United States
                   Environmental Protection
                   Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
                   Research and Development
EPA/600/S7-88/023  Dec. 1988
&EPA         Project  Summary
                   Effects of Sorbent Injection for
                   Sulfur  Dioxide  Removal on
                   Particulate Control Systems  for
                   Coal-Fired  Boilers
                   J. P. Gooch, J. L. DuBard, M. G. Faulkner, G. H. Marchant, Jr., R.
                   S. Dahlin, and R. Beittel
                    Various studies were undertaken to
                  quantify the effects  of  dry  SO2
                  sorbent injection on ESP  operation
                  with  a coal-burning utility  boiler.
                  The specific dry sorbent operation of
                  interest was EPA's LIMB process.
                  The combination of spent sorbent
                  and fly ash has a higher resistivity, a
                  higher mass concentration, and a
                  finer  particle-size distribution  than
                  the ash alone; all of these  factors
                  diminish the effectiveness of  ESP.
                  Also  investigated was  chemical
                  conditioning to reduce the resistivity
                  concern, the only one  of  the three
                  concerns stemming from sorbent
                  injection  that  can  be readily
                  mitigated. Other topics studied were:
                  the recycle, disposal, and  utilization
                  of waste ash-sorbent mixtures; the
                  selection   and  modification  of
                  sorbents to improve SOg  capture in
                  the furnace; and the reactivation of
                  spent sorbent by humidification to
                  achieve supplemental  post-furnace
                  capture of SO2-
                    This Project Summary was devel-
                  oped  by  EPA's Air and Energy
                  Engineering Research Laboratory,
                  Research  Triangle Park, NC,  to
                  announce  key  findings of   the
                  research  project that  is fully
                  documented in a separate report of
                  the same title (see Project  Report
                  ordering information at back).
Introduction
  Limestone Injection, Multistage Burners
(LIMB) is being developed by EPA as a
way to control  SC>2 emissions from
coal-fired  boilers. This technology  is
one of several  based in  dry-sorbent
injection  that are  expected  to  be
adaptable in  retrofit  applications to
boilers  that  predate new-source
performance standards. It provides only
partial  reduction of SOj emissions at
economical operating costs;  however, it
requires a lesser capital investment than
more efficient conventional  processes,
such as wet scrubbing, and  is  thus a
candidate for  application  where old
boilers continue in service.
  One  concern m implementing LIMB
technology is the difficulty introduced  in
maintaining  acceptable  levels of
particulate emissions. The concentration
of suspended solid matter that must be
removed is greatly increased; moreover,
the fraction of the solids consisting of
small  particles  (and thus  especially
difficult to remove) is substantially
increased.  These changes in suspended
solids  will  have  unfavorable
consequences  in any particulate-control
device. Their impact  is most serious,
however, in terms of the performance of
electrostatic precipitators (ESPs), since
most of the boilers where LIMB may be
applied operate in conjunction with ESPs
for particulate control.

-------
  The primary concern of  this research
was  the  problem  of  operating  ESPs
successfully in  utility  plants  equipped
with LIMB. Three of the five major tasks
addressed  this  central  issue.  A
secondary concern  was  to  provide
technical  support  of the full-scale
demonstration  of  LIMB  that EPA  is
undertaking at the Edgewater  plant  of
Ohio Edison  at Lorain, Ohio. The two
remaining tasks were in this area.

Project Scope and Objectives
  The five technical tasks  that provided
the organization structure of this project
were:

  1. Development  of a predictor of  the
    electrical  resistivity of mixtures of fly
    ash and  sorbent  that result from
    sorbent  injection into coal-fired
    boilers.

  2. Determination of  the effect of  high
    mass  loadings  of  ash/sorbent
    mixtures on ESPs.

  3.Designing  and  implementing a
    pilot-scale  program  of ESP  with
    furnace sorbent injection.

  4. Investigation  of  recycle, disposal,
    and utilization  of  LIMB ash/sorbent
    mixtures.

  5. Studies  with   a  pilot-scale
    combustor to characterize sorbents
    that may  be applicable to the LIMB
    process.

