Prepared for:

             Walter H. Stevenson
     Standards  Development  Branch (MD-13)
             Michael G. Johnston
      Industrial 'Studies Branch  (MD-13)
 Office  of  Air  Quality Planning and Standards
     U.S. Environmental Protection Agency
Research Triangle Park> North Carolina  27711
                 Prepared by:

     Kristina  L.  Nebel  and David M.  White
              Radian Corporation
            Post Office Box 13000
Research T'riangle Park, North Carolina  27709
                September 1991

This report provides information on mercury emission rates

and control technologies applicable to municipal waste combustors
(MWC's). section 2.0 presents emissions data for MWC's located
in North America. Discussions on apparent relationships between
mercury removal and various parameters such as particulate matter
(PM) control device inlet temperature and levels of carbon in the
fly ash are included.
Section 3.0 discusses mercury control technologies currently
being used in Europe and Canada, and includes a review of MWC'~,
that use sodium sulfide (Na2S) injection, activated carbon

. .
injection, and wet scrubbing to limit mercury emissions.
.1, .
Section 4.0 provides a listing of the references ~~ed.
kl j/145

Table 2-1 lists mercury emissions data tor MWC's in
North America that are not equipped with specific mercury control
technologies. The data cover 47 MWC units at 32 different
plants. For those facilities reporting both inlet and outlet'
mercury concentrations (in microgramsjdry standard cubic meter
[~gjdscm] at 7 percent 02), mercury removal efficiencies are
calculated. Table 2-1 also lists the type of combustor, type of
air pollution control device (APCD), PM control device inlet
temperature, and inlet dioxinjfuran (CDDjCDF) concentrations.
Uncontrolled (i.e., APCD inlet) mercury concentrations are
. .f
reviewed in section 2.2. Section 2.3 reviews the ~ontrolled
. I
(i.e., APCD outlet) concentrations for different APCD types and
discusses the apparent relationship between controlled mercury
levels versus APCD type and ~M control device temperature. The
temperature entering the PM control devic~ is important because
mercury exists in a vaporous form at temperatures greater than
300 of and does not effectively condense9nto PM. The
relationship with inlet CDDfCDF concentrations is also examined
in this section. Inlet CDDjCDF levels serve as a surrogate for
estimating residual carbon in the ,fly ash, which may enhance
mercury removal due to the adsorption of mercury onto carbon. A
summary of the results and conclusions drawn from these
relationships is pres~nted in S~ction 2.4.
As listed in Table 2-1, mercury levels prior to APCD's range
from roughly 200 to r400 ~gjdscm. Most of the concentrations
range from 400 to 1000 ~gjdscm and average roughly 650 ~gjdscm.
Based on the data, there is no clear distinction in inlet mercury'
levels at mass burn plants and refuse-derived fuel (RDF) plants., /'
kl j/145

       (F) (ng/dsCII) (ug/dsCIII) (ug/dsCII) (X)  
Detrol t (7/89)  9 Runs RDF ESP 600    653  1a 
Detroit (11/89)*  3 Runs RDF ESP 600    96  1a 
Detroit (11/89)**  14 Runs RDF ESP 600    193  1a 
Detroit {12/89-1/90)*** 13 Runs RDF ESP 600    172  1a 
Detroit (3/90)  9 Runs RDF ESP 600    194  Ib 
Hi 1 lsborough   . 3 Runs HB/W ESP     823  1e 
Oneida County  1-3 H/S ESP HM    2060  2 
            \1 '.  
Pigeon Point  1-3 H/E ESP 410    363  2 
Pinellas County  1-3 HB/W ESP 543 54 -- ~i.,. 847  2 
Pope/Douglas  1-3 H/E ESP 482    133  2 
Quebec City  2,10,11 HB/W ESP 406    918  2 
Quebec City  5,6,12 M8/W ESP 417    685  2 
Tulsa    1-3 MB/W ESP 375    418  2 
Tulsa    3 Runs MB/W ESP     1000  1e 
Tulsa    3 Runs MB/W ESP     746  1e 
Tulsa    3 Runs MB/W ESP  --   600  1e 
Tulsa    3 Runs MB/W ESP,     711  1e 
Tulsa    3 Runs MB/W ESP     97  1e 
Tulsa    2 Runs MB/W ESP     466  1e 
Dayton    1-3 MB/R ESP 560 252 962  1016 -5.6 2 
Dayton    4-6 M8/R ESP 4,01 328 1055  1150 -9.0 2 
Dayton    10-12 Ma/R FSIIESP 394 38 973  757 22.2 2 
Dayton    13-15 Me/R FSIIESP 298 14 907  709 21.8 2 
Dayton    16-18 18/1 DSI/ESP 306 5 716  491 .31.4 2 
~ .              
'. A 1 exandri a, Unit 1 (12/88) 1-3 MB/W FSIIESP     517  Ie 
Burnaby, Unit 3 (11/88) 1-3 MB/W DSIIFF 313 78 527  485 8.0 3 
Burnaby, Unit 1 (4/89) 1-3 MB/W DSIIFF 307  1360 b  3 
Burnaby, Unit 1 (9/89) 1-3 Me/W DSIIFF 324    470  3 "
     . .
Burnaby, Unit 2 (9/89) 1-3 Ma/W DSIIFF 325    368  3 
Burnaby, Unit 3 (9/89) 1-3 MB/W DSIIFF 319    1088  ..3 
Outchess County, Unit 1 1-3 MS/w' DS II FF 430    1080 . 2 
Dutchess County, UnI t 2 1-3 MB/W DSIIFF 365    85  2 
Quebec City - Pilot 5-6:" MB/W DSIIFF 400 1597 451  614 -36.1 2 
Quebec City - Pilot 1.2,11 MB/W DSIIFF . '285 2277 320  16 95.0 Z 

