-------
Uranium in Coal

     The data presented in Table 3-52  indicate that the  uranium content of
the four major coal types does not vary significantly.   However,  lignite
coals have a slightly higher mean uranium concentration  than the remaining
three coal types.   Bituminous and subbituminous coals  have a wider reported
range of uranium concentrations, up to 59 and 76 ppm for these two coal
types, respectively.   The means listed in Table 3-53 may be viewed as typical
values for uranium in coal because they are based on the most complete data
set currently available.   However, the standard deviations about the means
are greater than the means themselves, indicating variability in the data
set.
     Table 3-54 lists means and ranges of uranium in coal by geographic
region.  There is not a large difference in mean uranium concentrations
among coals from these regions.  But coal from the Western Interior and the
Gulf Province have higher mean concentrations of uranium than do coals  from
other regions.  The means listed  in the table can be regarded ^s typical  for
coal from each region.
     The uranium-238 concentrations in coal from five states and one  region
are given in Table 3-55.   Highest uranium-238 concentrations are seen in
coals from Kentucky and Colorado, 0.91 and 0.877 pCi/g, respectively.

BEHAVIOR OF TOXIC POLLUTANTS DURING COMBUSTION

     Trace metals contained  in  fuels  are  released  during  the  combustion
process.  They may be retained  in the  bottom  ash,  or  they may be  emitted  via
the flue gas.  Trace  elements present  in flue gas  may be  contained in the fly
ash or they may be in vapor  form.   Polycyclic organic matter  (POM) is also
formed during combustion  and emitted  to  the  atmosphere.    This section
describes the behavior of trace metals and radionuclides  during combustion
processes and discusses  the  formation/transformation of POMs  and
formaldehyde.
MCH/007       .                        3-59

-------
          TABLE 3-52.  CONCENTRATION OF URANIUM IN COAL BY COAL TYPEC
Coal Type
Bituminous
Subbituminous
Anthracite
Lignite
Number of
Samples
3527
640
52
183
Uranium Concentration (ppm)
Mean
1.85
2.13
1.94
3.37
Standard Deviation
2.71
3.84
3.38
10.3
  Data presented in White et al..  1984; based on data in the USGS National
  Coal Resources Data System (NCRDS) as of 1982.  Arithmetic means from this
  study may be used as typical values for uranium in coal.
      TABLE 3-53.  RANGES OF URANIUM CONCENTRATION IN COALS BY COAL TYPE
   Coal Type
Uranium Concentration

     Range (ppm)
  Bituminous

  Subbituminous

  Anthracite

  Lignite
       <0.2-59

       0.4-76

      0.3-25.2

      0.5-16.7
  a
   Lowest  and highest values reported in the literature reviewed.  Note:  The
   White e_t  al..,,  (1984) study does not list the range of values in the NCRDS.
   The Swanson  et al.. (1976) study, a subset of the NCRDS containing data on
   about 800 coal samples does provide ranges.  This table is based primarily
   on the  Swanson et  al.. (1976) study and Valkovic, (1983a).
MCH/007
                                    3-60

-------

















z
o

a
as

£M
ea
j
8

z


z
o
1— 1
H
PS
ty>
U
CJ
z
8

13
>— I
25
2
3


^
in
I
CO
Cd
H




























41
u
C
4)
U
4)
IM
41
OS





e
•o o
1-1 -H
CO *J
•a eo
C -H
<0 >
/^ AJ 41
e co o
a
c
0

•U U
CO -H

U m C
e e  > O O O
4) CD O 4> O co 4) CO 0) to
ue.* uji e ue --> c
•H CO *— * '^ •** CO -^ CO -^ CO
-C 3 co .Ceo 3 J=3 J= 3
3 co > 5 > en 3 co 3 co



r-. I I p~ I I -» I «* I
00 1 1 O 1 1 vO 1 CM 1
• 1 1 • 1 1 • 1 • 1
— < in CM CM








« <0 co CO c
\0 oo r^ ^^
\O ^f ^3 O^ ^ v/^ O C*4 u^ C^
^*^^^^ CS<^ ^ CO f^ «^«4O






•
iT»
* ^^ \o r^ ^^
O ^^ u^ * * *
1 -H *-H 1 1 ^ I \O 1 CS
III 1 O 1 l-«ll
1 CS -H | CS ~H II 1 CN
• • • f^ l/^ *
o o o • o
V o o v






ON •»* 1 CM 1 \o OO *^ *^ CO
^^ CO 1 ON 1 tf^ CO CO P^ C^N
r^ co I in i co
CM


T3 CD
e c
• * 4) -^
CO U CO
c co e -->
a « •••* a,
••^ >
j= u ca a e
o o --i I* i*
CO -H O 04 41
-" 1- C JZ
co 
41 <0 O
4-1 C -^
•-( CO —
J= 3 to
3 co >



5 ! !
• i i
«^








0
5 vO 00
CM — 1 «N







00
vO
co r-
1 CM I
1 CM O
• •
O 0
V






CM «3- 1
~4 CO 1
m — i I




e
• i«4
CQ
4J
c
3
Q
T?

X
^
u
o
ad
MCH/007
3-61

-------






















y*^V
"O
41

C
C
o
u
z
o
M
S
od

^4
CO
t 1
^
s

z
1— 1
z
o
t— 1 '
H

H
z
u
CJ

1
x

^
2
3
B
3


^
1
e*^

Ed
-J
B































41
U
C
01
kl
01
41
05
















^N
g
a


e
a
^
«
u
4-1
fi
0
e

t^j

\

t
o
jt







e
•^ o

ea •"
c — <
ca >
u 4)
01 a





u

SB
ca
JS 41
71 s
u









4)
oo
B
CO
O5








<4-4
O ca
41
u •-•
4i a.
|I
3 oj
z













c
o

oo
41
05




en
a
05
o
z
rS
^* ^^
00 «^
Q\
-^ *
- -1
J «'
ea 4j|
oi
4-1
cj e
0
oi co
4-> B
— t CO
J= 3
3 01