  Tasks  1-3  addressed  different
aspects  of  the concern  of  ESP
performance  in  conjunction  with a
LIMB-equipped  furnace: the  high
electrical resistivity  of the ash/sorbent
mixture  resulting  from the  LIMB
operation;  the  substantially elevated
concentration of particulate  matter to be
collected;   and   the   long-term
maintenance of  a   pilot-scale   ESP
attached to  a  moderate-  or full-size
furnace. An objective  under Task 5 was
the evaluation of sulfur tnoxide gas or
water vapor as  a  means of lowering the
high resistivity of an ash/sorbent mixture.
  Task 4 addressed issues of importance
to the LIMB demonstration  at Edgewater.
Can spent sorbent in the form of calcium
oxide be rejuvenated  by  hydration  and
reinjected as  calcium  hydroxide  to
achieve more complete utilization? Can
spent sorbent be disposed of in a landfill
without contaminating  groundwater? A
third question, less related to Edgewater
than to the broader future application of
LIMB, can a commercial market be found
for LIMB wastes?
  Task  5  addressed  a number  of
additional  issues  of  concern   at
Edgewater: the  selection of  a sorbent
from those available  in the market;  the
enchancement of sorbent activity  with
lignosulfonate (a dispersing agent) as  an
additive;  and the reactivation of spent
furnace-injected  sorbent by humid-
ification of flue gas in a low-temperature
duct, to achieve post-furnace capture of
SOa This task also addressed the issue
of chemical  conditioning  of ash/sorbent
mixtures to lower their resistivity.

Results and Discussion
  Task 1. Development of a resistivity
predictor for ash/sorbent mixtures. The
first  step was  to  obtain a  variety  of
samples  of  ash/sorbent  mixtures  and
characterize them with respect to mineral
composition, particle  size, surface area,
density, and electrical resistivity.  Efforts
were then made to establish correlations
between  resistivity and one or more of
the other properties.  In  particular,  an
attempt  was  made to  establish a
relationship between  resistivity and  the
sum of lithium and  sodium contents,
which is the basis of an earlier correlation
for  ordinary  fly  ash; this  attempt failed.
During the project, it was discovered that
the  procedure  used  for  resistivity
measurements altered the surface  area
of the ash/sorbent mixtures  and  thus
altered the resistivity.
  The need  for a revised procedure  for
measuring resistivity  was thus evident.
Work toward that end will be described
in the report of another project. The effort
to develop  the  resistivity  predictor,
however,  remains  unfulfilled;  the  pros-
pects for  completing this  effort
successfully cannot be  evaluated from
the data now available.
  Task 2. Investigation  of high  mass
loadings in  ESPs. This  task  was
addressed through  both experimental
and  theoretical  approaches.  The
experimental effort centered  on the  use
of a specially instrumented ESP that was
operated in  conjunction  with a  pilot-
scale combustion  facility,  which
produced an ash  of high resistivity (2 x
1012  ohm-cm), an elevated  concentra-
tion  (higher by  a  factor of three), and a
particle-size  distribution with a  much
greater  proportion of fine material  The
primary  observations were  that:  there
was an  ESP malfunction owing  to  the
high  ash resistivity,  and there  was a
strong  space-charge suppression  of the
corona  (observable  when  the  ast
resistivity was lowered with 803) due t(
the  fineness  of  the  particle-siz*
distribution. The data were used  in the
existing ESP  model to project the effec
on ESP emissions  of adding  the  LIME
process to a typical boiler,  shown ir
Figure  1.
  The  theoretical  study  consisted o
development  of a revised  version  of the
ESP model,  which  is responsive to  £
wider range of ESP  operating conditions
Comparisons  of the  revised model with z
limited  number of field data sets indicate
that  the  revised model  functions satis-
factorily  as  a   predictive  tool
Nevertheless, there was  an  indicated
need for certain further refinements in the
model.
  Task 3. Performance  of a pilot-scale
program  of ESP  performance  with
furnace sorbent injection.  Arrangements
for pursuing this task at Base  Gagetown
of the  Canadian armed forces in New
Brunswick were completed,  and certain
background data were  obtained.
Ultimately,  however, the  pilot facilities
assembled at Gagetown  were dismantled
at EPA's  direction  before  the planned
measurement  program  could be
completed. A crucial factor that led to this
action  was the abnormally high  carbon
content of the  ash produced  in the
Gagetown burners, which would have had
a serious impact on ESP  performance,
regardless of the  impact  produced  by
sorbent injection.
  Task 4. Studies of LIMB ash recycle,
disposal, and utilization. The  results of
these studies demonstrate that LIMB ash
can be reactivated by hydration and can
then be  reinjected  to  achieve  further
utilization for SO2  capture  A  major
stumbling block  to  this  approach,
however, is the excessive  mass loadings
that will occur unless the sorbent can be
separated from the fly ash. Potential
separation methods  that deserve  further
study are based on a cyclone or an ESP,
a dense-phase  classifier, or a  steam
elutriator.
  Both laboratory and field tests showed
that LIMB ash can be safely handled and
used  for  landfilling.  Cementitious
reactions  in  the  moist  ash  produce a
nearly  impermeable product, which has a
minimal  potential for  groundwater
contamination.  LIMB   ash  cannot  be
classified  as  a hazardous waste under
any of  the RCRA criteria.
  Synthetic aggregates  represent  one  of
the most  promising  uses  for LIMB ash
Such aggregates are likely to conform