        (F) (ng/dsCIII) (ug/dsCII) (ug/dsCIII) (X) 
Charleston, Units A & 8 1-3  M8/W SO/ESP  -- e   .723  4
Haverlll, Unit A (6/89) 1-3  M8/W SO/ESP 285    247  1e
Haverill, Unit B (6/89) 1-3  MB/W SO/ESP 285    208  1e
Haverlll, Unit 8 (3/90) 1-3  M8/W SD/ESP     567  1e
Honolulu, Unit 1  1-3  RoF SD/ESP 300 -- d   5  5
Honolulu, Unit 2  1-3  RDF SO/ESP 293. -- d   7  5
141 11 bury,  Unit 1  1-6  M8/W SD/ESP 249    .565  2
Mi 11 bury, . Unit 2  1-3  MB/W SO/ESP 240 170 e   95-4  2
Portland, Unit 1 (12/89) 4-6  M8/W SO/ESP 308 -- f .J. 550  6
Portland, Unit 2 (12/89) 1-3  M8/W SO/ESP 285 --. f -_ : 382  6
SEMASS, Unit 1  1-3  RoF SO/ESP 287 --g   59  7
SEMASS, Unit 2  2-4  RDF SO/ESP 293 --g   105  7
West Palm Beach, Unit 1 3 Runs  RoF SO/ESP. 275    56  8
West P.lm Beach, Unl t 2 3 Runs  RoF SD/ESP 278    23  8
Babylon, Unft 2  1-3  MB/W So/FF 331    451  9
Babylon   3 Runs  MB/W So/FF  --   323  1e
Blddeford   1-3  RDF SD/Ft: 278 903 389  ND >99 2
Bristol   3 Runs  148/W. So/FF     99  Ie
Bristol   3 Runs  MB/W So/FF     105.  1e
Bristol   3 Runs  148/W So/FF     64  1e
Bristol   3 Runs  MB/W SD/FF     399  1e
Carmeree (1987)  11.13.14  MBtW SD/FF 270 h 28 450  570 -26.7 2
Ccmneree (1988)  3.5.9  MB/W SD/FF 290 h 446. 453  39 .91.4 2
Carmerce (1988) MB/W SD/FF 290 h 783 . 261  68 74.0 2
Fai rfax   3 Run.  I48LW SD/FF     406  1c.
Fairfax   3 Run.  MB/W SD/FF     466  1e
Fal rfax   3 Run.  M8/W SD/FF     331  1c
Fairfax   3 Runs  148/W SD/FF     514  lc.
~stead. Unl t 1 (9/89) 1-3  MB/W SD/FF 3-10 h    9  1"0
~st8ad. Unit 2 (9/89) 1-3  148/W SD/FF 310 h    25 -- .. 10
IieqIstead. Unl t 3 (10/89) 1-3  M8/W SD/FF 310 h    25  10
Huntsville   3 Runs  M8/W SD/FF     1275  1e
Huntsville   3 Runs  148/W SD/FF     463  lc

.a[[[a. I
        (F) (ng/dsaa) (ug/dsCII) (ug/dsCIII) (X)  
Indl anapolls. Unit 1 1-3 MB/W SD/FF 307    283  11 
Indianapolis   3 Runs MB/W SD/FF     200  Ie 
Indianapolis   3 Runs' MB/W SD/FF     277  Ie 
Long Beach     1-3 MB/W SD/FF 298 305   180  2 
Marton County   4-6 MB/W SD/FF 272 43 e   239  2 
Mid-Connecticut (7/88) 1-3 RDF SD/FF ;76 1019 1008    2 
Ifl d-Connect I c-ut (7/88) 1-3 (Hg)1 RDF SD/FF 284  884  50 94.3, '. 2 
Mid-Connecticut ( 2/89) 12-14 RDF SD/FF  436 668  9 98.7. 2 
Quebec City - Pilot  7-8 MB/W SD/FF 282 176<' 187 .r 10 94.7 2 
Quebec City - Pilot  9-10 MB/W SD/FF 284 2157  ,'/    
 360  19 94.7 2 
Stanislaus County, Unit 1 14,16.19 MB/W SD/FF 295    499  2 
Stanislaus County, Unit 2 38.40.42 MB/W SD/FF 290    . 462  2 
Stanislaus County  3 Runs MB/W SD/ FF     508  lc 
Stanislaus County  3 Runs MB/W SD/FF'     481  lc 
Stanislaus County  3 Runs MB/W SD/FF     427  Ie 
Kent     3 Runs MB/W SD/FF     166  lc 
Kent     3 Runs MB/W SD/FF  --   248  Ie 
M/E - Modular/Excess Air
M/S - Modular Starved Air
MB/R - Mass Burn/Refractory
MB/W - 'MassBurn/Water.all
ND - Not detected.
RDF - Refuse Derived Fuel
. Secondary trammel. bypass.
.. Lime fed at 1200 lb/hr.
... Both ".' and "..- apply.
.a Results reported at 71 02.
b Inlet values reported for cCJll)&rlson purpose.; outlet values Influenced by Na2S Injection..
c Outlet CDD/CDF value.: Unit B - «.2 ng/dlal (average of three run. conducted during s- test c811P8lgn).
d Outlet CDD/CDF values: unit 1 - 6.3 ng/dSC8; unit 2 - 3.8 ng/dsC8. .'
e Inlet CDD/CDF s~les collected during ~rate run. fral Hg, but' duro ~1 ... tat c4111p4lgn and at'shanar
operating condition..

2.3.1 Electrostatic Precipitators (ESP's)
Nine of the facilities reviewed are equipped with ESP's.
with the exc~ption of the data from the Pope/Douglas and the
Oneida County MWC's, and one set of runs from the Detroit and
'Tulsa MWC's, mercury outlet levels range from approximately 200
to 1200 ~g/dscm. The Dayton MWC was the only ESP-equipped plant,
that measured inlet mercury concentrations, and the corresponding

removal efficiencies for the ESP-only averages at Dayton were
both negative (-9.0 and -5.6 percent).
Based on typical uncontrolled levels, the measured outlet
. ,
mercury levels at ESP-equipped MWC's.suggest little or no removal
of mercury by this control device' type. Flue gas ~emperatures

reported for these tests range from 375 of to 600 of. At these
temperatures no mercury qondensation will have occurred;

, ,
therefore, no relationship between flue gas temperature and
mercury outlet levels were ob.served.
2.3.2 Sorbent Iniection/ESP
Two facilities equipped with sorbent injection for acid gas
control followed by an ESP have been tested for mercury
emissions. One set of data was collected from the stack of the
Alexandria MWC, which is equipped with a furnace sorbent
injection (FSI) system. The system generally operates at an ESP
inlet temperature of 360 to 370 ~F., The average emission
concentration from three runs was 517 ~g/dscm. Inlet CDD/CDF
levels were not measured, but were probably low based on
information from other Martin grate systems.2 This suggests that
levels of carbon in the fly ash at Alexandria were probably low.
Three sets of da~a were collected at the Dayton MWC., Two of
the data sets were with FSI, and the third data set was with duct
sorbent injection (DSI). Inlet and outlet mercury levels were
measured during all three test sets. The mercury removal
efficiencies for Runs 10-12 (FSI at an average ESP inlet
temperature of ,394 OF) and Runs 13-15 (FSI at an average ESP
inlet temperature of 298 OF) are both 22 percent, which