3 I
. i









ca
00

• •
-4 X









eN
1 !A
1 1
1 -*
•
O










r- oo
0 -H


















ed
eo
ea
,••4
«ti
01
u
ai eg
u«
eg eg
4^
u en eg js
O Q T3 U
14-1 05 eg
U B 01
to z Q
41 ca S
3 01 B O
— i u eg ki
eg --4 3 i*-(
* BM co
-4 41 41 01
CO f >-*
O 4) £-• CX
•-i j= S
Q. 4-1 eg
x * to
4-1 g ca
O 
•H -i-l
X 4J 41 -O
ca 41 i-i c
8 JSI-
•0 4-1 S
B -^ X"<
ca u B
§4J
B
r-4 01
A . eo 4J
ea ui S B
—4 Q -H O
•- • aet ca u
CQ U *J
5 z § §
• 41 cj -ft
X 01 JS B
2 S? -2 2
B 3 B art 3
41 -^ O
hi X Z M
V4 J3 T) O
3 41 41 iw
u -o •« js
V 3 w ea
4-1 B -< 01
41 •* O 4t 3
CO eg B U 1-1
4-1 -H B eg
eg B -^ >
u -P4 eu ea
eel eg u <4-i
-3 1 ea - 0
B
V B X O 41
> CD -a ea oo
••4 « 3 e B
ea co 4j eo eg
B X (0 9 M
41 Ol 01
JS *~* 41
41 • co vo g j;
u co 4j r^ o 4-1
a. c eg -. 0
U 00 4) 
SO O eo 4) -^
O U CJ B > 00
«M eg -^
•d —34-10
4) •-' eg c/1 eg 4J
ca eg B u
eg O O 41 B 41
J3 U ._( js 4) ki
4-1 4-1 CO 41
41 B eg 41 js
kl -H Z S U
eg o Ou -o
S U ki 41 41 •
ca 3 u-c ki 13 C
41 ••* C/J 3 O
3 B Q eg 41 -•* -^
—" eg oci 4-1 ki U 00
eg ki U eg O B 01
> 3 Z d S •-< u
eg J3 u
MCH/007
3-62

-------

















z
o
i
OS

act
0

H
H
CO

%4
00

J
^
8

z
1—4

00

CM
1
X
s

o2
a

S

z
o
1— 1
H
as
H
Z
U

z
8


•
tn
10
i
m

Cd
pHJ
3
C_!






























01
u
c
41
14
41

as











y^
01

•H
a
a B
>-• CO
B X
0

^
0

^
B
41
U
B

00 01
 • eo co
•M • •! • as as •
X ^ <-* \ -1 *•"
CO (01 > A
oo ca •* -o * o
C B O CO U co
•p4 a) B W 3 ^
S 4-1 -H O 4J CO
i o c •— --< c a.
i >> o — ' o 4i a.
i s x P-I o * *t


to
to
5

5,
>^
4.J
ao
r-i en
0} CS
O 1
u s
9
lj -i-i
0 B
B U
0 9
00 41
01 f
14 "
^"o

rM B

-

C U <*4

•O eo -^
•^4 oi •a
^ 09
o 8

Ou U )-i
9 *W
4J 4J
01 to «
09 14 »•*
01 CO
CO 4J O
4J -H O
•S "^ s
0) -H
4) £
> *J 09
• ,4 B
09 S O
B -4 -M
01 4.1
A T3 «0
4) S h
14 3 4-1
a. o e
g u-i ID
o <•>
u u B
o o
•< C 0
a]
MCH/007
3-63

-------
Partitioning and Enrichment Behavior of Trace Metals during Combustion

     The concepts of particioning and enrichment are frequently used  co
characterize the behavior of trace elements in combustion processes.
Partitioning generally refers to the split of the trace element among the
various boiler outlet streams:  bottom ash, fly ash, and flue gas.
Enrichment refers to the difference in trace element concentration betveen
different streams or to the change in trace element concentration of bottom
ash or fly ash as a function of particle size.
     One method of describing partitioning behavior is by reporting the
fraction of che total elemental mass input that leaves the boiler via each
of the outlet streams.  Another method is to compare the trace element
concentration of one outlet stream to that of another through enrichment
ratios (or enrichment factors).  In general, enrichment ratios are
calculated by the following equation:
                              CL./C
                     ERy  -   1J
                              Cic/CRc
where
      ER.. - enrichment ratio for element i in stream j
      C..  - concentration of element i in stream j
      CD.  - concentration of reference element R in stream j
       RJ
      C.   - concentration of element i in fuel
       xc
      CD   - concentration of reference element R in fuel
       K.C

      An  enrichment  ratio greater than 1 indicates that the element  is
 "enriched"  in the given stream, or, expressed another way, chat  the element
 "partitions"  to the given stream.  Different reference elements  commonly used
 by various  authors  are aluminum, iron, scandium, and titanium.   These
 elements  are  chosen because their  partitioning and enrichment behavior  is
 often comparable to that for the total mass.  That is, their concentration
 by weight in  all ash streams and size fractions is constant.
 MCH/007       .                       3-64

-------
     Various classification schemes  have  been developed to  describe
partitioning or enrichment behavior  (Klein,  ec al.,  1975b;  Coles  ec  al.,
1979; Baig et al.,  1981).   The classification scheme used by Baig et al.
(1981) is as follows:

          Class 1.   Elements which are  approximately equally distributed
          between fly ash and bottom ash,  or show little or no  small
          particle enrichment.
          Class 2.   Elements which are  enriched in fly ash  relative  to
          bottom ash,  or show increasing  enrichment with decreasing  particle
          size.
          Class 3.   Elements which are  intermediate between Classes  1 and 2.
          Class 4.   Elements which are  emitted in the gas phase.

     Because of factors such as differences in classification schemes used
by different investigators, different and ill-defined dividing lines between
the classes, sampling and analytical errors in the data used to determine
classification, and variations in the behavior of an element in different
studies, it is not possible to make  an absolute classification of the
elements.  However, such a classification scheme is useful  in indicating
general trends in the behavior of the elements.  Several of the elements
have shown behavior characteristics  of each of the first three classes in
different studies.  These elements were assigned to Class  3, since  Classes 1
and 2 represent the extremes  in behavior' and Class 3  is  intermediate between
them.
     Based on  information in  about 20 previous studies,  Baig et  al.  (1981)
classified arsenic and cadmium as Class 2 elements.   Beryllium,  chromium,
manganese, and nickel were placed in Class  3.  Copper was  not  included  in
the Baig et al. (1981) study, but may also  be  placed  in Class  3.  Mercury
behaved as a Class 4 element.  Brief descriptions  of  the behavior of each
element follow:
     As..  Arsenic has exhibited Class 2 behavior in almost every study.
Therefore, As  is considered  to be a Class  2 element (Baig  et al., 1981).
MCH/007                               3-65

