United States
                    Environmental Protection
                    Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
                    Research and Development
EPA/600/S7-89/001 Sept. 1989
&EPA         Project  Summary
                    Humidif ication of Flue  Gas to
                    Augment SO2 Capture  by  Dry
                    Sorbents

                    J. P. Gooch, R. Beittel, E. B. Dismukes, and R. S. Dahlin
                      In a  coal-burning power plant,
                   humidification of the flue gas in a low
                   temperature duct is a possible way to
                   increase SO2 removal by dry calcium-
                   based  sorbents. In  particular,
                   humidification may be a desirable
                   modification of EPA's LIMB process,
                   which is based on the injection  of
                   limestone or  hydrated  lime  in the
                   furnace; it could  augment SO2 re-
                   moval  by adding post-furnace re-
                   moval  to that occurring  in the
                   injection zone.  Southern Research
                   has investigated  certain  aspects  of
                   low-temperature  S02  removal  in
                   humidified flue gas as part  of the
                   research effort funded by the EPA.
                      This Project Summary was devel-
                   oped by  EPA's  Air  and  Energy
                   Engineering Research  Laboratory, Re-
                   search 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
                     EPA has been developing the dry-
                   sorbent technology known as LIMB over
                   the past several years  as  a means  of
                   lowering  S02  emissions from coal-fired
                   power boilers.  The  acronym stands for
                   Limestone Injection, Multistage  Burners.
                   Originally, the process was envisioned as
                   one based on limestone as the specific
                   calcium-based  sorbent;  currently,  how-
                   ever, the process emphasizes the use of
                   hydrated lime — that is,  Ca(OH)2 — as a
                   more reactive sorbent in place of  lime-
                   stone. LIMB based on  either  sorbent,
                   however, depends on the injection of the
                   sorbent in the furnace at a temperature
                   above 1100°C, which  leads to the
calcination  of the  sorbent and  the
reaction of the resulting CaO with SO2 in
the presence of O2 to produce CaSO4.
  The utilization of the sorbent under
these conditions does not exceed 25 to
30% on the mole basis  and, thus,
accounts for removal of 50 to 60% of the
S02 present at a practical 2:1 Ca/S mole
ratio. Such  utilization levels are accept-
able  for retrofit  operations in boilers
predating new source performance regu-
lations. Still, increased  utilization is
obviously desirable if it can be attained.
  Humidification  and cooling of  the flue
gas in a post-furnace duct, beginning at a
temperature of about 150°C,  is  one con-
ceivable way to enhance the utilization of
furnace-injected sorbent. At relatively low
temperatures, SO2  can react  with the
calcine of limestone or hydrated lime to
produce CaSO3; whereas, at high tem-
peratures in the furnace, it reacts only in
combination  with 02 to produce CaSO4.
The critical issue, however, is not temper-
ature and the thermodynamic possibility
of sulfite formation but humidification and
the kinetic enhancement of low-temper-
ature reactions,  which would  be slow
without humidification.
  Humidification is a way to enhance SO2
removal with furnace  injection of dry
sorbents and to remove SO2 with direct
injection of  Ca(OH)2 in  either the dry or
the wet state. Extensive research under
the auspices of the Department of Energy
has been recently performed and is  now
continuing on duct-injection  processes,
which take two basic forms: (1) separate
injections of dry  Ca(OH)2 and  a water
spray, and 2) simultaneous injection  of a
Ca(OH)2 and water  in a slurry. In either
process,  the  sorbent reaches  the
particulate collector (ESP or baghouse) in

-------
a dry state, after the water has evapo-
rated.  Maintaining  proximity  to  vapor
saturation,  however,  is  the key to
effective S02 removal.