correspond to average mercury outlet levels of 7 57 ~g/dscm., and
709 ~g/dscm, respectively. Mercury reductions during Runs. 16-18
(~~I at an average ESP inlet temperature of 306 OF) average 31
percent, cor-responding to a mercury outlet level of 491 ~g/dscm.
Inlet CDD/CDF levels reported for all of these tests average less
than 40. ng/dscm. ...
The data from Alexandria and Dayton indicate that sorbent.
injection/ESP systems may have a small impact o~ controlling
mercury emissions, and that little or no additional benefits are
achieved by lowering flue gas temperature. The limited mercury
control achieved by these sorbent injection/ESP systems may h~ve
been in~luenced by the lack of carbon in the fly ash.
2.3.3 Sorbent In; ection/Fabric Fil ter :i,.
Test data are available for six MWC units at!three sites
with DSI/fabric filter (FF) systems. Operating conditions and
outlet mercury emissions varied considerably during these tests.
During testing of a pilot-scale DSI/FF system at Quebec
City, reported outlet mercury emissions varied from 13 to
614 ~g/dscm. At the lower readings (13 to 40 ~g/dscm), the flue
gas temperature. at the FF in.let was less. than 300 of, and the
inlet CDD/CDF levels averaged between 900 and 2400 ng/dscm. The
high mercury reading (614 ~g/dscm) was at a FF inlet flue gas
temperature of 400 of, and the inlet CDD/CDF concentration
averaged 1600 ng/dscm. Based on:measured inlet mercury
concentrations, mercury removal efficiencies range ~rom
. -

essentially zero at FF inlet temperatures of 400 of or more, to
over 90 percent at FF temperatures less than 300 of.
At Dutchess County, average stack mercury emissions reported
. .-
from Unit 1 (operating at 430 OF) and Unit 2 (operating at
365 OF) are 1080 ~g/dscm and 85 ~g/dscm, respectively. No inlet
mercury or CDD/CDF measurements were made.
At Burnaby, outlet mercury emissions range from 368 to
1086 ~g/dscm. Inlet mercury levels were measured during two sets
of runs and suggest little or no mercury reductions. Flue gas
inlet temperatures at the FF reported for these tests are between
lel j/145
2-6 .

307-325 0p. Inlet CDD/CDP levels, were measured during the. first
set of tests only, and were relatively low at 78 ng/dscm.
These data suggest that mercury removal may be a function of
flue gas temperature and inlet fly. ash carbon content.. At high
flue gas temperatures (greater than 400 0p) or low CDD/CDP levels
(less than 200-300 ng/dscm), mercury control is low. At lower
temperatures and higher CDD/CDP levels, however, the level of
mercury control increases. At Quebec City, mercury reductions'
exceed 90 percent at flue gas temperatures of less than 300 op
and inlet CDD/CDPlevels exceeding 800 ng/dscm.
2.3.4 SDrav Drver/ESP
Outlet mercury emissions data for SD/ESP systems are
available for 14 units at seven MWC facilities. Pour of the
. I
MWC's, Charleston, Haverill, Millbury, and Portland, are mass
burn combustors. , The other three, Honolulu, SEMASS, and West
Palm Beach, are RDP units. o,utlet mercury levels at the mass
burn MWC's range from approximately 210 to 950 ~g/dscm. Reported
ESP inlet temperatures are approximately 300 op or less. The RDP
units operated at similar temperatures. .~utlet mercury levels
for these units, however, range from 5 to 105 ~g/dscm. Due to
the suspension. firing of fuel in. the combustor, RDP units
generally have higher PM loadings.and higher carbon contents at
the combustor exit than do mass burn units. The results from the
RDP units support the theory that increased levels of carbon in
the fly' ash e~hance mercury removal.
2.3.5 SDrav Drver/Fabric Filter !SD/PF)
Mercury data were obtained from 14 MWC's (17 units) that use
SD/PP's. with the exception of one test average from the.
Huntsville MWC, outlet mercury levels from these plants vary from
below detection to 570 ~g/dscm. The high test average from the
. .
Huntsville MWC (1275 ~g/dscm) was due to one high run of over
2700 ~g/dscm.
Plue gas temperatures entering the PP's were less than
30Q.op at all of the SD/FF-equipped facilities. As with the





. , '
The test data indicate that mercury emission levels from
municipal solid waste combustion vary significantly from site to
site. Factors such a~ waste composition, combustion efficiency
(carbon burnout), and'APCD type and operation may affect mercury
removal. Their exact effects are unclear. However, it appears
that good PM control, low temperatures in the APCD system, and
-- ,

significant carbon in the fly ash are necessary to achieve
mercury control. As discussed previously, the combination of a
low PM control device operating temperature and a high level of
carbon in the fly ash (as indicated by the level o~ CDD/CDF at
the combustor exit) enhance mercury adsorption onto particles
which are removed by the PM control device.
SD/ESP data, the lowest mercury outlet levels and highest removal
efficiencies occurred at the two RDF plants, Biddeford and'Mid-
Cqpnecticut. The average inlet CDD/CDF levels at these two
I;>lants were 903~g/dscm and 436 ~g/dscm, respectively., The
Quebec City mass burnjwaterwall (MB/WW) pilot-scale SD/FF test
achieved 94.7 percent removal and had an inlet CDD/CDF level 9f
2157 ng/dscm.
The lowest reported mercury removal efficiency (-27 percen~).
was during the 1987 tests at the Commerce MWC. During this
testing, inlet CDD/CDF levels were low, averaging 28 ng/dscm.
During subsequent testing the following year, average mercury 'f '
outlet levels were approximately 40 ~g/dscm (91 percent removal)
and 70 ~g/dscm (74 percent removal). During the ~988 testing the
mercury inlet levels were similar to those during 'the 1987 test,
but the CDD/CDF levels were much higher, averaging 450 ng/dscm
and 780 ngjdscm. These data further support the theory that
increased levels of carbon in ~he fly ash enhance mercury
It should be noted that the Hempstead MWC, which is a mass
MWC, also had low mercury emissions'~- below 25 ~g/dscm.

levels were not reported.

