United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-81-103 Aug. 1981
Project Summary
Disposal  of  Flue  Gas
Desulfurization  Wastes:
EPA  Shawnee  Field
Evaluation,  Final  Report

P. R. Hurt, P. P. Leo, J. Rossoff, and J. R. Witz
  This report summarize* the results
of the Flue Gas Desulfurization (FGD)
Waste Disposal Field Evaluation Pro-
ject sponsored by the U.S. Environ-
mental Protection Agency at the
Tennessee Valley  Authority (TVA)
Shawnee Steam Plant, Paducah,
Kentucky. This pilot scale project,
which was initiated in 1974 and com-
pleted in September 1980, evaluated
methods and  costs for disposing of
wastes produced from wet, nonregen-
erable scrubbing of  sulfur dioxide
from coal-fired utility boiler flue gases.
The environmental  effects of various
disposal techniques were studied.
including evaluations of untreated,
chemically treated, and oxidized
wastes utilizing lime or limestone
scrubber absorbents.
  Because  water quality and  land
reclamation are of  principal interest,
analyses of leechate, supernate, run-
off, and ground water were conducted
and physical properties of the wastes
were evaluated. No measurable effect
on the ground water quality at the
disposal site was detected during the
course of the program. Chemical
treatment and underdrainage of un-
treated waste yielded structurally
sound materials. Cost-effective and
environmentally sound disposal meth-
ods for FGD waste and slurried gypsum
appear to be ponding with underdrain-
age and cKemfealtreatment/landfilling
of the FGD waste. Disposal costs for
these methods range from about 0.8
to 1.5 miHs/kWh. based on a high-
sulfur (e.g., 3%) coal application.
  This Project Summary was devel-
oped by EPA's Industrial Environmen-
tal Research Laborabory, 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
  This report summarizes activities and
results of the EPA Shawnee Flue Gas
Desulfurization (FGD) Field Disposal
Evaluation Project. This project,  initi-
ated in September 1974, was designed
to evaluate various ponding and landfill
alternatives and to develop cost esti-
mates for the  disposal of  by-products
from wet, nonregenerable scrubbing of
sulfur dioxide (SO2) from coal-fired
utility boiler flue gases. The environ-
mental effects of various disposal tech-
niques,  as well as scrubber reagent,
operations, weather, and field operating
procedures, were assessed with the
goal of determining environmentally
sound disposal methods. Water quality
and land reclamation are  of principal
interest and, to that end, periodic samp-
ling, analyses,  and assessments were
conducted of leachate, supernate, run-
off, ground water, and soil and waste

-------
cores. The Aerospace Corporation was
responsible for the project planning,
coordination, selected water and solids
analysis, assessment and evaluation of
the various disposal  methods, and
reporting. Site construction, mainte-
nance,  coring, water sampling, and
routine water analyses were performed
byTVA. The technical effort described in
this report was completed in September
1980.
  Ten field sites were evaluated in this
project, which is located  at the TVA
Shawnee Steam Plant near Paducah,
Kentucky. Of the ten sites, eight ranged
in size up to 560 m2 (6000 ft2) with  a
waste depth of 0.9 m (3 ft) to 1.2 m (4 ft),
while two are  surface disposal sites at
about 190 m2 (2000ft2). Waste materials
for this  project were produced by two
scrubber systems. Using lime or lime-
stone slurries as the  S02 absorbent,
each of the scrubbers, a UOP, Inc.,
turbulent contact absorber (TCA), and a
Chemico, Inc., venturi spray tower
(VST), treated a flue gas slipstream
equivalent to 10 MW from one of the
boilers. Waste  produced by both scrub-
bers was evaluated. The disposal tech-
niques consisted of the following: (a)
untreated waste  ponding, (b) chemical
treatment and  landfilling, (c) chemical
treatment and ponding, (d) calcium
sulfite oxidation to gypsum and subser
quent disposal, and (e) untreated waste
ponding with underdrainage. Figure  1
illustrates the  relationship  of the FGD
waste to the disposal alternatives that
were evaluated at Shawnee. This project
has provided a broad data base for the
evaluation of the control of SOz scrubber
wastes by combining analyses of results
from field disposal operation and labo-
rgtory tests.