-------
     Be.   Beryllium has exhibited Class 1 behavior in some studies, Class  2
in others, and Class 3 in others.  This difference in classification could
be due in part to differences in criteria used to assign elements  to one
class over another, or could be due to differences in the behavior of  Be  in
different combustion systems.  For this study, Be is considered as a Class 2
element (Baig et al., 1981).
     Cd.   Cadmium has exhibited Class 2 behavior in every study examined,
and is therefore considered to be a Class 2 element (Baig et al.,  1981).
     Cr.   Chromium, like Be, has shown Class 1, 2, and 3 behavior  in
different studies, and is considered as a Class 3 element (Baig et al.,
1981).
     Cu.  Copper has shown Class 2 behavior in most studies (Klein et  al. ,
1975b; Mann et al., 1978; Radian Corporation, 1975a; Cowherd, 1975).
However,  Class 1 and 3 behavior has also been reported (Davison et al.,
1974; Natusch et al., 1974; Coles et al., 1979).  Copper is considered a
Class  3 element, but resembles Class 2 more closely than the other Class  3
elements  do.
     Mn.  Manganese has also shown Class 1,'2, and 3 behavior, and will be
considered as a Class 3 element.  However, since  it has been reported  to
show Class 1 behavior more  frequently and Class 2 behavior less frequently
 than the  other Class  3 elements, it may come closer to Class 1 behavior than
 to Class  2 and resemble Class  1  elements more than the other Class 3
 elements  do  (Baig et  al.,  1981).
      Ni.   Nickel  has  shown Class 1, 2, and 3 behavior, and will be
 considered as  a  Class  3 element  (Baig et al., 1981).
      Hg.   Mercury is  a Class 4 element ac normal  stack temperature of  150 C
 (300 F).   Lower  temperatures,  however, will  cause condensation of  some of
 the gaseous mercury so that it can be considered  as Class 2 (Baig  et al.,
 1981).

 Theories Explaining Trace Metal  Behavior  in  Coal  Combustion Systerns -
      Volatilization/condensation mechanism.   One  of the most widely held,
 fundamental theories that has  been proposed  to  explain the behavior of trace
 elements in coal combustion systems  is  the volatilization/condensation
  MCH/007       ".                       3-66

-------
mechanism (VCM).   This theory suggests that volatile  species in the  ash are
vaporized in the firebox, where peak temperatures of  1650 C (3000 F)  are
typical for pulverized coal-fired boilers.   As the flue gas cools to
370-430 C (700-800 F) in the convective heat transfer section and further Co
150 C (300 F) in the air preheater,  the volatilized species condense.   These
species may condense or adsorb onto  existing particles according to  the
available surface area or they may condense homogeneously,  forming fine
particles.  The elements thus volatilized would be depleted in the bottom
ash and concentrated in the fly ash, since the fly ash has more relative
surface area than the bottom ash and since the bottom ash does not come in
contact with the.volatilized elements long enough for the elements to
condense on the bottom ash (Baig et al., 1981).
     The VCM primarily explains the behavior of the Class 2 elements, but it
also explains the behavior of the other classes of elements.  The Class 1
elements are the nonvolatile matrix elements that do  not vaporize in the
boiler.  These elements form the fly ash matrix on which the volatilized
elements condense.  The Class 1 elements are thus equally distributed
between bottom ash and fly ash, and show no small particle enrichment.  The
Class 3 elements apparently are partially vaporized in the boiler, and  thus
show behavior intermediate between Classes 1 and 2.  The Class 4 elements
are highly volatile.  They do not condense or  condense only partially  as the
flue gas cools to normal stack temperature  (Baig et al., 1981).
     The VCM also explains the enrichment  of Class 2  elements  on small
particle sizes.  Because smaller particles have  a higher surface area,
relative to  their mass,  than  the larger particles, they have  more available
area on which Class  2 and  3 elements  can condense.  The  Class  1  elements  are
not vaporized, and  thus  show  no dependence  of  concentration on particle
size.
     Compound boiling points.  Kaakinen et al.  (1975)  have compared
enrichment ratios for several elements to  various  measures of element
volatility,  including melting points,  boiling points,  and vapor pressures of
elemental and  oxide  forms, and reported that  the oxide properties generally
showed good  agreement.
MCH/007                               3-67

-------
     All of the Class 2 and Class 4 elements included in che current studv
(As,  Cd, and Hg) have elemental or oxide boiling points less than  1650 C
(3000 F).   Class 1 elements, such as Al,  have boiling points greater than
1650°C (3000°F).  The Class 3 elements also generally have elemental and
oxide boiling points greater than 1650 C (3000 F),  and so would be expected
Co behave like the Class 1 elements.
     A simple correlation of the element or oxide boiling points thus does
not explain the behavior of all trace elements.  A fraction of these
elements, however, may form compounds other than oxides (such as chlorides
or carbonyls) that are volatile.  Reducing conditions can exist during the
Initial combustion stage that might contribute to the formation of such
compounds.  Moreover, the compounds formed and the fractions of the element
forming the volatile and nonvolatile compounds might vary under different
combustion systems and different conditions of furnace temperature, coal
time/temperature history, excess air, and coal composition.  Such variations
could explain  the observed variation in the behavior of these elements in
different combustion systems (Baig et al., 1981).
     Elemental  association  in coal.  The association of trace elements in
coal  (with the  organic fraction or inorganic matrix) has also been suspected
of playing a key  role  in the fate of elements upon combustion (Mann et al.,
1978; Edwards  et  al, 1980a).  The theory is that trace elements bound in the
organic phase  are atomized  during combustion, while those occluded with the
mineral matter in the  coal  are  less likely to be vaporized.  Moreover,
actual  volatilization  of the organically associated elements may not be
necessary for  trace  element enrichment.  Deposition of the nonvolatilized
trace elements associated with  the  organic fraction, on the remaining
mineral inclusions  that  form the  fly ash, will give a similar inverse
dependence  of  concentration with  size.  This  theory may explain the behavior
of certain elements, but not all  (Baig et al., 1981). .

Theories Explaining Trace  Metal Behavior  in Oil  Combustion Systems -
      Since no  bottom ash is formed  from oil combustion, it can generally be
 assumed that all of the  trace  elements present in the oil are emitted with
 the fly ash or in the  gas  phase.  There are few  data on particle size
 MCH/007       -                       3-68

-------
association of trace metals emitted from oil combustion systems.
Volatilization/condensation mechanisms may play a role in the behavior of
elements in oil combustion systems.  However,  oil fly ash particles  have
irregular, honeycombed surfaces as opposed to coal fly ash particles which
have smooth, round surfaces.  Therefore,  surface area will not necessarily
have a strong dependence on particle size, and trace metal enrichment on
small particles may not be as pronounced for oil combustion as for coal
combustion (Baig et al.,  1981).