.Based on compliance test data, the PM operating temp~rature
that is critical for mercury capture is not clear, but appears to
be in the range of 300 to 400 of. Based on published literature
,citing the~retical calculations, however, a temperature of around
300 of appea~s to be necessary.12 The data indicate that RDF
units (which generally have higher uncontrolled PM and CDD/CDF at
the combustor exit, and thus, higher expected fly ash carbon
levels) can achieve mercury emissions less than 110 ~g/dscm when
equipped with a good PM control system operating at less than' '
300 of. The mercury emissions data for mass burn units range
from 10 to over 1000 ~g/dscm, even when equipped with ,the best
acid gas/PM control systems. These. units are generally"
characterized as having low organic emissions, low fly ash carbon
, .J.
content, and low PM loadings relative to RDF units.
, I
kl j/145

Mercury control technologies include the injection of Na2S,
activated carbon or modified activated carbon into the flue gas

. . .
prior to the DSI or SD-based acid gas control system.
Alternatively, wet scrubbing can be used for mercury control.
These technologies' have' not be.en used on U. S. MWC' s, but have
been applied to MWC's in Europe, Canada, and Japan. Brief
discussions of these technologies are presented in this section.
3.1.1 Chemistry ~
Sodium sulfide is a crystalline solid that dissolves in
water to form a solution of up to 10 weight per~~nt Na2S in 34 of
water and 15 weight percent Na2S at 60 of. The/resulting Na2s
solution is sprayed into the flue gas prior to the acid gas
control device. Aqueous Na2S is caustic and will off-gas toxic
hydrogen sulfide (H2S). Th~ reaction of Na2S and Hg precipitates
solid HgS that can be collected in the PM control device. The
specific reactions of Na2S and Hg are not totally understood, but
appear to be:
HgO (gas) + Na2SeH20 --> HgS (solid) + NaOH, and
HgC12 (gas) + Na2SeH2o -->.HgS (solid) + NaCleH20
While (1) flue gas temperature, or (2) lime or ammonia injection
for acid gas or NOx control may affect these mercury reactions,
their effects are uncertain. Testing is currently being
conducted that should provide insight on some of these
3 . 1. 2. . Existina Use of Na~S Bv MWC' s .
Sodium sulfide is or has been used for mercury control by
MWC's in Avesta, Koping, and Hogdalen, Sweden; Kempten, and
Munich (South), Germany; and Burnaby, British COlumbia.
Injection of Na2S has been used at the Hogdalen MWC since 1986~
The Avesta, Koping, and Kempten plants began Na2S injection in
1989. The Munich plant began operation with Na2S injection in
1990. The Burnaby MWC began ,testing of Na2S in 1989 and began
. .
kl j/145

continuous operation with, a temporary system in December '1989.
In October 1991, however, the Burnaby plant is intending to
.switch to activated carbon injection for mercury control.
. All of' these facilities use DSI/FF systems supplied by Flakt
for. acid gas and PM control. Injection of Na2s occurs prior to
the DSI system a~ flue gas temperatures of 265-480 of. Hogdalen
reduces flue gas temperatures prior to Na2S injection with a heat
exchanger that provides hot water for district space heating. , ,
The Burnaby and MunichMWC's use water quench towers for flue gas
cooling. Flue gas temperatures at the stack at Burnaby normally
range from 260-300oF. Additional information on the Munich plant
was not available.
Flakt reports that Na2S feed rates vary from 0.05 to
0.5 kg/Mg (0.1 to 1 lb/~on) of MSW, depending on site-specific
conditions such as the amount of mercury in the flue gas, the
level of control required, and the level of carbon present in fly
ash.'3 As discussed in Secti~n 2.0, residual carbon in the fly
ash is believed to promote mercury removal through adsorption
onto the carbon. As a result, if a plant has little carbon in
its fly ash, it may be nec~ssary to increase the amount of Na2s
Mercury control performance data with Na2S injection are
shown in Tables 3-1 and 3-2. The data have been compiled from
information provided by Flakt, the Burnaby MWC facility owner
(the Greater Vancouver Regional District (GVRD», and from trip
.. . 3 13 14 15 16 17
reports to the. Hogdalen and Burnaby MWC's. ' , , . .
Mercury levels prior to Na2s injection at the Burnaby MWC
(400-1400 ~g/dscm) .are higher than general inlet values. reported'
at European MWC's (55-560 ~g/dscm). The objective of the testing
conducted at the Burnaby MWC was to evaluate key system
parameters. During the initial .tests, 1 to 3 k9/hr (2 to
7Ib/hr) of Na2S was fed as 10-15 percent concentration solutions
. and achieved mercury reductions of 50-65 percent. Subsequent
. .
tests conducted at a feed rate of 2 to 6 kg/hr (4 to 13 lb/hr) of
kl j/145

(ug/dscm) (ug/dscm) (%)
Burnaby 265 DSI/FF Run 1 . 1.0  1465  570  
3/89b   Run 2 . 2.0  993  407  
(10% Na2S)   Run 3 . 2.0  1151  393  
     AVG 1203 AVG 457 AVG 62
Unit 1. 4/89   All runs.  1423  670  
(15X Na2S)    3.0  1443  750  
      1205  473  
     AVG 1357 AVG 632 AVG 53
B/89b Run 1 . 2.5 406  98
(2-4% Na2S) Run 2 . 6.0 775  ..91
 Run 3 = 2.0 670  84
 Run 4 .. 3.0 793  101
 Run 5 . 6.0 661  103
  AVG 661 AVG 95
Unit 1. 12/89 All runs.  HR  138
(2X Na2S)  4.0 HR  67
    NR  146
     AVG 117
Uni t 2. 12/89 All runs.  NR  149
(2X Na2S)  4.0 HR  115
    NR  118
     AVG 127
Uni t 3. 12/89 A 11 .runs .  NR  152
(2% Ha2S)  4.0 NR  159
     AVG 155
AVG 86
...............8....................................8................................... .
NR - Hot reported; HA - Hot applicable
a Results reported at 121 C02 (assummed to be equal to 7X 02).
b Unit not specified.

     (kg/hr) (ug/dlCll) (ug/dlCII) (I)  
Hogdal.n. Unl t 3 265 051/FF        
 (02/86)   343    96   
 (05/86)   316  181  168  7. 
 (08188)   304 1.2   65   
 (08/86)   343 1.2 344  37  89 
 (08/88)   345 1.2 463  57  88 
 (09/88)   304 0.9   51   
 (12/88)   334 1.2   28   
 (02/87)   253 1.2 310  3  99 
 (03/87)   334  124  LOS . 15 t
 ( 04/87)   334 1.2 207  54  74 
 . (09/87)   338 1.2 388  30  92 
 (12187)   338   .I." 139   
 (04/88)   336  288 ' 138  49 
 (05/88)   325 .1.2 369  22  94 
 (08188)   , 311 1.2 121  18  85 
 (08/88)   311 1.2   8   
 (08/88)   311    2Z   
 (10/88)   291 1.8 102  13  87 
 (10/89)   270 1.2 158  142  10 
 (11189)   280 2.4 192  4  98 
 (02190)   280 1.2   9   
 (02/90)   282 1.2.   8   

Na2S and a solution concentration of 2-4 percent achieved average
mercury reductions of 86 percent and outlet mercury
,~oncentrations between 84 and 103 ~g/dscm. Testing conducted at
a Na25 feed' rate of 4 kg/hr (9 lb/hr) and a solution
concentration of 2 percent achieved average outlet mercury

