNAs).  B&W-
                            EPRI/Alliance  stream samples
                            analyzed by HPLC/flourescence include
                            combustor  flue gas fine  (<3 /L/m) and
                            coarse (>3 //m) particles, nonvolatile
                            organics (adsorbed on XAD-2, extracted
                            with CH2Cl2>, and cyclone hopper ash.
                            Appropriate blanks were also run, and
                            data were blank corrected.
                              Results of analysis of combustor flue
                            gas samples are shown in Table 8. Most
                            compounds are well below the approp-
                            riate DMEG. Only benzo(a)pyrene may
                            be present  at a concentration near the
                            goal of 0.020 fjg/m3. However,  this
compound  cannot  be  differentiated
from  the  much  less  hazardous
benzo(e)pyrene, so its actual concentra-
tion may be below 0.02/yg/m3 in which
case none of the PNAs measured in the
flue gas exceed their DMEGs.

Bioassay Results
  Table 9 summarizes test data from
bioassays of five effluent stream sam-
ples  from the  B&W-EPRI/Alliance
AFBC. Included  are results of testing
three combustor flue gas samples and
two bulk solids (cyclone catch and spent
combustion bed material). Results are
shown as positive/negative for mutage-
nicity in S. typhimurium for the Ames
test, and as the appropriate toxicologic
parameter (LCso, ECso, etc.) for the re-
maining test. These data have been as-
sessed qualitatively using tire currently
recommended Level  1 criteria for bio-
assay response.
  Of the five stream samples tested for
mutagenicity, only combustor flue gas
organics (XAD  extract)  had   positive
response.  Data reveal mutagenic activ-
ity for this sample in frameshift tester
strains TA 1537,1538, and 98. The low-
est recorded concentration (dose) that
elicited a positive response (LRPC) was
12 //g CH2CI2-extractable material. Rel-
ative mutagenic activity of this sample,
calculated using the number of rever-
tants at the LRPC for the most active
strain  (TA 98), is eight revertants///g
organics.
  Ames test data also reveal higher spe-
cific mutagenic activity in the absence
of S-9 microsomal mix. Addition of the
activation system generally   reduces
mutagenic response, similar to results
observed from testing of Exxon PFBC
samples. These results  indicate  that
polynuclear  aromatic hydrocarbons
(PNAs) cannot be the sole mutagenic
species responsible for the  observed
activity. The very low levels of PNAthat
were measured by HPLC/flourescence
also  indicate  that  other  compounds
should be investigated.
  Mutagenic activity in the Ames assay
is  not  a  quantitative   measure  of
potential  hazard to  health   or  the
environment.  The Ames assay is only
one of the battery of Level 1 screening
tests  in  which  a positive response
indicates  the   need  for   further
investigation  and  analysis   of  the
sample. B&W-EPRI/Alliance Ames test
results can be compared with those
from  other  units,  but  absolute
judgments  of  the  numerical
relationships  of these results  are not

-------
 Table 6.     Combustor Flue Gas Organic Analytical Results
               Sample
          fraction designation
            Measured
         emissions, fjg/m3
  4
  5
  6
  7

Paniculate
  Coarse f>3 urn)
  Fine (<3 (imj
                10.4
                17.2
               110.8
               144.7
              no data
                  30
    Organic compounds of
     potential concern3
Gaseous - boiling point range^'4
-160 to -100°C
- 100 to -50° C
-50 to 0°C
0 to 30°C
30 to 60° C
60 to 90°C
90 to 100°C
Volatile liquid chromatography
1
2
3
4
5
6
7
Nonvolatile liquid chromatography fraction
1
2
3