Behavior of Radionuclides During Combustion

     Naturally occurring radionuclides present in coal include uranium-238
(U-238), uranium-235 (U-235), thorium-232 (Th-232), and potassium-40 (K-40)
as well as their daughter products.  Some of these include Th-230, Th-228,
radon-228 (R-228),  R-226, lead-210 (Pb-210) and polonium 210 (Po-210).  For
the purposes of this study, U-238 and Th-232 will be used as indicators of
radionuclide emissions.  These two species have the longest half-lives
         q                             10
(4.5 x 10  years for U-238 and 1.4 x 10   years for Th-232) and are the
parent species of the two predominant decay chains.  They have been selected
as indicators of radionuclides in previous risk assessments  (Environmental
Research and Technology, Inc., 1983; U. S. Environmental Protection Agency,
1984a).
     Radioactive uranium and thorium contained  in the  coal  feed is
partitioned between the bottom ash and fly ash  during  combustion.  Very
little, if any, radionuclides are  emitted to  the  atmosphere in vapor  form
(Roeck et al., 1983)
     Several studies have  found  that U-238  is  enriched in the small (<1  urn
diameter) fly ash particles  (Coles et al.,  1978;  Klein et al.,  1975b;
Roeck et al., 1983).  Uranium-238  would be  termed a Class 2 element using
the terminology developed  previously.   It has been postulated that  a portion
of the uranium in coal is  associated with the silicate (i.e.,  coffinite) and
follows the alumino-silicate minerals which melt and drop out as  slag during
the combustion process.  Another fraction of the U-238 is dispersed in the
coal as uranite and becomes volatile  as uranium oxide (UO,) during
MCH/007       .                        3-69

-------
combustion and continues along with the flue gas and fly ash.   At  normal
stack temperatures the UO. condenses out on the fly ash, preferentially
concentrating on the smaller fly ash particles because of  their larger
surface area to mass ratio (Coles et al., 1978).
     Some studies have found that for Th-232, there is little  preferential
partitioning between the slag and the collected or discharged  fly  ash
(Coles et al., 1978; Klein et al., 1975b).  Other studies  have indicated
small particle enrichment in the fly ash (Roeck et al., 1983).   Thorium-232
would be termed a Class 3 element using the terminology developed  in
previously.

Formation and Transformation of POM and Formaldehyde During Combustion

Formaldehyde-
     Formaldehyde  is  formed and emitted during combustion  of
hydrocarbon-based  fuels such as coal and oil.  Formaldehyde is present in
the vapor phase of the  flue gas.  Since formaldehyde is subject to oxidation
and decomposition  at  the  high  temperatures encountered during  combustion,
large  units with  efficient: combustion  resulting from closely regulated
air-fuel ratios,  uniformly high combustion chamber  temperatures, and
relatively  long  retention times should have  lower formaldehyde emission
rates  than do small,  less efficient combustion units  (Hangebrauck et al.,
 1964;  Rogozen et al.,  1984b).

 Polycyclic Organic Matter-
      The term polycyclic  organic  matter  (POM) defines  a broad  class of
 compounds which generally includes  all organic  structures  having two or more
 fused aromatic rings (i.e.,  rings which  share a common border).  Polycyclic
 organic matter with up to seven fused rings  have  been  identified.
 Theoretically, millions of POM compounds  could  be  formed;  however, the list
 of  species that have been identified and studied  is more  on the order of
 approximately 100 (U. S.  Environmental Protection Agency,  1980b).
      Nine major categories of compounds  have been defined by the U. S.
 Environmental Protection Agency to constitute the class  known  as POM
 (Shih et al., 1980a).  The nine categories  are  as  follows.

 MCH/007       .                       3-70

-------
     1.    Polycyclic  aromatic  hydrocarbons  (PAHs)  -  Che  PAHs  include
          naphthalene,  phenanthrene,  anthracene,  fluoranthene,
          acenaphthalene,  chrysene, benzo(a)anthracene,
          cyclopenta(c,d)pyrene,  the  benzpyrenes,  indeno(l,2,3-c,d)pyrene,
          benzo(g,h,i)perylene,  coronene, and  some of  the alkyl derivatives
          of these compounds.
     2.    Aza arenes  -  aza arenes are aromatic hydrocarbons containing  a
          ring nitrogen.
     3.    Imino arenes  -  these are aromatic hydrocarbons containing  a ring
          nitrogen with a hydrogen.
     4.    Carbonyl arenes - these are aromatic hydrocarbons containing  one
          ring carbonyl group.
     5.    Dicarbonyl  arenes -  also known as quinones,  contain two ring
          carbonyl groups.
     6.    Hydroxy carbonyl arenes -  these are  ring carbonyl arenes
          containing hydroxy groups and possibly alkoxy or acyloxy groups.
     7.    Oxa arenes  and thia arenes  - oxa  arenes contain  a ring oxygen
          atom, while thia arenes contain a ring sulfur atom.
     8.    Polyhalo compounds - these  include  the polychlorinated
          dibenzo-p-dioxin (PCDDs),  polychlorinated dibenzofurans (PCDFs),
          and polychlorinated biphenyls (PCBs),  and also brominated analogs
          of these compounds such as polybrominated biphenyls (PBBs).
     9.    Pesticides -  including aldrin, chlordane,  and DDT.

These categories were developed to better define  and standardize  the types
of compounds considered to be POM.
     The two POM chemical groups most commonly found in emission  source
exhaust and ambient air are PAHs, which contain carbon  and hydrogen only,
and the PAH-nitrogen analogs.  Information available in the literature on
POM compounds  generally pertains  to  these  PAH groups.   Because of the
dominance of PAH  information  (as  opposed to other POM categories) in the
literature, many  reference sources have  inaccurately used  the  terms POM  and
PAH interchangeably.  The majority of  information in  this  report on POM
physical/chemical properties, formation  mechanisms, and emissions pertains
to PAH compounds.

MCH/007       -                        3-71

-------
     Polycycltc organic compounds are formed in stationary combustion
sources as products of incomplete combustion.  The rates of POM formation
and emission are dependent on both fuel characteristics and combustion
process characteristics.  Emissions of POM can originate from POM compounds
contained in fuels that are released during combustion or from high
temperature transformations of organic compounds in the combustion zone
(Shih et al., 1980a; National Academy of Sciences, 1972; National Research
Council, 1983).
     Two important fuel characteristics affecting POM formation in
combustion sources are (1) the carbon to hydrogen ratio and molecular
structure of the fuel and (2) the chlorine and bromine content of the fuel
(Shih et al., 1980a).   In general, the higher the carbon to hydrogen ratio,
the greater the probability of POM compound formation.  Holding other
combustion variables constant, the tendency for hydrocarbons present in  a
fuel to form POM compounds is as follows.

               aromatics > cycloolefins > olefins > paraffins

Based on both carbon to hydrogen ratio and molecular structure
considerations, the tendency  for the combustion of various fuels to form POM
compounds is as follows (Shih et al., 1980a).

       coal>  lignite > wood >  waste oil > residual oil > distillate oil

      In the  formation of chlorinated and brominated POM compounds during
 stationary source  fuel  combustion, the chlorine and bromine content of the
 fuel  plays a major role.  Based on the chlorine content of fuels, the
 tendency to  form  chlorinated  POM compounds during combustion is:

       bituminous  coal > wood  > lignite > residual oil > distillate oil

 Similarly, based  on the bromine content of fuels, the tendency to form
 brominated POM compounds  during combustion is:
       bituminous  coal  > lignite >  residual oil > distillate oil > wood
 MCH/007       .                       3-72

-------
     The primary combustion process  characteristics  affecting  POM compound
formation and emissions are (Shih et al.,  1980a;  Hangebrauck et  al.,  1964;
Barrett et al.,  1983):

          combustion zone temperature,
          residence time in the combustion zones,
          turbulence or mixing efficiency  between air  and fuel,
          air/fuel ratio, and
          fuel feed size.