, ,
concentrations between 117 and 155 ~g/dscm; inlet mercury
concentrations were not measured during these tests, therefore
percent reductions could not be calculated. The improved mer~ury
reduction at lower Na2S concentrations is believed to be the
result of improved atomization and mixing when feeding higher
volumes of low concentration solution versus lower volumes of
, ,

high concentration solutions.
Mercury performance data for the Hogdalen ~Unit 3) and
Kempten MWC' s are presented in Table 3-2.14,15,16,17/ Mercury testing
with Na2s injection at'the Hogd'alen facility began in the s~er
of 1986. Testing"prior to the installation of the Na2S inj~ction
system indicated mercury lev~ls as, high as 165 ~g/dscm. Testing
in 1986 with a Na2S feedrate of 1.2 kg/hr (2.6 lb/hr) and
0.9 kg/hr (2 lb/hr) decreased mercury emissions to between 37 and
65 ~g/dscm. When inlet levels were measured, emission reductions

were 88 and 89 percent. Subsequent testing with Na2S injection
in 1987 resulted in similar emission levels, with emission
reductions between 74 and 99 percent. For those tests conducted
without Na2S injection, minimal mercury control was achieved.
In the summer of 1988, a, heat exchanger was 'installed,
replacing a precooler, which resulted in improved performance of
the Na2S system. Other changes to the system have been made over
the course of operation (see discu~sion in section 3.1.3) which'
also improved performance. Test results from 1989 and 1990,
excluding the October 1989 testing, show mercury levels between "
approximately 5 and 25 ~g/dscm. The poor performance during th~
October 1989 testing may be due to the operating conditions
during the test--the boiler load was constantly increased from
low load to full load, which increased the boiler,wall
temperatures and potentially yolatilized mercury adsorbed on
Itl j/145

collected soot. Also note that the high level of performance
during the August 1988 testing without Na2s injection was
measured immediately after a measurement with Na2S injection.
- . . 17
Therefore, the results are uncerta~n.
. Injection of Na2S on the two older units at the Hogdalen
plant was also investigated, but was discontinued since these
units were able to achieve low mercury levels (< 4 ~g/dscm)
without using the Na2S system. The high collection of mercury. J
has been attributed to the high content of unburned carbon in the
flue gases from these. two older units.14,17
Limited information i~ availa~le on the Kempten, Germany
MWC, but outlet mercury levels at Kempten when u~ing Na2S
injection averaged less than 56 Jlg/dscm., :'
As indicated by the. data, typical inlet mercury levels at
the European facilities are lower than those in the U.S. and
Canada. This may be the reason for Hogdalen's lower Na2S
feedrates and the resulting Hg outlet levels at Hogdalen and
K~mpten. The reduction efficiencies, however, are generally
similar for all facilities currently us~ng Na2s injection for
mercury control.
3.1.3 Potential Technical Limitations with Na~s Iniection
All of the existing MWC's using Na2S injection are equipped
with DSI/FF systems. As a result, some uncertainty exists
regarding the applicability of Na2s injection to other APCD
configurations, such as SD/FF'sand SD/ESP's. Potential problems
related to applica~ion of Na2S to spray drying systems include
the existence of adequate time for reaction between'mercury and
. .
aqueous Na2s and pos!ilible reactions between Na2s and acid gas
control sorbent. If Na2S is injected into hot flue gas (e.g.,
450 OF), the associated water may evaporate rapidly, leaving a
dry Na2s particle, which may be less reactive with mercury. If.
Na2S and calcium in the sorbent react to form calcium sulfide
(CaS), the availability of S for reaction with mercury would be
diminished and reduce the mercury collection efficiency. Flakt
stated that they do not belieVe this was a problem, but do not
. "
kl j!145

have any actual operating experience with application of 'Na2S to
spray drying systems. Flakt did indicate, however, that' it would
'probably be necessary to have separate Na2S and calcium sorbent
feed and injection systems to avoid CaS scaling of the sorbent
feed ~ine. A concern related to' the use of ESP's is the
collection capability o~ an ESP if the HgS precipitates a~ a very
f. . 1 t 3,13
~ne part~cu a e. .
The only operating problem reported by Flakt was gas-side:
corrosion of cool surfaces (such as the Na2s piping and nozzles)
by condensation of Hel from the humid flue gas. To prevent this,
hastalloy steel has been used on these surfaces. An initial'
concern raised by the BurnabyMWc operator was a~ apparent
increase of roughly 50 percent in lime consumpttbn rates'
following installation ~f the Na2S system. However, due to
limited operating experience at that time, the plant operator was
not certain whether this increase was due to Na2s use or was
caused by other changes in'plant operations. The plant operator
h~d carefully inspected the DSI/FF system during the plant's last
outage prior to beginning continuous injection of Na2S. The
plant had not been out of ,service since restarting operation,
therefore, the operator had not been able to re-examine the unit
for corrosion or. other problem~ .3,13 Recent information from the
plant did not indicate the occurance of any such problems.'8
Further, recent discussions with plant personnel indicated that.
the lime consumption increase was not related to the use of NA2S
. . t. 19
~nJec ~on.
The Hogdalen plant encountered a problem due to moisture
buildup which resulted in the clogging and plugging ~n ~he screw'
conveyor that was used to transport the sodium sulfide to the mix
tank. Additionally, the sodium sulfide caked up and solidified'"

. .
due to pressurization of the storage silo when transferring the
sodium sulfide. To overcome these problems, Hogdalen eliminated.
the storage silo and developed a system in which 500 kg bags of
sodium sulf ide are emptied directly into the mix tank. 14,17

Another problem encountered at Hogdalen was sludge .buildup
occurring in the mix tank as a result of the presence of.
inorganic salts in the mixing water. To remedy this, treated
boiler feed.water is now used for mixing. other operational
modifications at the Hogdalen plant include the use of piston-
. .