392
<123
5,388
22,260
<294
<351
<408

0.8
0.9
1.8
not detectable
1.2
14.5
4.4

59.7
1.4
6.7

One-ethylene (ecological-base)
none
two - formaldehyde, vinyl chloride
none
one - acrolein
none
—

none
none
one - PCBs
none
one - N-nitroso dimethylamine
none
none

—
none
three, B(ajP, 1 ,2:5,6-dibenzanthracene.
3-methyl cholanthrene
none
none
one - dibenzo (c,g) carbazole
two - 2,4,6-trinitrophenol,
3,3-dichlorobenzidine

Insufficient information to assign
probable compound types
  These organic compounds were not specifically detected in flue gas samples but they are within the indicated boiling point range.
  Emission goals of these samples are below measured total emissions; therefore, they have the potential to be present.
 Table 7.    Emission of Total Organics from the B&W/Alliance AFBC Compared to a Conventional Coal-Fired Boiler
                                                                      Organic emissions, /jg/m3
Conventional coal-firing
Organic component
Volatile (bp 100 - 300°C)
Nonvolatile fbp >300°Cj
Total organic component
B&W/ Alliance"
117
673
890
Scrubber inlet
810
5400
6210
Scrubber outlet
652
788
1440
a Average of two test runs.
meaningful.  Test  protocols  vary
between laboratories, and even those in
the research field have not decided how
Ames data can best be presented to
facilitate  just   these  types  of
comparisons. However, a positive Ames
test response,  using  agreed-upon
criteria, is a valuable screening tool.
  Bearing  in  mind the  limitations
mentioned above,  some comparisons
between Alliance results and other data
are still instructive. The result for TA 98
(8 fjg/g organic)  is similar to  that
calculated from  the  testing  of a
comparable  sample from  the  Exxon
Miniplant PFBC.3 All positive responses
detected were  of  the direct-acting
frame-shift   category.  Metabolic
activation did not enhance mutagenic
activity.
  Mutagenic activity has recently been
reported  in  paniculate samples
obtained from other FBC units. Clark et
al.6  have  obtained positive Ames
response in testing flyash samples from
the45.7-cm(18-in.)diameterFBCatthe
Morgantown   Energy  Technology
Center. Kubitschek and Williams7 have
reported mutagenic activity in fly ash
samples from an experimental 15-cm
   (6-in.)  diameter  FBC operated at the
   Argonne' National  Laboratory.  This
   response was seen to vary in specific
   activity   at  lower  temperatures
   presumably because  of the increased
   deposition  of  mutagenic  species by
   condensation mechanisms on particle
   surfaces.  The  higher  paniculate
   sampling  temperatures  at  B&W-
   EPRI/Alliance would result in trapping
   of volatilized substances by the XAD-2
   resin.
     It  must  be emphasized that  the
   positive mutagenic response observed
   for these  several FBC effluent streams

-------
is a  result  unique neither  to FBC
processes  nor to  coal  combustion
effluent streams in general. A survey of
the literature reveals that production of
atmospheric  mutagens   is  not  a
phenomenon limited to certain specific
air emission sources. Mutagenicity has
been  reported  in the Ames assay of
conventional combustion and ambient
air samples collected near various sites
both in the U.S. and elsewhere.
  Of the four samples assayed, the fine
particles (<3 fim) from the  combustor
flue gas stream elicited the highest toxic
response in RAM cells based on viability
index. The coarse paniculate (>3 /urn)
sample  exhibited  a relatively  lower
toxicity.  Several  investigators have
demonstrated a  direct  relationship
between decreasing  particle size and
increasing cytotoxic response.  This is
corroborated by these data.
  Three B&W-EPRI/Alliance samples
were  assayed  using the  CHO test
system.  Combustor  flue  gas  XAD-
extract  was  the  most active  of  the
samples tested. This sample caused 50
percent reduction in plating efficiency at
a concentration of only 1.2 //I/ml. This
response corroborates the biological
activity observed for this sample in the
Ames test.
  Data indicate that  both the cyclone
catch  and  bed reject samples were
Table 8.    PNA Emissions from B&W-EPRI/Alliance AFBC Compared to DMEG
           Emissions3
PNA species*
Naphthalene
Anthracene
Fluoranthene/P//-e/ie
Benz (a) anthracene/
Triphenylene/Chrysene
Benzo (k) fluoranthene
Benzo (a) pyrene/Benzo(e)pyrene
7,12-Dimethylebenzoanthracene,
1 ,2,3,4-Dibenzoanthracene
1.2,5, 6-Dibenzoanthracene
Indeno (1,2,3-cd) pyrene
Emission
goal (/jg/rrff
50,000
56,000
90,000