With adequate residence time and efficient mixing,  temperatures  in  the
800-1000°C (1472-1832°F) range will  cause  complete destruction of  POM
compounds such as polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated
dibenzofurans (PCDFs),  and polychlorinated biphenyls (PCBs).   Concentrations
of polyaromatic hydrocarbons (PAHs)  also decrease rapidly with increasing
temperature (Shih et al., 1980a).
     The most important reason for incomplete combustion of fuel,  thereby
resulting in POM formation, is insufficient mixing between air,  fuel, and
combustion products.  Mixing is a function of the combustion unit's
operating practices and  fuel firing configuration.  Hand- and stoker-fired
solid fuel combustion sources generally exhibit very poor air and fuel
mixing relative to other types of combustion sources.   Liquid fuel units and
pulverized solid fuel units provide good  air and fuel mixing  (Shih et al.,
1980a; Hangebrauck et al.,  1964; Barrett  et al., 1983; Kelley,  1983).
     The air/fuel ratio  present  in  combustion environments is important  in
POM formation because certain  quantities  of air  (i.e., oxygen)  are needed to
stoichiometrically carry out complete combustion.   Air supply is
particularly important  in  systems with  poor air  and fuel mixing.   Combustion
environments with  a poor air supply will  generally  have  lower combustion
temperatures and will not  be capable of completely  oxidizing  all  fuel
present.  Systems  experiencing frequent start-up and  shut-down  will  also
have poor air/fuel  ratios.  Unburned hydrocarbons,  many  as  POM  compounds,
can exist in such  systems  and  eventually  be  emitted through the source
stack.   Generally,  stoker  and  hand-fired  solid  fuel combustion sources  have
 MCH/007       .                       3-73

-------
problems wich insufficient air supply and tend to generate relatively  large
quantities of POM as a result (Shih et al., 1980a; Kelley, 1983;
Barrett et al.,  1983) .
     In solid and liquid fuel combustion sources, fuel feed size  can
influence combustion rate and efficiency, therefore, POM compound formation
is affected.  For liquid fuel oils, a poor initial fuel droplet size
distribution is conducive to poor combustion conditions and an enhanced
probability of POM formation.  In most cases, fuel droplet size distribution
is primarily influenced by fuel viscosity.  As fuel viscosity increases,  the
efficiency of atomization decreases and the droplet size distribution  shifts
to the direction of larger diameters.  Therefore, distillate oils are  more
readily atomized then residual oils and result in finer droplet size
distribution.  This behavior combined with distillate oil's lower carbon  co
hydrogen  ratio means that residual oil sources inherently have a  higher
probability of POM formation and emission  then distillate oil sources
(Shih et  al.. 1980a; Hangebrauck et al., 1964; Kelley, 1983).
     For  solid fuels, fuel size affects POM formation by significantly
impacting combustion rate.  Solid  fuel combustion involves a series  of
repeated  steps, each with the potential to form  POM compounds.  First,  the
volatile  components near the surface of a  fuel particle are burned followed
by burning  of the  residual solid structure.  As  fresh, unreacted  solid
material  is  exposed, the process is repeated.  Thus, the larger the  fuel
particle, the greater the number of times  this sequence is repeated  and the
 longer  the  residence time required to complete the combustion process.  With
 succeeding  repetitions,  the  greater  the probability of incomplete combustion
 and POM formation.   Again, stoker  and hand-fired solid fuel combustion units
 represent the greatest  potential for POM  emissions due to fuel size
 considerations  (Shih et al.,  1980a).
      Polycyclic  organic matter  can be emitted  from fuel combustion sources
 in both gaseous  and particulate phases.   The compounds are initially formed
 as gases, but as the flue  gas  stream cools,  a  portion of  the POM
 constituents adsorb to  solid fly ash particles present in the stream.   The
 rate of adsorption is  dependent on temperature,  and on fly ash and POM
 compounds characteristics.   At  temperatures  above 150°C  (302°F),  most  POM
 MCH/007      -                        3-74

-------
compounds are expected to exist primarily in gaseous form.   In several types
of fuel combustion systems,  it has been shown that POM compounds are
preferentially adsorbed to smaller (submicron)  fly ash particles because of
their larger surface area to mass ratios.   These behavioral characteristics
of POM emissions are important in designing and assessing POM emission
control systems (Shih et al., 1980a;  Kelley,  1983;  Griest and Guerin,  1979;
Sonnichsen, 1983).

EFFECTS OF COMBUSTION SOURCE DESIGN AND CONTROL TECHNOLOGY ON EMISSIONS

Characteristics of the Boiler Population

Boiler Design-
     Boiler design influences the rate of trace metal and POM emissions.
Types of coal-fired boilers used in the utility, industrial, and
commercial/institutional sectors include pulverized coal-fired, cyclone, and
stoker units.  Pulverized units are characterized by ash removal method as
dry bottom or wet bottom.  There is little variatiAi in the design of
oil-fired units, and almost all are tangentially fired.  Table 3-56 shows
the prevalence of each boiler type (in t^.rms of 1978 fuel use) in the
utility, industrial, and commercial/institutional sectors.
     The utility sector is dominated by pulverized  dry bottom  coal-fired
units.  In the future, the percentage of these units is expected to
increase.  Coal-fired pulverized wet bottom  and cyclone boilers are no
longer sold due  to  their  inability to meet NO   standards.   Stoker boilers,
currently accounting for  less  than one percent  of  the  total,  are obsolete
due to their  inefficiency and  are being  retired.
     In the  industrial sec.tor,  more  natural 'gas  is  used  relative to coal  and
oil.  Pulverized coal-fired  units  are  the  most  common type of coal-fired
unit; however,  stoker units  (mainly  spreader stokers)  also account  for a
large percentage of total coal use.
     The commercial/institutional sector consumes a greater proportion of
oil and natural  gas relative to coal consumption than the other two sectors.
Small underfeed stokers  are  the predominant type of coal-fired boiler.  Some
 MCH/007       .                        3-75