Summary
  The data obtained in the Shawnee
FGD waste disposal evaluation are
summarized and discussed in this re-
port. The physical characteristics of the
wastes were obtained by analysis of the
materials prior to disposal  and of core
samples after disposal.  Water samples
from each site were analyzed (i.e.,
leachate or  underdrain, runoff, super-
nate, and ground water) for the evalua-
tion of the chemical  characteristics.


Site Description
  A description of the disposal sites is
summarized in Table 1. A view of the
disposal site is shown in Figure ? Seven
ponds or surface sites contained un-
                                                      Landfill
     Chemically
       Treated
       Waste
      Untreated
        Waste
     Force-Oxidized
        Waste
 Figure 1.    FGD waste disposal alternatives..
Figure 2.     Overview of disposal sites with Shawnee Steam Plant in Background. ,

-------
 Table 1.    Shawnee Disposal Sites
           Fill       Scrubber  	
 Site     Date         Type*    Absorbent
                                          Waste
                                        Source
                                       Solids
                                      Content,
                                       wt%c
                                         Treatment
                                                Remarks
  A    10/8/74

  Al   5/10/76
   B

   C
4/15/75

4/23/75
  D   2/5/75

  E   12/7/74

  F   2/3/77


  G   10/5/76

  H'  9/2/77

  H9  9/30/77


  J   12/31/78

  K   3/29/79
VST

VST


TCA

VST


TCA

TCA

TCA
Lime

Lime


Limestone

Lime


Limestone

Limestone

Limestone
F

F


CU

CE


CU

CU

CU
                VST     Lime             CE

                VST     Limestone        CU

                VST     Limestone        F


                VST     Limestone        F

                VST     Limestone        F
46      Untreated    Out of service 4/15/76

46      Untreated    Control pond, transferred from
                      Site A

38"      Dravo        Underwater disposal

55"      IUCS        Site converted to runoff mode
                      3/79

38      Untreated    Control pond

38"      Chemfix      Covered 11/77

47      Untreated    Underdrained site, covered
                      11/77

47      Untreated    Underdrained site, covered 4/80

33      Untreated1

86      Untreated1    Surface site: unreacted
                      limestone, 13% by dry weight

81       Untreated*    Surface site: adipic acid additive

80.      Untreated*    Surface site: unreacted
                      limestone, 25% by dry weight
 'Venturi and spray tower (VST);   bulent contact absorber (TCA).
 "Fitter (F), clarifier underflow (C /. and centrifuge (CE).
 "Site H is ash-free. All others: fly ash is approximately 40 wt% of solids content. In Site G, half of fly ash was mixed with waste;
  the remaining portion was placed in six equally spaced layers.
 "Prior to chemical treatment.
 'Lower (below grade) portion.
 'Upper (above grade) portion.
 9Force-oxidized to gypsum.
treated waste, and three were filled
with chemically treated material. The
untreated sites include Al, D, F, G, H, J,
and K. Sites Al and D were considered
control  sites with  lime and  limestone
(absorbent) waste disposed in ponds
containing indigenous soil. Sites F and
G were Underdrained ponds containing
limestone and lime scrubbing wastes.
Sites H, J, and K contained wastes that
were force-oxidized to gypsum; Site H
had a lower ponded portion  of under-
drained clarifier  underflow slurry and
an above-ground stack of gypsum filter
cake; and Sites  J  and  K had gypsum
filter cake stacked on the surface. A total
of 14 ground water wells were situated
in and around the disposal  area  and
were sampled periodically to monitor
any potential effects of  the disposal
sites on the ground water quality.

  Sites  B,  C, and  E were filled with
chemically treated material. Site B was
filled in April 1975, with clarifier under-
                                 flow (limestone absorbent), treated by
                                 the Dravo Corporation. Site C (Figure 3),
                                 also filled in April 1975, contained
                                 centrifuge cake (lime absorbent), treated
                                 by IU Conversion Systems, Inc. Site E
                                 was filled in December 1974 with
                                 clarifier underflow (limestone absorb-
                                 ent), treated by Chemfix, Inc.
                                   Several of the sites were modified
                                 since  initial  placement  (Table  1)  to
                                 evaluate site retirement and reclamation
                                 methods.  Chemically treated Site E and
                                 Underdrained Sites F and G were covered
                                 with clay, which was then contoured
                                 and planted with grass. Site C was con-
                                 verted  to a  runoff configuration in
                                 March  1979. As  a separate EPA/TVA
                                 project, the crown of the clay covering of
                                 Sites E and F was removed, and several
                                 species of small trees were planted in
                                 the spring of  1979.