-------
          20
                      SCA, m*/m3/s

                  30      40      50
                                          60
1.0




0.6

0.5

0.4


0.3



0.2
   I
   ^
   1
        0.1
       0.08
   E  0.06

   I  0.05
   3

   °  0.04
      0.03
      0.02
       0.01
           -      \
\
                                        sh Plus CafOHh
                                     (1) Coal Ash Only
                           I
                           I
                                         400


                                         300



                                         200






                                         100

                                         80


                                         60

                                         50




                                         30
                                                                 c

                                                                 g
                                                                 •Q
                                                                 si
                                                                 o
                                                                         
-------
  vapor alone cannot activate the solid
  appreciably. Physical wetting, on the
  other hand  which occurs  when dry
  entrained sorbent particles and water
  spray droplets collide,  has  the
  potential for strongly activating the
  solid. The  practical limitation of the
  wetting process is the infrequency of
  particle-droplet collisions, which
  allows most of the sorbent particles
  to remain dry and hence unreactive.

• Charge-augmented   sorbent
  humidification (CASH) was  studied
  briefly.  In  this process, sorbent
  particles were charged  negatively
  and water  droplet  positively,  to
  enhance the collision frequency.
Further study of the concept was
pursued under another project. In the
final analysis, no beneficial result of
the process  seems to be  in  the
offing.

The effects of SOa gas and water
vapor   were   compared  as
conditioning agents, to  alleviate the
concern  with high resistivity  of
ash/sorbent mixtures in an ESP. At
concentrations  of  about 60 ppm,
SOa  was  found   to  be  largely
ineffective - much  less promising
than  had been  indicated  by prior
studies. At addition  levels of  around
5%  by  volume  (which   gave
approaches to adiabatic saturation of
    11°C),  water vapor, on  the  othe
    hand,  produced  a  much  mor<
    promising result  -  a reduction ii
    resistivity from 1 x 1013 to 1 x 101
    ohm-cm.  The effect of water  vapo
    is shown in Figure  2.

Recommendations
  Further research should be done tc
complete  or augment  work  initiate(
during  this  project.  Examples  o
objectives  not  yet satisfied  are  the
development  of  a resistivity  predicto
(Task 1) and the long-term evaluation o
a pilot-scale ESP  operating  with  LIME
particulate matter   generated  in  <
moderate-  or  full-scale combusto
(Task 3). Efforts to complete Task 1 have
/ V
7014
;o13
I
I
:v.
>•
S
? ,_12
« 10*
7011

/O10
WOO/Tfl
°
0
fill I I I I
A Spark Data
O V-IData
A*
— A —
<9
o
A
^ ^ A^
* Vs
AA°0* A
- •— * 11
°0 *
o°o g
%o°
Humidification • • No Humidification
\
\ \ \ \ \ \ \ \ \
0 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.
C 40 60 84 112 144 182 227 283 352 44
f 103 141 183 233 291 359 441 541 666 82
                                   Temperature
Figure 2.   Resistivity of ash/sorbent mixture as a function of temperature and humidification.

-------
been made under  another project,  but
have not yet reached a satisfactory  end
point.  The long-term ESP test is now
rescheduled  under another project  as
part  of the  LIMB demonstration  at
Edgewater. Further developments  of the
revised ESP  Model from Task 2 have
been made under another project.

  Certain  findings under Tasks  4  and 5
can be implemented  at  Edgewater  if
required:  The LIMB wastes  may  be
safely  disposed  of in a  landfill;  the
Ca(OH)2 sorbent selected for Edgewater
may  be treated  with  lignosulfonate
additive, either in solid or liquid form, to
achieve  enhanced activity; and  the
concern with high electrical resistivity of
the waste-ash/sorbent  mixture  may be
dealt  with adequately  with water-vapor
conditioning.

  Certain results under Tasks 4  and 5
indicate  that certain  modifications  to
LIMB (to  enhance the  process)  may
require  further  work.  Further work is
needed  to  make  the  recycle  or
commercial utilization of LIMB wastes
practical  and  economical.  Additional
studies are needed to permit  effective
post-furnace  reactivation of  spent
sorbent by  humidification; the data now
available,  however,  reveal  that the
primary limitation on the effectiveness of
humidification is  the  infrequency  of
collisions of sorbent particles and  water
droplets.  The   charge-augmented
humidification   process  that   was
investigated seems to offer no benefit.

-------