Scope of the Investigation
  This research project addressed  four
topics within the general area of sorbent
reaction with S02 under  conditions of
low-temperature flue-gas humidification,
•  Pilot-scale  investigation  of  post-
   furnace humidification  to  achieve
   sorbent activation. This investigation
   was carried  out  with Southern  Re-
   search's 1 x 106-Btu/h  (300-W)  coal
   combustor  to  support  EPA's LIMB
   demonstration  at  Ohio  Edison's
   Edgewater power plant at Lorain, Ohio.
   The central  issue was  to determine
   what humidification  conditions  are
   required for  sulfated calcine from the
   furnace  to undergo  low-temperature
   reaction with S02. Is humidification by
   water  vapor  alone  adequate, or  is
   humidification with  a  water  spray
   required because of the  need  for
   physical wetting of  the  sorbent
   particles?
•  Pilot-scale investigation of charge-
   augmented  sorbent  humidification
   (CASH). Once again,  the  Southern
   Research pilot-scale  combustor  was
   used to address a question relevant to
   the Edgewater demonstration.  Can
   opposite electrical charging of sorbent
   particles and water droplets increase
   collisions sufficiently to  enhance  S02
   removal?
•  Laboratory  studies of  water-vapor
   adsorption by Ca(OH)2 with and with-
   out additives.  A  brief  experimental
   program was conducted to quantify
   the effects  of certain  additives on
   water-vapor  uptake  by  Ca(OH)2.  If
   Ca(OH)2 is  to be used in a duct-
   injection process,  use of a  deli-
   quescent additive may  be  desirable.
   Alternatively,  if  partially   sulfated
   calcine from furnace-injected sorbent
   is to be rehydrated outside  the power
   plant and reinjected in a duct process,
   it may be advantageously treated with
   a water-attracting additive.
•  Mathematical modeling of humidifica-
   tion and S02  removal processes. A
   mathematical model was developed to
   treat two processes: (1) the collision of
   sorbent particles  and water droplets
   from a spray nozzle,  and (2) the
   capture of S02 by the wetted sorbent
   particles. Interaction parameters  con-
   sidered  included  particle and droplet
   sizes, relative  velocities,  proximity to
   saturation  residence  time, and Ca/S
   ratio.

Procedures and Results
  Mechanisms   of  post-furnace
humidification. The  first  of the pilot-scale
investigations  listed  above  was
concerned  with how flue gas should  be
humidified to achieve effective capture of
S02  in  a  post-furnace duct.  Three
humidification  procedures  were con-
sidered: (1) addition of water vapor alone,
with separate  cooling  of  the gas  by
conductive or convective processes;  (2)
addition of a water spray,  resulting in
simultaneous humidification  and cooling
of the gas by evaporation; (3) reinjection
of sorbent in an aqueous slurry.  The third
procedure would provide chemical reacti-
vation  of  the sorbent  —  conversion of
CaO to the more reactive Ca(OH)2 —  and
a humidified and cooled reaction environ-
ment. To  evaluate  these  procedures, a
large batch of  partially sulfated calcine
was generated by  burning  S02-doped
natural  gas in  the  combustor,  injecting
Ca(OH)2 in the furnace, and collecting the
resulting solid on a fabric  filter  at  the
point of gas discharge to the  atmosphere.
The solid was then  reinjected either as a
powder in  a low-temperature duct that
was humidified with water vapor of spray,
or in a slurry. Reaction conditions were
adjusted to provide reaction  with  S02 at
the  same temperature, humidity  level,
and approach to saturation.
  The molar extent  of conversion of CaO
to CaSO4  prior to reinjection of  the solid,
was 18%. The  molar increments in con-
version  produced by further,  low-
temperature reaction with S02 were:

                        Increment in
  Humidification method    conversion, %

 Cooling to enhance the
 effect of water vapor
 already present                1

 Cooling and
 humidification with a
 water spray                   6
 Slurrying with liquid water
 and soray injection
33
  Very  little  reaction  of  S02  with  the
calcine,  producing  an  increment  in
conversion of only 1%, occurred when
water vapor already present was cooled.
A sixfold  increase in  reaction, giving  an
increment  of  6% occurred when the dry
             solid was subjected to  the  effects of
             water spray. An increment in  excess
             30% occurred  with slurry injection. Cor
             parison  of the  results obtained with tl
             water  spray   and  the  slurry  sprs
             indicates  that the extent  of  lov
             temperature reaction was controlled  t
             the fraction of  sorbent particles actual
             colliding with water droplets  and becor
             ing physically  wet. The added increme
             due to the  water spray  (5%) divided  t
             the increment   due to  the slurry  sprj
             (33%) indicates a  collision efficiency
             15% (5/33 x 100 = 15).
               Further experiments  were  conducte
             with  fresh Ca(OH)2 injected in the lo\
             temperature duct in place of the partial
             sulfated  calcine.  These experimen
             yielded:
                                     Increment in
               Humidification method    conversion, %