                                 Waste Characteristics
                                  The disposal sites were filled with
                                                        FGD waste representing a cross section
                                                        of scrubber effluent conditions. The dis-
                                                        cussion of the various methods of waste
                                                        disposal is divided  into the following
                                                        categories: untreated wastes, chemically
                                                        treated wastes, gypsum (oxidized calcium
                                                        sulfite) filter cake, and untreated wastes
                                                        with underdrainage. Specific properties
                                                        related to those wastes and methods of
                                                        treatment are included.

                                                          The chemical composition of waste
                                                        input liquor, water from site  runoff,
                                                        supernate,  leachate/underdrainage,
                                                        and ground water from 14 wells in and
                                                        around the disposal  area was analyzed
                                                        for a variety of chemical species. The
                                                        composition  of  the waste input liquor
                                                        (before treatment, if any) is summarized
                                                        in Table 2 and shows a wide range in the
                                                        concentration of chemical species for
                                                        the different FGD wastes.
                                                          The physical properties considered in
                                                        the  disposal  of FGD  waste include
                                                        viscosity, bulk density, moisture content.

-------
Table 2.    Chemical Analysis of Disposal Site Input Liquor
Concentration (mg/ll
Total
Dis-

Site'
A
B
C
D
E
F
G
W
W
J
K

pH
8.3
8.9
8.9
9.2
9.4
12.2
7.8
7.1
(b>
5.8
6.9

Calcium
2,100
1.060
2.720
1.880
1.880
1.990
150
1,300
1.510
1.250
550

Sulfate
1.525
1.875
1.575
1.500
1,400
1.100
6.600
1.930
1.875
1.438
2,250

Chloride
4,600
1.850
4.700
2.950
2.700
2.000
3.600
3.500
6.600
3.500
2,450

Sulfite
4
3
45
56
32
Ib)
w

17
46
56
41
70
12
62
116
107
68
Sele-
nium
0.005
0.020
0.018
0.014
0.014
0.042
0.63
0.14
(b)
0.008
0.035

Mercury

-------
 Table 3.    Solids Analysis of Disposal Site Untreated Input Wastes
Site*
A
Bc
Cc
D
Ec
F
G
Ha
rf
J
K
Solids
Content
(% By Weight)
46
38
55
38
38
47
47
33
86
81
80
Percent Solids by Dry Weight
Calcium
Sulfite
(b)
30.3
38.8
29.4
27.9
33.3
W
0.2
0.1
0.1
0.1
Calcium
Sulfate
(b)
10.7
9.7
10.9
9.4
14.8
(b)
98.0
86.8
63.7
47.1
Calcium
Carbonate
(b)
14.2
0
18.6
18.3
8.3
W
1.8
13.1
1.2
25.0
Fly
Ash
(b)
35
40
33
34
40
20
0
0
35
28
 "Table 1 lists type of waste at each site
 "Not analyzed
 "Subsequently treated chemically
 "Lower (below grade} port/on
 "Upper (above grade) portion
   6000
           1975      1976       1977      1978
                                       Year
                 1979
1980
 Figure 4.    Concentration of TDS and major species in Site D leachate.
minor species have shown only slight
reductions after 5 years of seepage. For
nearly 900 untreated leachate analyses,
4 showed the concentration of minor
species to be greater than 10 times the
National Interim Primary Drinking Water
Regulations  (12,  14, and 20 times
greater for arsenic and 11 times greater
for selenium).  Under the new Hazardous
Waste Management System, all FGD
wastes are temporarily exempt from the
Federal regulations covering hazardous
materials. However, they are subject to
state and local regulations; these may
include Federal criteria for nonhazardous
wastes.
  The dewatering characteristics of
FGD wastes are important to the various
disposal techniques in that they affect
the volume of the disposal basin, the
waste handling methods, and the condi-
tion of the wastes in their final disposal
state. The effectiveness of the dewater-
ing methods used and the ability of FGD
waste to be dewatered are a function of
a number of solids characteristics,
including the size and distribution of
particles and the crystalline structure of
the particles, which are a function of the
scrubber system and its operating pa-
rameters. In laboratory dewatering
tests, four methods were evaluated,
using FGD wastes from Shawnee: set-
tling and decanting, settling by free
drainage, vacuum filtration, and centrif-
ugation  (Table 4).
  The highest  density  was  obtained
principally by vacuum-assisted filtration.
However, the density differences be-
tween filtration and centrifugation were
relatively small.  For most wastes, set-
tling by free drainage yields a slightly
greater  density than by settling  only.
Table 4 shows that the wet-bulk densities
of limestone scrubbing wastes ranged
from a low of approximately 1.45 g/cm3
for settled wastes to a high of  1.65
g/cm3 for vacuum filtered. Values for
lime scrubbing  waste were approxi-
mately 7% less than those for limestone
scrubbing wastes.
  Confined load-bearing strengths of
untreated FGD wastes  (including gyp-
sum) as a function of solids, absorbent,
and fly  ash content are presented in
Figure 6. Among other  considerations,
the data highlight the criticality of solids
content  on the load-bearing strength of
untreated wastes. This is of particular
importance when considering the dis-
posal  of slurried gypsum or untreated
FGD wastes that are underdrained,
because these types of disposal depend
on  dewatering  to achieve  material
bearing  strength. The test data indicate
that these wastes may be dewatered to
a narrow range of solids content, above
which the load-bearing strengths in-
crease rapidly to values well above the
minimum for safe access of personnel
and equipment. In addition, the critical
concentration appears to be unique for
each type of waste tested. Figure 6 also
illustrates the effect of the absorbent
and fly  ash on dewatering and load-
bearing  characteristics.  Limestone
scrubbing wastes have critical solids
content higher  than lime scrubbing
wastes.  Likewise, the presence of fly
ash enhances dewatering in both types
of wastes; however, for any specific
solids content of a given FGD waste, the