45
1,600
0.02

0.26
0.093
1,630
B&W-EPRI
Alliance
<2.6
<0.043
0.34

0.13
<0.0003
0.024

<0.078
<0.0022
0.34
a Represents gaseous emissions adsorbed on XAD-2 resin and extracted with
b Where two or more species co-elute, the emission goal of the most hazardous is
 listed (underlined).
c Air, health Discharge Multimedia Environmental Goals (DMEGs).
highly toxic (less than 20 percent; see
Table  9) to  fathead  minnows  at
concentrations  of   12.3  and  13.5
percent, respectively. LCso values in the
daphnia  test for  both samples were
similar to those obtained in  the fish
bioassay tests. The 48-hour LC50 values
for the bed reject and cyclone samples
were  12.3  and   10.7  percent,
respectively, which is considered high
toxicity  High toxicity to algae was also
demonstrated by both samples.
  Even though testing of leachates in
these ecological effects assays resulted
in uniformly high toxicity responses in
all tests for both leachate samples,
these results are of limited use. The
high toxicity was probably a  result of
increases in pH, to levels fatal to the test
species.  Although increased  pH is  a
valid endpoint to be measured in these
tests, toxicity resulting from increased
pH masks the effects that result from
chemical constituents of these leachate
samples.

Conclusions and
Recommendations
  Air emissions of trace elements from
the  B&W-EPRI/Alliance AFBC do not
appear  to  present  a  significant
environmental concern when these
emissions are evaluated using available
emission  DMEGs. This  conclusion is
based  on particulate  measurements
which were made before the use of any
control device and to which a "control
factor" of 99.9 percent was applied.
Measurements after a final particulate
Table 9.    Results of Biological Testing of B&W/Alliance Effluent Stream Samples

                                         Health effects testing
Mutagenicity Toxicity
Sample identification
Fine particles f<3 (im)

Coarse particles f>3 fjm)

Flue gas organics

Cyclone catch'

Spent bed material*

Ames RAM
(+/-f (LCsof
60
(MODf
785
(LOWf
+ NT

435
(LOW)"
295
(LOWf
CHO
(LCso)
NT

NT

1.2 (fjl/ml)
(HIGH)
142
(LOWf
358
(LOWf
Ecological effects testing
Fish
(LC50f
NT

NT

NT

12.3
(HIGHf
13.5
(HIGHf
Daphnia
(LCso)
NT

NT

NT

10.7
(HIGHf
12.3
(HIGHf
Algae
(LCso)
NT

NT

NT

8.7
(HIGHf
7.9
(HIGHf
"+/- = positive or negative mutagenic response.
b/.C50 reported as fig/ml except where indicated.
cLCso for fish tests is 96 hour; for Daphnia 48 hour.
"LOW = low toxicity; MOD = moderate toxicity; HIGH - high toxicity.
e NT = sample not tested in the indicated assay system.
leachates of these materials were tested in ecological effects tests.
                                  a