type pumps rather than impeller pumps in order to maintain ~ more
consistent feed rate, and the injection of the sodium sulfide
downstream of the heat exchanger to prevent clogging of the
None of the problems discussed above appear to be of a
magnitude to raise concerns abo~t the ability of Na2S injection
to~control mercury on a continuous "basis.
3.1.4 Cost Estimates for NaaS Iniection }
Available cost data are based on estimates from Flakt,
. .
information provid~d for the Burnaby plant, and supporting
chemical. costs from PPG. The Burnaby MWC operator estimated
capital costs for a Na2S system for the Burnaby plant, which .has
a MSW combustion capacity of 800 TPD, at.$150,OOO-250,OOO (1990
d~llars).3 The chemical costs for the sodium sulfide, as quoted
by Flikt, range from,$0.10~0.50/ton of .MSW.13 This cost is
dependent upon the uncontrolled mercury level and the level of
reduction required. The chemical cost reported for the Burnaby
MWC is $0.30/ton of MSW, and the chemical cost (without shipping)
reported by PPG is $0.30/ton of: MSW, both of which are consistent
with Flikt's estimate. Based on this information, annualized
costs (based on a capital recovery factor of 0.1315 and 8,00.0
hours of operation per year) for.Burnaby are estimated at $0.20-
0.60/ton of MSW.
Another mercury control technology used in Europe is the
injection of powdered activated carbon prior to the APCD. It is.
believed that the activated carbon is a catalyst for the
oxidation of elemental mercury to mercuric oxide, which can be
captured in the APCD.zO This technology has been used
commercially on an MWC locate~.in Zurich, Switzerland, and during
kl j/145
" .

test programs at MWC's in Amager, Oenmark; Kassel, Germany; and
Burnaby, ,British Columbia.
,,. The 'Zurich MWC is equipped with an SO/ESP system. Powdered
activated carbon is injected into the flue gas ahead of the SO,
and the temperature entering the SO is between 430 and 540 of.
Test results from the Zurich plant are shown in Table 3-3. ,
Testing was conducted with and without activated carbon injection
and at SD outlet temperatures between 230-284 of. For tests ~un
without activated carbon injection, the lowering of temperature
did not result in a substantial increase in mercury capture. The
addition of activated carbon, however, increased the average':
percent removal efficiencies from the mid-40's to over
87 percent. It was observed that fluctuatingmeicury inlet
levels did not affect performance when activatedl carbon was used.
, '
The affect of increasing additive was investigated during the
testing at a SD outlet temperature of 248 of. The increase from
9 mg/dscm to 20 mg/dscm to3~ mg/dscm did not have a significant
impact on mercury outlet levels or removal efficiency. In all
cases, the average removal efficiencies were greater than
85 percent, and average outlet levels were between roughly 30 and
90 ~g/dscm. 21 .
The Amager,MWC is equipped with a SD/FF system and operates
similarly to the Zurich plant. Testing was conducted with
temperatures at the SD exit of 284 of and at 260 of. As shown in
Table 3-4, results from the testing with activated carbon
injection at the higher temperature indicate outlet mercury
levels between 23 and 77 ~g/dscm, corresponding to removal
efficiencies between 82 and 95 percent. without activated carbon
injection, outlet mercury emissions were between 67 and
195 ~g/dscm, with removal efficiencies between 15 and 65 percent., .:-
The highest removal efficiencies when using activated carbon
occurred with increased additive levels (70 mg/dscm vs.
7 mg/dscm). 21 '
Testing at the lower APCD inlet temperatures shows greater
control of mercury, especially'when activated carbon injection
kl U145

(mg/dscm) (F) (ug/dscm) (ug/dscm) (X)  
o 284 703 510  27 
  449 310  31 
  890 546  39 
  730 542  26 
  531 438  17 
  1403 1003  28 
 AVERAGE 784 559  29 
9. 248 656 122  81 
  574 .62  89 
 AVERAGE 615  92  85 
20 248 300  43  86 
  230  14  94 
  313  29  91 
 AVERAGE 281  29  90 
39 248 705  51  93 
  771  41  94 
  202  '26  87 
  148  30' . 80 
 AVERAGE 457  37  92 
. '     
o 239 648 304  53 
  841 271  68 
  306' 153  SO 
  1242 877  29 
  963 623  35 
  525 327  38 
 AVERAGE 754 426  44 
39 239 461  58  88 
  462  58  88 
  368  38  90 
 AVERAGE 430  51  88 

(mg/dscm) (F) (ug/dscm) (ug/dscm) (X)  
o 230 326 162  50 
  293 173  41 
  453 277  39 
 AVERAGE 357 204  43 
 230 636  B9  86 J"
  851  59  93 
  171 " 58  66 
  546  52  90 
  352  67  81 
  511  65  87 

TABLE 3-4.
-(mg/dscm) (F)
a.......m...............................................a............ .
o 284 203 154 24 
  229 195 15 
  219 86 61 
  202 74 63 
  165 67 S9 
 AVERAGE 204 115 44 
7 284 378 58 85 .1.
  227 40 82 
 AVERAGE. 303. 49 84 
20 2B4 214 31 86 
  248 35 86 
  - 336 36 89 
 AVERAGE 266 34 B7 
70 2M 1516 77 95 
  318 23 93 
 AVERAGE 917 50 94 
o 260 421. 32 92 
  196 48 76 
  163 30 82 
  189 54 71 
 AVERAGE 242 41 80 
23 260 201 24 88 
83 260 198 6 97 
  220 7 97 
 AVERAGE 209 6.5 97 

was not used. With activated carbon, outlet mercury levels
ranged from 6 to 24 ~g/dscm (88 to 97 percent removal), and
,without activated carbon the outlet levels were between 30 and
53 ~g/dsc~(72 to 92 percent removal) .
. .

The MWC in Kassel is equipped with an ESP followed by a
. ..
SD/FF, and the system has the capability of operating in either
the single-pass of partial-product recycle mode. (Zurich and
Amager are single-pass systems.) The recycle design of the'
Kassel system results in an increased chloride content of the
lime slurry, therefore, the temperature exiting the SD must, at a
minimum, be kept at 275 of. Recent test results at a temperature
of 279.oF and carbon additive feed~ates of between 0 and
84 mg/dscm are in shown in Table 3-5. ~ithout a~tivated carbon
injection average mercury outlet levels exceeded 750 ~g/dscm .
(35 percent removal). By using carbon injection at levels of
25 mg/dscm and higher, mercury levels less than 75 ~g/dscm were
achieved, with removal efficiencies. exceeding 80 percent.
Recently available information on the use of carbon
injection at the Burnaby MWC indicate significant mercury
reductions. Results'of te~ting from June 1990 through January
1991 show mercury removal efficiencies averaging 84 percent. It
'has been determined that the final installation at the Burnaby
plant will be a carbon based injection system. The annual
operating costs with such a system will be less than an NazS
injection system, and the health risk to employees (caused by the
offgassing of NazS) will be reduced. It is expected that the
final mercury control system will be operational in October
1991. 18
List prices for activated carbon range from $0.50-1.00 per
pound depending on the raw material used to produce the carbon
and the available s.urface area. 22 Based on a carbon feed rate 0+
1 to 2 kg/hr, these costs correspond to $0.lS-0.35/ton of MSW."
Estimates of capital costs for the construction of an activated
carbon injection system at the Burnaby MWC are on the order of
$200,000 (1990 dollars). With this information, the annualized
Id j/1105

TABLE 3-5.
 o 279 1175 762 35
 12 273 440 229 48
 26 279 424 75 82
 62 279 234 25 89
 84 279 389 68 82

costs (based on a capital'recovery factor of 0.1315 and
8000 hours of operation per year) for Burnaby are estimated at
'$0.30 - o.SO/ton of MSW. One of the criteria for selecting the
additive used at the Zurich MWC was its low cost.n
. .