-------
      TABLE 3-56.   POPULATION CHARACTERISTICS  OF UTILITY,  INDUSTRIAL AND
         COMMERCIAL BOILERS  IN TERMS  OF BOILER DESIGN AND  FUELS,  1978


Boiler Type
Coal-Fired Boilers
Pulverized Dry Bottom
Pulverized Wet Bottom
Cyclone
Stoker
Oil -Fired Boilers
Gas-Fired Boilers
Otherd
Total fuel consumption
by external. combustion
sources (10 Btu)

Percent o:
Utility*

49.6
7.2
7.4
0.7
21.6
13.6
-
16,761

f Total Fuel Use (Heat Input)
for Each Sector
Commercial/
Industrial Institutional

7.1 0.4
1.7 0.02
0.4
7.1 2.4
19.6 51.6
57.4 43.6
0.01 0.04
8,236 4,777
  Source:  Shin et al., 1980b
  'Source:  Suprenant et al., 1980a
  'Source:  Suprenant et al., 1980b
  Other  includes wood and refuse.
MCH/007
3-76

-------
of the larger institutional sources in this  sector  are  pulverized  coal-fired
boilers and spreader stokers.

Control Status-
     All coal-fired utility boilers are equipped with some  form of
particulate emissions control  device.   High  efficiency  electrostatic
precipitators (ESPs) are the most common.  Data on  the  distribution of
control techniques for coal-fired utility  boiler particulate emissions  are
shown in Table 3-57 (Radian Corporation, 1983).   A  study of coal-fired
utility boilers larger than 100 MW and placed in service since 1950 showed
that in 1980 about 92 percent  of the generating capacity was controlled with
ESPs, 2 percent with fabric filters, 1 percent with scrubbers, and the
control status of 5 percent was unknown (Barrett et al., 1983). New  units
subject to NSPS must control particulate emissions  by about 99 percent, so
the control status of coal-fired utility boilers is expected to improve over
time.  More current (1984) data on the control status of utility boilers is
contained in the POWER data base maintained by the  Utility Data Institute
(UDI) in Washington, D.C.
     The Utility Data Institute is a private data base management  group
under contract to the Edison Electric Institute (EEI) to manage their
"POWER" data base.  The data base contains power plants utilizing coal, oil,
and other fuels organized alphabetically by State.   Information included for
each plant includes about 300 parameters including name, location, latitude
and longitude, capacity, fuel type, fuel use, criteria pollutant  emissions,
control status, and stack parameters.  Most of the data are obtained from
DOE/EIA Form 767.  The utilities send UDI a copy of these  forms when they
return them to DOE.  Other data comes from direct contacts  and surveys  of
utilities.
     In 1984, about 17 percent of  the utility coal generating capacity was
equipped with flue gas desulfurization  (FGD) systems.   The majority  of these
were lime or limestone scrubber systems.  It is  predicted  that by 1992,
about 31 percent  of coal generating capacity will be equipped with FGD
systems (Melia et al., 1984).
     Oil-fired utility boilers are often  uncontrolled;  however some  are
equipped with mechanical precipitators, cyclones,  or ESPs  (Shin et al.,

MCH/007                               3-77

-------
   TABLE  3-57.   BREAKDOWN OF CONTROL TECHNIQUES FOR REDUCING PARTICULATE
                        EMISSIONS  FROM COAL-FIRED UTILITY BOILERS

Control Device Type
ESPa
Wet Scrubber
Baghouse
Mechanical Collector
Number of Boilers
With This Control
979
32
47
137
Percent of Total
Generating Capacity
Represented
92.6
4.2
2.1
1.1
aESP category also includes units listed as having a combination of control
 techniques, units using flue gas conditioning to improve ESP performance,
 and a small number of units for which no control method was listed.

 Does not include units with scrubbers for flue gas desulfurization (FGD)
 unless the scrubber is the only particulate control device.

°Includes units which have only mechanical control techniques (cyclones,
 multicyclones).

Source:  Radian Corporation, 1983.
 MCH/007                             3-78

-------
1980b).   The POWER data base contains  current  information on the  control
status of oil-fired utility boilers.
     Coal-fired industrial boilers are less  well  controlled than  utility
boilers.  Based on a 1976 survey of over 2,500 units,  about 14  percent  were
controlled with ESPs,  47 percent with  cyclones, 4 percent with  scrubbers,
1 percent with fabric  filters,  and 33  percent  were uncontrolled
(Suprenant et al., 1980a).   The applicability  of  these percentages  to  the
entire industrial boiler population is unknown.   In general,  larger units
are more likely to be  controlled than  smaller  units,  and pulverized coal  and
cyclone boilers are more likely to be  controlled  than stokers (Suprenant
et al.,  1980a).  The NSPS for industrial boilers  (>100 million  Btu) and
small boilers (<100 million Btu) will  result in improved emissions  control
in the future.  Oil-fired industrial boilers are  typically uncontrolled.
     Commercial and residential boilers and furnaces are typically
uncontrolled.  However, cyclones are in place  at  some of the larger
commercial/institutional coal-fired boilers (Suprenant et al.,  1980b).

Trace Metal and Radionuclide Emissions

     Boiler design affects the amount of ash entrained in  the  flue gas.
Since all of the  trace metals and radionuclides reviewed,  except mercury,
are emitted predominantly in particulate form, the amount  of fly ash emitted
will influence the amount of trace metals emitted.  Table  3-58 presents the
fraction of coal  ash emitted as fly ash for different combinations  of  boiler
firing configurations  and coal  types  (Baig et al., 1981).  The fractions  for
bituminous coal-fired  boilers are based on several tests.  The values  for
lignite and anthracite are much less  certain.  Further  testing is  necessary
to determine  if the three types of coals generate  different  ratios of  fly
ash to bottom ash when burned  in  similar boilers.
     Boiler configuration may  also affect the volatilization/condensation
behavior of trace elements,  and hence their emission rates.  This  is
especially true for Class 3  elements  which  show  enrichment in  the  fly  ash in
some studies  and  not  in  others  (Baig  et  al.,  1981).   Elements  may  be  more
likely  to be vaporized in large pulverized  coal-fired boilers  where
MCH/007       .                        3-79

-------
             TABLE 3-58.  COAL ASH DISTRIBUTION BY BOILER TYPE3

Furnace Type
Pulverized dry bottom
Pulverized wet bottom
Cyclone
Stoker
Percent
Bituminous
Coal6
80/20
65/35
13.5/86.5
60/40
Fly Ash/Percent
Lignite
Coal
35/65
—
30/70
35/65
Bottom Ash
Anthracite
Goal0
85/15
—
—
5/95
aSource:  Baig et al., 1981
 Based on several studies of coal ash from large and intermediate size coal-
 fired boilers.
CBased on an analysis of uncontrolled particulate emissions.
MCH/007      -                      3-80