-------
     100
  0.0001
            1975       1976       1977      1978

                                        year
                1979
1980
 Figure 5.    Concentration of minor species in Site D leachate.
load-bearing strength is less when fly
ash is present.
  The pollution potential of waste liquor
seeping into ground waters is governed
by the mobility of leaching waters; this
mobility is limited by the coefficient of
permeability of the various  media
through which this leachate must pass.
The permeability coefficient of untreated
wastes (lime and limestone) containing
fly ash is approximately 2 x 10~4 cm/sec,
which is comparable to typical values
for silty sand (1CT4 cm/sec).
  The viscosity of FGD waste is indica-
tive of its pumpability, which could
affect both the mode and cost of trans-
port.  The results of  viscosity tests
conducted in the laboratory for various
untreated wastes define dewatering
limits for certain waste materials if they
are to be pumped. The tests also show
that pumpable mixtures (considered to
be less than 2 pascal-sec (20 poise) for
slurries) range from a high solids content
of 55 wt% to a low of 30 wt% (Figure 7),
depending on the absorbent and ash
content. For example,  the viscosity of
the limestone scrubbing at a solids
content of 50% is less than lime scrub-
bing waste at the same solids content.
Table 4.    Bulk Densities of Untreated FGD Wastes
The results also show that FGD wastes
of a given type are less viscous if they
contain fly ash.

   Disposal of untreated  material in a
pond is usually the least costly methoc
of FGD waste disposal.  If the pond does
not have a base material  considered tc
be impermeable, a liner must be added
to prevent seepage, thereby increasing
the disposal costs significantly. Clay or
polymeric liners  may be  placed at the
base and on the slopes of such ponds.
Any pond continually exposed to
weathering, however,  may eventually
be subject  to a  degree of seepage
because liners are not completely im-
permeable and, in the case of polymeric
membranes, may be  subject to varying
degrees of physical damage and deteri-
oration. Because FGD wastes are thix-
otropic in nature, ponds  of untreated
wastes are nonstructural  and are gen-
erally difficult to reclaim, except possibly
in areas of low rainfall and high evapo-
ration.