              Cooling to enhance the
              effect of water vapor
              already present               9.6

              Cooling and
              humidification with a
              water spray                   12

              Slurrying with liquid water
              and soray injection            28
  The  results  with  the hydrated  lirr
differ  from  those  with  the  partial
sulfated calcine primarily in terms of tr
effects produced by cooling alone and t
the combination of spray humidificatic
and  cooling.  Cooling alone  gave
utilization  of  nearly  10%  with  tr
hydrated lime but an  increment of ju
1% with the calcine.  The presence
water spray, on the other hand, increase
the utilization of hydrated lime only fro
10 to  12% but increased  that of tf
calcine  from 1 to 6%. Slurrying  ga\
about the same result with both sorbenl
a utilization of about  30% totally with tt
hydrated lime or about 30%  incremei
tally with the calcine.
  The  primary conclusion  from th
investigation, insofar as the  Edgewat
demonstration is concerned, is that wat
vapor alone is not effective for activatir
the removal of  S02  by partially sulfate
calcine at low temperature. Humidificatic
of the sorbent with a water spray is  mo
effective, but even  its effectiveness
limited by  the  infrequency of  collisioi
between the sorbent particles and wat
droplets. The rapid and extensive rate
reaction of sorbent in a slurry, even aft

-------
very brief contact  between sorbent and
 vater, reveals that  the key to successful
numidification is increasing the frequency
of sorbent-droplet collisions.
  Water vapor alone is more effective for
activating  hydrated lime than for activa-
ting  partially sulfated calcine.  Clearly,
however, an increased frequency of colli-
sions between Ca(OH)2  particles  and
water droplets would  be even more
effective.
  Charge-augmented sorbent humidifica-
tion. In studies of CASH, the usual proce-
dure was to charge  sorbent  particles
negatively and water droplets positively.
The  sorbent was injected  in the furnace
as Ca(OH)2 with either gas- or coal-firing;
it was subsequently charged  on passing
through  a low-temperature  duct.  Two
types of particle chargers were investi-
gated: a disc-type  ionizer  or  a  wire-pipe
ionizer. Water droplets were dispersed in
the gas  stream  downstream from  the
particle  charger; the spray  nozzle  was
maintained at  a high positive  potential
and thus the droplets were charged at the
time of dispersion.
  A substantial part of the effort devoted
to CASH  involved equipment modifica-
tions. For example, the  disc-type ionizer
first  used to charge sorbent  particle was
limited by sparking at  low currents; the
 vire-pipe  ionizer  was developed  to
overcome these shortcomings.  Further-
more, considerable effort was required to
overcome current  leakage in  the spray
unit and ensure adequate charging of the
water droplets. Measured charge  den-
sities were m the range of 2 to 9 ^C/g for
the sorbent particles and 0.2 to 2 iiC/g for
the water droplets.
  The  usual procedure  for  determining
the effect of CASH on SO2 removal was
to operate the sorbent and water injection
apparatus for  20-30 minutes and  then
activate the two charging devices for  a
comparable  period. The SO2 concentra-
tion at the exit of the humidification  zone
was  monitored continuously  during  both
modes of operation. The effect of CASH
was expected to appear as an increase in
the slope of SO2 versus time;  generally
speaking,  however, the  effects, if  any,
were miniscule. Additionally,  solids  were
occasionally collected with and without
the charging devices  in operation,  and
the  samples were then analyzed  to
determine the extent of reaction. Compar-
ison of S/Ca ratios  in the solids indicated
that no  difference was  produced by
charging.
  In summary, no favorable effect  of
 1ASH on  sorbent utilization was evident.
  he  reasons for  this disappointing result
are not clear.  A possible  explanation is
that  the  attractive  force  between
oppisitely  charged sorbent particles and
water droplets is small; it decreases with
the square of the separating distance and
is  small in comparison with  other elec-
trical  forces except when  the separating
distance is of the order of the particle-
dimensions. The  other electrical  forces
arise  from the  lack of precise  electrical
balance between  positive and  negative
charges,  thus  creating  a net  "space
charge" field. In the absence of zero as a
net charge in the cloud of particles and
droplets, an electrical force  operates  on
each, bringing  about a separation and