-------
combustion is more efficient due to higher temperatures,  longer iresiclef.ee
times, and efficient mixing of air and fuel;  and they may be volatilized to
a lesser degree in smaller, less efficient,  lower temperature combustion
systems.  The temperature of the stack gas and fly ash characteristics
influence the condensation behavior of volatilized trace  metals and their
adsorption onto fly ash particles.
     The efficiency of control devices in removing trace  elements  depends  on
whether the elements are in vapor or particulate form and on the size of the
fly ash particles with which the elements are associated.  Typical
particulate controls on industrial and utility boilers include multicyclones
and ESPs.  Scrubbers are applied to some utilities for S02 (and particulate)
control.  For elements such as manganese, which tend to show an even
distribution on all sizes of particulates, collection efficiency of
particulate control devices should be similar to overall particulate control
efficiency.  However other elements such as arsenic, cadmium, copper and
U-238 are enriched in the smaller particulate fractions  (<1 urn).  Mechanical
collection devices such as cyclones and multicyclones generally show
decreasing collection efficiency as particle size decreases; therefore, the
collection efficiency of trace elements concentrated on  small particles will
be less than overall particulate collection efficiency.  Although not as
severe as for cyclones, this condition also exists for scrubbers and ESPs.
ESPs often show a minimum collection efficiency  in the 0,1  to  1 urn diameter
size range (Ondov et al.,  1979a).
     Furthermore, ESPs and cyclones will not reduce emissions  of elements,
such as mercury, emitted in the vapor phase.  A  portion  of  the  other trace
metals, especially the Class 2 elements, may also remain in vapor  form  in
the flue gas, and may thereby escape collection.

Polvcyclic Organic Matter  Emissions

     Polycyclic organic matter emission  rates  are also influenced by boiler
design.  As noted previously, POM formation  depends  on temperature,
residence time, efficiency of air and  fuel mixing,  air/fuel ratio,  and fuel
feed size.  Based on these criteria, pulverized dry bottom and wet bottom
MCH/007       .                        3-81

-------
coal-fired units would have the lowest POM emission factors of any
coal-fired units.  These units are generally large, temperature of  the
combustion zone is high [around 1,650 C (3,000 F)], residence time  in  che
combustion zone is relatively long (0.5 sec), air/fuel ratios are constant
and adequate for efficient combustion, and the coal feed is pulverized  into
small particles.  Cyclone-fired boilers would have the next lowest  POM
emission rates.  Stokers would have higher emission rates, with overfeed and
underfeed stokers having slightly higher emission  rates than spreader
stokers.  Stoker units are usually smaller, temperatures in the combustion
zone are lower due to the 30 to 60 percent excess  air present, mixing
between air and  fuel is less efficient, the on-off cycle results  in
fluctuations in  the air/fuel ratio, and fuel feed  size is larger.   These
factors lead to  increased POM formation.  Hand stoked units would have  the
highest emission factors of all coal-fired units  (Shih et al., 1980a;
Barrett et al.,  1983).
     Oil-fired units have less of a tendency to form POM than coal-fired
units due to fuel characteristics.  Based on fuel  characteristics,  residual
oil fired units  are more likely to form POM than  distillate oil fired  units.
Based on boiler  design characteristics, large oil-fired utility boilers
would have  the  lowest POM emission rates, followed by industrial  boilers.
Based on design, home heating units would have higher POM emission  rates;
however,  these  are usually  fired with distillate  oil which would  tend  to
 reduce  emissions (Shih et al., 1980a).
      Polycyclic organic matter is emitted in both vapor and parciculate
 phases, with che vapor phase  generally predominating, and the particulate
 phase showing small  particle  enrichment.  Particulate POM, particularly fine
 particles,  would be  controlled most effectively by baghouses or ESPs.   No
 control of gaseous  POMs would be  achieved by baghouse and ESP systems.  Wet
 scrubbers could potentially be effective for controlling particulate and
 gaseous POM.  Scrubbers would condense  the  POM compounds existing as vapors
 and collect them as  the  gas stream  is saturated  in the scrubber.
 Multicyclones would be the  poorest  control  system for POM emissions because
 they are ineffective on fine  particles  and  would  have no control  effect on
 gaseous POM (Kelley, 1983).
  MCH/007       .                       3-82

-------
     Wet FGD/ESP systems, while providing for the control of POM condensed
on particulate matter at the entrance to the ESP, have been shown to  be poor
at controlling vapor phase POM.  Tests examining benzo(a)pyrene showed that
condensation of the vapor phase POM compound would occur in the scrubber,
but significant collection of POM particles remaining in the gas flow
through the scrubber was not achieved (Kelley,  1983).
 MCH/007      .                        3-83

-------
                                  SECTION 4
                    TOXIC AIR POLLUTANT EMISSION FACTORS
                         FOR COAL AND OIL COMBUSTION

     This section contains emission factors for selected  toxic  air
pollutants from coal and oil combustion.   Factors are  presented for  arsenic,
beryllium, cadmium, chromium, copper, manganese,  mercury,  nickel, lead,
formaldehyde, POM, and selected radionuclides (uranium-238,  thorium-232).

EMISSION FACTORS FOR OIL-FIRED COMBUSTION SOURCES

     The literature was reviewed for measured and calculated oil emission
factors.  A summary of emission factors for the nine trace metals,  POM,  and
formaldehyde emitted from the combustion of residual and distillate oil are
presented and discussed below.  No data were identified for radionuclide
emissions from oil combustion.
     The summarized emission factors should not be construed to represent a
fully characterized or representative emission rate for the given combustion
source situation.  Extensive data quality assurance procedures, necessary
to reasonably characterize a data set as representative of a particular
source, were not performed in  this study because of time  and budgetary
constraints.  Instead, the summarized factors  are simply  straightforward
calculations of emission  factor averages and ranges based on data presented
in the literature.  The summarized factors are not  to be  considered as
suggested emission factor values  for use  in  other activities such as regula-
tory development  or specification of acceptable  ambient  concentrations.

Summary of  Emission Factors

     A  summary of toxic pollutant emission factors  for  residual and
distillate  oil combustion are  presented in Table 4-1.   These  are uncon-
trolled emission  factors  that  could  be used in efforts  such as emission
 MCH/007                               4-1

-------
              TABLE 4-1.  SUMMARY OF TOXIC POLLUTANT EMISSION
                              FACTORS FOR OIL COMBUSTION*
                                                     r  (lb/1012 Btu)
Pollutant
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Manganese
Nickel ,
POM
Formaldehyde
Residual Oil
19
4.2
15.7
21
280
28C
3.2
26
1260
8.4b
405*
Distillate Oil
4.2
2.5
10.5
48
280
8.9d
3.0
14
170
22.5
405e
 All emission factors are uncontrolled,  and are  applicable  to oil-fired
 boilers and furnaces in all combustion sectors  unless  otherwise noted.
 This value was calculated using all available residual oil data given
 in Table 4-35.  If the upper end of the range of available data is
 excluded when calculating an average value (which could be used in this
 table), the average factor for POM from residual oil combustion becomes
 4.1 lb/10   BTU.
Applicable to utility boilers only.