-------
control device should be conducted to
confirm that 99.9 percent control can be
achieved.
  The  positive  mutagenicity  and
cytotoxicity results require some further
investigation. Additional bioassays and
chemical analyses are recommended.
References
   1. Lentzen, D.E., D.E. Wagoner, E.D.
     Estes, and W.F. Gutknecht. IERL-
     RTP Procedures Manual: Level 1
     Environmental Assessment
     (Second  Edition). Prepared  by
     Research Triangle Institute, Rea-
     search Triangle Park, NC, for the
     U.S.   Environmental  Protection
     Agency. Report No. EPA-600/7-
     78-201  (NTIS  PB  293795),
     October 1978.
   2. Cleland, J.G. and G.L Kingsbury.
     Multimedia  Environmental Goals
     for  Environmental Assessment,
     Volumes I  and II.  Prepared  by
     Research  Triangle  Institute,
     Research Triangle Park, NC, for
     the U.S. Environmental Protec-
     tion Agency.  Reports No. EPA-
     600/7-77-136a and b (NTIS PB
     296919 and 276920), November
     1977.
   3. Kmdya, R.J., R.R. Hall, G.H. Hunt,
     W. Piispanen, and P.P. Fennelly.
     Environmental  Assessment:
     Source Test  and  Evaluation
     Report — Exxon  Miniplant Pres-
     surized Fluidized-Bed Combustor
     with Sorbent Regeneration. Pre-
     pared by GCA/Technology Div-
     ision,  Bedford, MA, for the U.S.
     Environmental  Protection
     Agency. Report No. EPA-600/7-
     81-077, April 1981.
   4. Harris,  J.C.,  M.J.  Hayes,"  P.L
     Levfns,  and   D.B.  Lindsay.
     EPA/IERL-RTP  Procedures for
     Level 2 Sampling and Analysis of
     Organic Materials.  Prepared by
     Arthur D. Little, Inc., Cambridge,
     MA, for the U..S. Environmental
     Protection Agency.  Report  No.
     EPA-600/7-79-033  (NTIS  PB
     293800), February 1979.
   5. Leavitt,  C.,  K. Arledge,  W.
     Hamersma,  R. Maddalane, R. Bei-
     mer, G. Richard, and M. Yamad.
     Environmental  Assessment of
     Coal -  and  Oil-Firing in a Con-
     trolled  Industrial Boiler, Volumes
     I, II, and III. Prepared by TRW, Inc.,
     Redondo Beach, CA, for the U.S.
     Environmental  Protection
     Agency. Reports No. EPA-600/7-
   78-164a-c  (NTIS  PB  289942,
   289941, 291236), August 1978.
   Clark, C.R., R.L. Hanson, and A.
   Sanchez. Mutagenicity  of Efflu-
   ents  Associated  with  the
   Fluidized-Bed  Combustion of
   Coal.  Inhalation  Toxicology
   Research Institute Annual Report
   1977-1978,  pp. 274-284.  Pre-
   pared by Lovelace Biomedical and
   Environmental  Research Insti-
   tute, Albuquerque, NM, for the
   U.S.  Department  of  Energy,
   December 1978.
   Kubitschek,  H.E.  and D.M.
   Williams. Mutagenicity of Fly Ash
   From a Fluidized-Bed Combustor
   During Start-up and Steady Oper-
   ating  Conditions.  Mutation
   Research 77:287-291, 1980.
R. J. Kindya. R. R. Hall, C. W. Young, and P. Fennelly are with GCA/Technology
  Division, Bedford, MA 01730.
John O. Milliken is the EPA Project Officer (see below).
The complete report, entitled "Environmental Assessment: Source Test and
  Evaluation Report—B&W Alliance Atmospheric Fluidized-Bed Combustor,"
  (Order No. PB 82-186 537; Cost: $12.00. 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, NC 27711
                                A U.S. GOVERNMENT PRINTING OFFICE: 1982-559-092/3412

                                    9

-------
                                               18
                                               |«
                                               •«>
                                                            3


                                                            I

                                                            o


                                                            *.
                                                            o\
                                                            N>
                                                            a>
                                                            co
nrvxc-o
orwm  w
t-iocio;
o  »-»
jxc
 T> n
                                                       -o * S <
                                                       i  3 (D =
                                                       W O O C
                                                       w< s. :
                                                       w   o :
                                                           3 5

-------