Another mercury control technique in use in Europe is the
injection of an activated carbon scrubber additive consisting of
approximately 95-97 percent lime and 3-5 percent activated
carbon. One of the first tests using this additive. was on the
MWC in Geiselbullach, Germany, in January 1989. The plant is
equipped with a DSI/FF system and has two lines, each capable of

. .
combusting 158 tpd of MSW. Despite high operati~g temperatures
during the initial testing (sometimes as high aS~~65 of exiting
the combustor), mercury emissions were reduced from inletle~els
of 250 to 330 ~g/dscm, 'to outlet levels under 110 ~g/dscm.
Subsequent testing with inlet levels between 140 and 640 ~g/dscm
resulted in outlet levels between 12 and 46 ~g/dscm. 24,25
Along with the Geiselbullach MWC, activated carbon/lime
injection has been used on other MWC's .in Germany including the
. .
Berlin-Ruhleben, WurZ'Durg,..and Siemens-KWU MWC's. During full-
scale testing using this addit~ve at the Berlin-Ruhleben MWC,
which is equipped with a SD/FF system, inlet mercury levels
averaged 444 ~g/dscm for Boiler 2 and 402 ~g/dscm for Boiler 3.
Outlet levels were reduced to average levels of 99 ~g/dscm and
83 ~g/dscm for Boil~rs 2 and 3, respectively. 25,26
At the WUrzburg MWC (DSI/FF), mercury emissions were reduced
by over 80 percent to levels under 65 ~g/dscm when using
activated carbon/lime injection. At the Siemen~-KWU MWG, which
is equipped with a wet scrubber followed by a FF, injection of
this additive is used as a final purification stage. Outlet
levels of mercury during testing were less than 25 ~g/dscm.25
Sorbalit (Manufactured by Marker Zementwerk GmBH) was
the commercial produced used. All test.results
discussed in this section are based on the use of this
kl j/145

Activated carbon/lime injection has also been used on
special waste incinerators in Germany (Schoneiche and Schwel-
Brenn-Anlage), where substantial reductions in mercury were
observed.~~~ .
Details on costs associated with the use of this additive
:were not available, but similar to activated carbon, low
investment costs and easy management are cited as two of the
advantag~s of the product.
.. Wet scrubbing is a form of acid gas and metals control that
has been used primarily at MWC's in Europe and Japan. Wet
scrubbing of MWC flue gases typically involves passing the flue
gas through an ESP to reduce PM, followe~ by a t~o-stage absorber
where flue gas is contacted with water to remove/HCl in the first
stage and an alkaline solution in the second stage to removeS02.
The absorber also saturates the gas stream and reduces flue gas
temperatures to as low as 13Q of. Several wet scrubber designs
include a fine PM collection system following the second absorber
to reduce aerosol and fine particulate emissions. The alkaline
solution, typically c~ntai~ing calcium hydroxide [Ca(OH)2]'
reacts with the acid gas to form salts, which are generally
insoluble and may be removed by sequential clarifying,
thickening, and vacuum filtering. The dewatered salts or sludges
are then landfilled.
Due to the low absorber operating temperature-that promotes
mercury condensati~n, wet scrubbing technology achieves high
mercury reduction. Mercury emissions can be reduced by up to
90 percent. The us~ of liquid chelating agents enhances. the
coagulating sedimentation and the fixation of the mercury
compound in the sludge.21 Disadvantages of wet scrubbing,
however, include the quantity' of. water required and potential

. .
difficulties with waste handling. To stabilize condensed mercury
compounds, use of additives, such as.TMT (trimercapto-s-triazine)
is required in some wet scrubber designs. Failure to stabilize
and remove collected mercury c.ompounds from the scrubber solution

cari result in revolatilization of mercury from the scrubber
solution and, thus, reduced collection efficiency. Further,
,while mercury control may be higher for wet systems, control of
organic em~ssions may be lower than that achieved w~th dry acid
gas controls.
Test results from three wet scrubber-equipped MWC plants
located in France and Switzerland are available. The two French
plants, Lyon-Nord and Lyon-Sud, began commercial operation in.'
1989 and 1990, respectively, and they are equipped with ESP's
followed by wet scrubbers. Mercury emissions results from these
plants are shown in Table 3-6. Average mercury outlet emissions
at Lyon-Nord were under 50 ~g/dscm for Unit 1 an~ 62 ~g/dscm for
Unit 2. Average removal efficiencies were great~r than
82 percent for unit 1 a~d62 percent for Unit 2.' At Lyon-Sud,
average mercury outlet emissions were less than approximately
60 ~g/dscm for both units, and average removal efficiencies were
. u.
greater than 86 percent. -
The Basel, switzerland MWC, which was originally equipped
with only ESP's, was retrofitted with wet scrubbing in 1989.
Mercury outlet emissi6ns, ,listed in Table 3-6, ranged from
16 ~g/dscm to 20 ~g/dscm at Unit 1, and from less than 13 ~g/dscm
. .

to 34 ~g/dscm at Unit 2. This corresponds to average removal
efficiencies between 90 and 96 percent for Unit 1, and between. 82.
and 96 percent for Unit 2.28 The higher mercury control
efficiency at the Basel MWC may reflect higher unburned carbon.
levels in the fly ash from the older Basel unit as compared to
the new combustors in Lyon.
In addition to the mercury control technologies discussed in
the previous sections, final stage activated carbon beds are .,
being investigated as an applicable mercury control technique. .
Activated carbon beds are "back .end" controls, positioned afte~ .
all other APCD's in the system. As .the flue gases pass through
the bed, pollutants adsorb onto the porous surface.