deposition on the container walls.
  Adsorption isotherms of water vapor.
Adsorption isotherms of water  vapor  on
Ca(OH)2 prepared with and without addi-
tives  were determined at 60°C  using a
commercial  surface-area analyzer, which
gives the  absolute pressure  of  water
vapor in equilibrium with a  solid. The data
were  displayed in  plots of a) the weight
ratio of water absorbed to  solid  substrate
(Ag/g) versus b) the relative partial pres-
sure of water vapor (p/p°,  the fraction of
the saturation value). The additives inves-
tigated  were LiCI,  NaCI,  Na2CO3, and
CaCI2, mole/mole  for the 1:1 salts and
0.05 mole/mole for the 2:1  salts. Three of
the salts (NaCI, Na2C03, and CaCI2) have
received  significant  attention  as deli-
quescent additives in Ca(OH)2 to achieve
enhanced  S02 capture; the  fourth salt
(LiCI) has not  been investigated  in this
sense but it,  like CaCI2, is more strongly
deliquescent material in  the pure state
than either NaCI or Na2C03.
  LiCI gave evidence  of  behaving as a
deliquescent  additive,  as  expected,
throughout the range humidity  values
investigated (p/po  = 0.2-0.9). CaCI2,  on
the other  hand,  appeared  to become
deliquescent only  at a  much  higher
humidity  threshold  than  expected (p/p°
 =  0.6-0.7).  Evidently, the  deliquescence
of  CaCI2 in the pure state is inhibited  by
the presence of Ca(OH)2, which converts
the normal chloride  to a  basic chloride.
NaCI, as  expected, exhibited  delique-
scence only in a range of high humidities
(p/p°  >  0.75). Na2CO3  also exhibited
weak deliquescence, which suggests  in-
complete  transformation to  highly deli-
quescent  NaOH  through  the  double-
decomposition reaction: Ca(OH)2  +
Na2C03 -» CaC03 + 2NaOH
  Deliquescence  is clearly not the only
mechanism by  which so-called  deliques-
cent additives enhance SO2  capture  by
Ca(OH)2.  None of the three compounds
found beneficial — NaCL, Na2CO3, and
CaCI2 — provides significantly enhanced
water-vapor  pickup  by  Ca(OH)2   at
humidity levels where they do  provide
significantly enhanced S02 pickup.  Deli-
quescence by some additives  may  be a
contributing factor in SO2 capture, but
some other unidentified mechanism  must
be important also and must be dominant
for these three salts.
  Mathematical  modeling.  A  mathe-
matical  model  was  developed  to  help
understand  the complex  set  of mech-
anisms involved in low-temperature
desulfurization processes.  The  "first-gen-
eration" model  treats the process in two
steps: (1)  the "activation"  of the sorbent
is  modeled in  terms of  the  collisions
between  sorbent particles and water
droplets; and (2)  the capture of  S02 by
the activated sorbent is modeled  in terms
of gas-phase  diffusion  of SO2  by the
sorbent/droplet  ensemble, chemically
enhanced  absorption  into  the liquid
phase, and liquid-phase mass  transfer of
the reacting species.
  The results of the modeling study show
that,  when lime and water are  injected
separately, the S02 removal efficiency is
governed  by the  degree of  sorbent
wetting  by  collisions between  sorbent
particles  and  water  droplets.  The
sorbent/droplet  collision  rate is  max-
imized by making the droplet size and
velocity as large as possible. In actual
practice, however, the droplet size  must
be well below the optimum value  in order
to ensure complete evaporation   and
avoid wetting of  the duct walls. If the
humidification process could be modified
to allow  larger  droplets, it  would  be
possible to significantly improve on the
sulfur capture achieved.
  The  sulfur capture is strongly   influ-
enced  by  approach to  saturation,  Ca/S
ratio,  residence time, and droplet  size
and velocity. To maximize S02 removal,
the approach  should  be  as close  as
possible.  However,  it may  be  difficult to
consistently maintain dry duct  walls at
approaches of less than 11°C.  Increasing
the sorbent injection  rate, or Ca/S ratio, is
also  beneficial, but  reaches  a point of
diminishing  returns due  to   increased
operating  cost  and  particulate loadings.
The effect of the latter on  ESP perform-
ance must also be considered.
  Increasing the residence time is bene-
ficial only  up to  the point of complete
droplet evaporation. Conversely, the resi-
dence time must  be at least sufficient to
allow  complete  evaporation.  For  more
retrofit situations, the residence time will
be fixed. This means that the droplet size
must be selected to  make maximum use
of the available  time  without allowing
droplets to reach the  ESP. The model
can  be used to predict the optimum

-------
droplet sizing for a given duct residence
time.