 Applicable to industrial, commercial, and residential  boilers.

 The formaldehyde factors are based on very limited and relatively old
 data.  Consult Table 4-37 and accompanying discussion  for  more  detailed
 information.
  MCH/007
                                       4-2

-------
inventory development.  They are applicable  to  all  types  of oil-fired
boilers in all four combustion sectors (utility,  industrial,
commercial/institutional,  and residential).

Derivation of Summary Trace Metal Emission Factors-
     The summarized emission factors for eight  of the nine trace  metals
studied were calculated from the typical level  of these metals in residual
and distillate oil assuming the entire mass  of  the  trace  metals entering the
boiler in the oil feed is emitted in the flue gas.   Typical values for the
trace element content of residual and distillate  oils presented in Section  3
were used in the calculations.  These were average values based on a review
of several previous studies of the trace element  composition of oil.
Typical trace metal concentrations in the oil feed (expressed in ppm)  were
                                                            12
converted to emission factors (Ib trace metal emitted per 10   Btu of oil
burned) assuming heating values of 150,000 Btu/gal for residual oil and
141,000 Btu/gal for distillate oil, and densities of 944 g/1 (7.88 Ib/gal)
for residual oil and 845 g/1 (7.05 Ib/gal) for distillate oil.  The heating
values are documented in Appendix B.
     Since oil combustion generates no bottom ash, the assumption that
100 percent of the trace metals entering the boiler in the oil feed are
emitted in the flue gas is reasonable.  The calculated uncontrolled emission
factors based on this assumption would be independent of boiler design and
combustion sector.
     Limited emission factor data for lead emissions from oil  combustion are
presented here.  The consideration of lead as a  trace pollutant  from  coal
and oil combustion was added  to this project by  EPA  late  in  the  data
analyses process.  For this reason, the  treatment  of lead,  including  the
availability of emission  factor data,  is very  abbreviated compared to the
other trace pollutants in the  document.  Only  a  limited  number of the
references listed in  the  report bibliography in  Section  6 were evaluated for
lead data.
     The general agreement between measured and  calculated emission factors
from several references lends  some  confidence  to the summarized  values.
However, they  should  be considered in light of the high  variability of trace
elements in oil.  Furthermore,  the  data base on  distillate oil was much less

MCH/007       -                        4-3

-------
complete Chan the data base on residual oil and coal.  For some metals,
there were only two or three available studies reporting their occurrence in
distillate oil.  The representativeness of the distillate oil emission
factors is, therefore, somewhat uncertain.
     Another data gap is the effects of particulate control  technologies  on
trace metal emissions from oil-fired boilers.  Many trace metals  are
enriched in the small particle fractions of the fly ash from coal combustion
sources.  However, oil fly ash has different characteristics, and whether
the volatilization/condensation theories predicting small particle
enrichment are applicable to oil combustion sources is uncertain.   There  is
a lack of literature on the form of trace emissions from oil combustion
(vapor or particulate) and on the association of trace elements with  various
size fractions of the oil fly ash.  Without this information, the efficiency
of particulate control devices at removing trace metal emissions  cannot be
calculated.  Almost all of the calculated and measured emission factors
reported in previous studies are uncontrolled.

Derivation of  POM and Formaldehyde Emission Factors-
     A qualitative discussion of theories of POM and  formaldehyde formation
and  behavior during combustion is presented in Section 3.  No methods for
calculating POM  and formaldehyde emission factors were found in the
literature.  The emission factors presented in Table  4-1 are average  values
derived  from test data contained in the  literature.
     More  test data are available for  POM emission  factors from residual  oil
 than from  distillate  oil.  Reported POM  emission factors for both types of
 oil vary over  two orders of magnitude.   The data show no clear pattern as to
 whether boiler type,  boiler size, combustion sector,  or oil  grade influence
 POM emissions.  Part  of  the observed variation may be due to variations in
 sampling and analytical methodology between studies.
      Only four measured  formaldehyde emission factors were available  in the
 literature.   While  these  are  in  fairly close agreement, the  scarcity  of data
 make the representativeness of the summarized emission factor highly  uncer-
 tain.   There are not enough data to derive separate formaldehyde  emission
 factors for residual versus distillate oil.
 MCH/007      -                       4-4

-------
     The efface of particulate control technologies on POM and formaldehyde
emissions is another area lacking data.   There are few measurements of POM in
controlled emission streams, and little data on the distribution of POM and
formaldehyde in the vapor versus particulate phases.  Theoretically,  a large
portion of POM and formaldehyde should be present in vapor form and would
therefore escape collection; however,  very limited test data for residual
oil-fired sources appears to indicate  lower POM emission factors for
controlled versus uncontrolled boilers.

        Emission Factors
     Based on a typical residual oil arsenic content of 0.36 ppm, the
summarized uncontrolled arsenic emission factor for residual oil combustion
is 19 lb/1012 Btu.  This is in the middle range of values calculated in five
                                                            12
previous studies, which range from less than 0.5 to 42 lb/10   Btu (see
Table 4-2) .  Eight measured arsenic emission factors from the literature are
shown in Table 4-3.  Uncontrolled emission factors reported by two authors
                          12
range from 4.2 to 37 lb/10   Btu, and are in good agreement with the
                             12
recommended value of 19 lb/10   Btu.  Since levels in fuels were often below
the detection limit, it is not possible to calculate mass balance closure for
the test runs.  Leavitt et al. (1980) reports higher emission factors,
despite the presence of control devices.  The reason for this is unknown.
     The summarized distillate oil arsenic emission factor  is
4.2 lb/1012 Btu based  on a typical level of 0.085 ppm in distillate  oil.
This is in good agreement with previously calculated factors of 3.0  and
         12
8.1 lb/10   Btu from two studies  summarized in  Table 4-4.   Only four
measured values are reported  in  the  literature, ranging from 1.5 to
3.5 lb/1012 Btu  (see Table 4-5).
 Bervlli'"B Eggxpn Factors
      The summarized uncontrolled beryllium emission factor for residual oil
             12
 is  4.2  lb/10   Btu.  This is in general agreement with previously calculated
                                                              12
 values  shown in Table 4-6 which range from 0.05 to 5.57 lb/10   Btu.  There
 is  some uncertainty regarding the calculated values reported in the

 MCH/007      -                       4-5