   (ug/dscm) (ug/dscm) (X)
LYOH-NORD.  1   168 <49 >71
FRANCE  2   289 <50 >83
  3   578 <49 >91
 AVERAGE    345 <49 >82
 2 4   177  49 72
  5   177  76 57
  6   140  60 57
  AVE~GE  165  62 62
LYON-SUO.  3   457  72 84
FRANCE  4   568 <49 :.!>91
  AVERAGE  513 <61 I >88
 2 1   438  69 84
  2   373 <49 >87
  AVERAGE  406 <59 >86
BASEL.  1   252  16 94
SWITZERLAND  2   168  17 90
  4   401  17 96
  7.   513  20 96
  8   187  19 90
 2 1   186 <13 >93
  2   224 <14 >94
  3   261  32 88
  4   224 <13 >94
  5   168 <13 >92
  6   168  21 88
  7   140 <13 >91
  20   363  13 96
  21   270  33 . 88
  23   75 <13 >82
  24   196 <13 >93

A potential disadvantage to the carbon bed technology
includes the possibility of-fires, since activated carbons are
'~aturally flammable and may self-ignite at temperatures as low ,as
175 OF.~ ~lso, if the bed is regenerated, special precautions
must be taken to ensure the capture of any resulting mercury
Another emerging mercury control technology includes the use
of selenium filters. Such filters are used in metallurgical,'
smelting operations, and consist of a cylindrical shell which
contains graded porous material impregnated with selenium. '
, , ,
Selenium has a strong affinity of mercury. Flue gas exiting' 'an
ESP would pass through the filter prior to the s~ack. The
filters would need to be replaced once they are:~pent.29

. 1c.
CE Resource Recovery Systems. Meeting Summary, Meeting on
Munic~pal Waste Combustors (MWC's) - Add-on Control of .
Mercury Emissions Attachment 10. U. S. Environmental
Protection Agency, Research Triangle 'ark, North Carolina.
February 7, 1990.
Telefax. Hartman, M., Combustion Engineering to D. White,
Radian Corporation. Detroi t Compl iance Tests. September.,
1990. . . .
Sussman, D. B. (Ogden Martin Systems). Testimony Before the
National Air Pollution Control Techniques Advisory
Committee. Research Triangle Park, North Carolina. January
3]., 1991. -
U. S. Environmental Protection Agency.
Combustors - Background Information for
Post-Combustion Technology Performance.
August 1989. -
Municipal Waste
Proposed Standards:
Trip Report. Burnaby ~C, British Columbia, Canada.
D., Radian Corporation. - May 1990.

Permit No. 0560-0196 for Foster- Wheeler Charleston Resource
Recovery, Inc. Municipal Solid Waste Incinerators A & B.
Charleston, SC. Bureau of Air Quality Control, South
Carolina Department of Health and Environmental Control.
Entropy Environmentalists, Inc. for Honolulu Resource
Recovery Venture. Stationary Source Sampling Final Report.
Volume I. Oahu, Hawaii. February 1990.
Woodman, D.E. -Test Report Emission Tests, Regional Waste
Systems, Portland, ME. February 1990.

Eastmount Engineering, Inc. Final Report, Waste-to-Energy
Resource Recovery Facility, Compliance Test Program, Volumes -
II - V. (Prepared for SEMASS Partnership.) March -1990.
Entropy Environmentalists, Inc. for Babcock & Wilcox Co.
North County Regional Resource Recovery Facility, West Palm
Beach, FL. October 1989.
- -
Ogden Projects, Inc. for Ogden Martin Systems of ~abylon,
Inc. Environmental Test Report~ units 1 and 2, Babylon
Resource Recovery Facility. Babylon, NY. February 1990

]..4 .
kl il145
Radian Corporation for American Ref-Fuel Company of
Hempstead. compliance Test Report for the Hempstead
Resource Recovery Facility, Westbury, NY. Volume I.
December 1989.
Ogden projects, Inc. Environmental Test Report;
. Indianapolis Resource Recovery Facility, Appendix
Appendix B, Volume I. (Prepared for Ogden Martin
Indianapolis, Inc.) August 1989.
A and
Systems of
Bergstrom, J. Mercury Behavior in Flue Gases. Waste
Management and Research. 1986 - 4~ pp. 57 - 64.
Telefax. Frame, G., Flakt, Canada, to D. White, Radian
Corporation. Summary from GVRD. January 1990.
( ,
Trip Report. Hogdalen MWC, Sweden. Hereth, M.
Limited. London, England. August 1990.
Telefax. Nilsson, B., Flakt Industry AS, to D. White"
Radian Corporation.,' May 1990.
Telefax. Carlsson, K., Flakt Industry AS, to D. White,
Radian corporation. June 1990.
Andersson, C. and B. Weimer (Hogdalen Plant). Sodium
Sulphide Dosing at the Hogdalen Plant in Stockholm.
Proceedings of the Second Annual International Conference on
Municipal Waste Combustion, Tampa, Florida.
April 15-19, 1991. .

Guest, T. and O. Knizek. Mercury Control at Burnaby's
Municipal Waste Incinerator~ Proceedings of the 84th Annual
Meeting and Exhibiion of the Air and Waste Management
Association, Vancouver, British Columbia. June 16-21, 1991.
Telecon. Knizek, 0., Greater Vancouver Regional District,
and K. Nebel, Radian Corporation. August 1991.

Schager, P. The Behaviour of Mercury in Flue Gases.
Department of :tnorganic Chemistry, Chalmers Institute of
Technology and University of Goteborg. Goteborg, SWeden.
Felsvang, K. S., Sander Holm, T., and Brown, B. Control of
Mercury and Dioxin Emissions from United States and European
MSW Incinerators by Spray Dryer Absorption Systems.
Proceedings of the Second Annual International Conference on
'Municipal Waste Combustion, Tampa, Florida. April 15-19,
. ,-

kl j/145
Letter from B~ Brown, Joy Environmental Equipment Company,
to T.G. Brna, U.S. Environmental Protection Agency,CRB/ORD.
May 10, 1991.
Memorandum from T. G. Brna, U.S. Environmental Protection
Agency, CRa/ORD, to W. H. stevenson, U.S. Environmental
Protection Agency, SDB/OAQPS. Mercury Emission Control from
Municipal Waste Combustors using Activated Carbon Injection
in Flue Gas. August 20, 1990. .

Nethe, L.P. Ein weg zu Weniger Queksilber and Dioxin.
Umvelt Magazin. November, 1990.
Information provided by MarkerZementwerk GmbH.
April 1991. .
Clarke, M.J. A Review of Activated Carbon Technology
Reducing MSW Incinerator Emissions. 1991 Proceedings
Second International Conference on.Municipa~ Waste
Combustion, Tampa, Florida. April 15-19, ~991.

Nakazato, K. Latest Technological Experience of the Removal
of Mercury in Flue Gas and the Management of Fly Ash from
MSW Incinerator. Takuma Co., Ltd.
of the
Belco Technologies Corporation. Presentation of Performance
Test Data to U.S. Environmental Protection Agency on
Municipal Waste Combustors. AUgus~ 1990.

Lindquist, B. Gas Cleaning in Connection with Waste
Incineration. 1991 Proce~dings of the Second Annual
International Conference on Municipal Waste Combustion,
Tampa, Florida. April 15-19, 1991. .