Chemically Treated Waste
Disposal
  Sites B, C, and E contain wastes that
were chemically  treated by the Dravo
Corporation IU Conversion Systems,
Inc., and Chemfix, Inc., respectively,
during late 1974 and  early 1975. Site B
simulates a disposal condition in which
the waste cures underwater and remains
underwater, except for periods of ex-
tended drought. Site  C initially repre-
sented a depression in  a landfill in
which  rainwater  collected, and Site E
represents a landfill that traps rainwater
that collects in a sump at its lower end.
Site B  has remained  as originally con-
figured; Site C was converted in March
1979  to a runoff configuration; and in
November 1977 Site E was covered
Shawnee
Source and
Sampling |
Date
Limestone
1 Feb 73
15 Jun 74
Lime
19 Mar 74
8 Sep 76
8 Sept 76
Fly Ash
«>/o by Dry
Weight)
20
40
40
40
~0
Settling and
Decanting
Solids Density
Why
Weight) (g/cm3)
49
53
42
45
47
1.45
1.46
1.34
1.34
1.37
Dewatering Method*
Settling by Free Centrifugation
Drainage
Solids Density Solids Density
(%by (%by
Weight) (g/cm3) Weight) (g/cm3)
56
58
43
58
51
1.51
1.53
1.36
1.50
1.41
60
63
50
53
48
1.56
1.60
1.44
1.44
1.38
Vacuum
Solids
(%by
Weight)
65
66
56
61
57
Filtration
Density
(g/cm3)
1.65
1.64
1.51
1.54
1.49
 *Using Laboratory Equipment

-------
    250
 I200
 I
 I
g  750
 
-------
  6000
            1975      1976       1977       1978      1979      1980
Figure 8.    Concentration of TDS and major species Site B in leachate.
   6000
   5000
   4000

5 3000
                                                       Converted
                                                       to a Runoff
                                                       Configuration
             1975
Figure 9.    Concentration of TDS and major species Site C in leachate.

                                 8
1980
material, increasing to a depth of 0.15 r
(6 in.) at present. These material
exhibit very low bearing capacities.
  Laboratory analyses of core sample
from these three chemically treate
FGD waste sites show widely vary in
coefficients of permeability as a result c
cracks  that were present  in some cor
samples (Table 6). In general, the coeff
cient of permeability for the wastes i
Sites C and E is approximately the sam
and is  somewhat less permeable tha
the Site B material. Although all mate
rials have a permeability of about 10"
cm/sec, which is at least one order c
magnitude less than the untreate
wastes, several samples  from Sites I
and E showed coefficients in the rang
of 10~* to  10"7 cm/sec. There does nc
appear to be a time-dependent trend ii
the permeability of core samples of th
waste.
  Other physical parameters includin
solids content, void fraction, and uncon
fined compressive strength are given ii
Table 6. Typical solids content  for th
cores from the three sites  were 45% fo
Site B, 61 % for Site C, and 52% for Sit
E. The average void fractions were 0.75
0.66, and 0.70 for samples from Sites B
C, and  E, respectively. The wide range o
values for the  unconfined compressivi
strength of free-standing samples o
these  materials may be  attributed ii
part to random cracks that  were presen
in some of the test  samples.
  Stabilization of FGDwastesbychemi
cal treatment  offers a solution to th<
disposal problem in that  the potentia
water pollution can be minimized aru
the site can be reclaimed. Chemica
treatment converts the waste into <
structural material, decreases its coeff i
cient of permeability and  seepage rate
relative to untreated wastes by one tc
three orders of magnitude, reduces the
initial  concentration of  soluble sal
constituents in the leachate by approx
imately 50%, is amenable to subgradc
or above-grade landfilling, and can be
contoured to  promote the runoff o
rainwater.
  A general procedure for managing
rainfall runoff  from a full-scale chemi
cally treated waste disposal site includes
collection  of the runoff in a periphera
ditch,  which  directs the water to E
settling pond. Depending on the quality
of water (e.g., concentration of TDS anc
total suspended solids (TSS) decantec
from this site), it can be discharged to s
stream or returned to the scrubbei
system. After closure of  the site, th^
waste is capped with soil and contourec

-------
   5000'
         0   20  40  60 80 100120140160180200220240260280300

                            M/eeAs After Pond Filling
Figure 10.     Concentration of TDS in chemically treated waste leachate (Sites B, C,
              andEJ.
    100
                                                       Converted
                                                       to a Runoff
                                                       Configuration
                                           1978      1979       1980
0.000 /11111111 111 111 111111 111 111 11111
          1975        1976
Figure 11.     Concentration of minor species in Site C leachate.
to support the growth of vegetation and
to minimize seepage of rainwater into
the waste.