Conclusions and Plans for
Further Work
   The  significance  of the  principal
findings are  discussed above in the
concluding  paragraphs  on each of the
four  experimental  tasks. Conclusions,
stated more succinctly, are:
•  Insofar as enhancement of S02 cap-
   ture at the Edgewater demonstration
   of LIMB  is concerned, humidification
   by water spray will  be  required.  A
   spray of pure water  would suffice if
   the frequency of collisions  between
   sorbent  particles and water droplets
   could be  increased.  Otherwise,
   isolation  of partially  sulfated sorbent
   followed by  rehydration and  rein-
   jection in a spray will  be necessary.
•  Charge-augmented  sorbent humid-
   ification (CASH)  - entailing the charg-
   ing of sorbent  particles and  water
   droplets  negatively  and  positively,
   respectively  - appears  unlikely to
   enhance sorbent reaction with  S02
   under post-furnace conditions.
        •  So-called deliquescent additives that
           are  known to enhance  S02 capture
           by Ca(OH)2  under low-temperature
           conditions clearly must operate to
           some  extent  by  mechanism other
           than deliquescence.

        •  The mathematical  model  gives an
           improved appreciation of factors that
           are  crrtical   to  S02  capture  by
           Ca(OH)2 when the sorbent particles
           are  subject to wetting.  The model
           gives an  improved rational basis for
           upgrading  the performance of  duct
           processes for  S02 removal.

           During  the  forthcoming Edgewater
        demonstration  of LIMB, with Ca(OH)2
        injected in the furnace, the flue gas will
        be  humidified   before  it  enters  the
        electrostatic precipitator. This will  be
        done to treat the electrical resistivity of
        the mixture  of sorbent-ash solids  and
        improve the efficiency of  precipitation.
        Relatively small  increases  in  SO2
        removal are expected, with the  size of
        the increases depending  primarily  on
        the droplet/sorbent  collision  efficiency.
        In  a pilot-scale adjunct to  the full-scale
        Edgewater operation, humidification will
be  tested in conjunction with the low-
temperature duct injection of a modified
calcium-based solid. This low-temper-
ature  process, developed by EPA,  is
known as ADVACATE;  it  employs the
product of hydrating LIMB sorbent-ash
mixtures under pressure and generating
a more reactive calcium-based sorbent.
   Subsequently, the  Department  of
Energy will use the Edgewater facilities
to demonstrate a commercial process
based on separate duct injections  of
Ca(OH)2 and water. Meanwhile, DOE will
be  sponsoring a  broad research  pro-
gram  on  generic duct-injection  pro-
cesses, ranging from laboratory-through
pilot-scale studies with ultimate full-scale
applications  in mind.
   A  specific need  during  the further
research is  elucidation  of the  role  of
additives  in the  category  known as
deliquescent materials,  which may  or
may not be  actually deliquescent under
the conditions of use,  as this  report
shows. Further  studies  are  needed  to
elucidate the mechanisms of action so
that  both  the compounds and  the
conditions of use may be selected more
intelligently.
 J.  P. Gooch, R. Be/He/, £. 8. Dismukes, and R.  S.  Dahlin   are with Southern
  Research Institute, Birmingham, AL 35255.
 Louis S. Hovis is the EPA Project Officer (see below).
 The complete report, entitled "Humidification of Flue Gas to Augment SO2 Capture
  by Dry Sorbents," (Order No. PB 89-169  841/AS;  Cost:  $15.95, subject to
  change) will be available only from:
         National Technical  Information Service
         5285 Port Royal Road
         Springfield, VA 22161
         Telephone: 703-487-4650
 The EPA Project Officer can be contacted at:
         Air and Energy Engineering Research Laboratory
         U.S. Environmental Protection Agency
         Research Triangle  Park,  NC 27711
 United States
 Environmental Protection
 Agency
Center for Environmental Research
Information
Cincinnati OH 45268
 Official Business
 Penalty for Private Use $300

 EPA/600/S7-89/001
         000085833    ||TEC«011  »GE»Cl
          CHICAGO

-------