Gypsum Filter Cake Disposal
  Sites J and K and the stacked portion
of Site H are limestone scrubbing FGD
wastes that were force-oxidized to
gypsum and filtered.^The evaluation of
the lower portion of Site H, containing
clairifier underflow slurried gypsum, is
discussed in the following section on
underdrained disposal). Input waste
characteristics are summarized in
Tables 1, 2, and 3. The composition of
filter cake samples represented a range
of limestone utilization of 97, 80, and
52% corresponding to  unreacted lime-
stone concentrations of 1, 13, and 25%
for  Sites J, H, and K, respectively.
Although the filter cake at Site K has a
relatively high unreacted limestone
content  of 25%,  it is considered to
approximate a portion of the gypsum
that will be produced by forced oxidation
at the TVA Widows Creek Steam Plant.
  These three sites were constructed to
determine the impact of piling or stack-
ing  FGD gypsum on the ground. Obser-
vations were made of the erosion,
runoff, water quality, and surface char-
acteristics of the filter cake and crust
bearing strength and strength loss
when moisture content is increased.
Sites H and J were generally conical in
shape and  approximately 10 ft high,
whereas Site K reflected the random
dumping of piles of gypsum over an area
of 150 m2 (1600 ft4).  Because of the
limited size of the piles, compacting of
the gypsum during stacking was not
performed.
  Chemical analyses were made of the
leachate collected in the underdrainage
system and of runoff samples from Site
H. The analysis of the underdrainage is
typical of an untreated  waste leachate,
with concentrations of major and minor
chemical species similar to those shown
earlier with an initial TDS value of 4000
mg/l, which is continuing to  decline
toward a saturated gypsum concentra-
tion. The concentration  of major species
in the runoff of Site H shows  a slight
decrease in TDS over time. The runoff
had TDS slightly in excess of 2000 mg/l
and TSS ranging between 4 and 300
mg/l, which  indicates the need to
control runoff in this type of disposal to
prevent seepage into underground
drinking water supplies or direct dis-
charge into streams. The minor species
(Figure 12) show a general decrease in
boron, lead, and mercury. However,

-------
 TableS.    Ultimate Bearing Capacity of FGD Wastes
                                                  Ultimate Bearing
                                                     Capacity
   Site
Waste Treatment
                                                 kPa
                                                                       psi
B
C
E
F
G
H
Chemical *
Chemical "
Chemical c
None (underdrained)
None (underdrained)
None (underdrained^
1035 to 2070
>2280
2070 to > 2280
410 to 520
1240 to 1650
930 to > 2280
150 to 300
>330
300 to >330
60 to 75
180 to 240
135 to > 330
  "Dravo Corporation
  b/t/ Conversion Systems, Inc.
  cChemfix Corporation
  aClarifier underflow
  Note:  Shawnee clay soil has an ultimate bearing capacity of 1650 to 2070 kPa.
 !
 c'
 .o
 **
 (0
 !
 1
    0.001
   0.0001
     0.01
                1977
                               7575            7373
                                       Year
Figure 12.    Concentration of minor species in Site H runoff.
                                             1980
 recent data indicate an increase in
 arsenic and selenium in the runoff
 samples. The most recent runoff sam-
 ples yielded values of arsenic approxi-
 mately 10 times the National Interim
 Primary Drinking  Water  Regulation
 standards and values of selenium nearly
 equal to those standards.
  At Sites H and J, the gypsum filter
 cake was stacked so that a natural slope
 occurred in a conical shape, producing a
 surface of  about 35 degrees to  the
 horizontal. These tests show that erosion
 would be a potential problem for FGD
 gypsum filter cake disposed of in this
 manner. For example, at Site H, after 18
 months of weathering, approximately
 20% of the mass flowed to the base. At
 Site J, which was stacked the following
 year, the same erosion trend developed.
  Field penetrometer readings were
 taken of ultimate bearing capacity on
 the gypsum stacks.  During periods of
 dry weather, a crust was observed with
 a thickness varying from approximately
 5 cm (2 in.) near the peak to 13 cm (5 in.)
 at midheight, and 28 cm (11  in.) at  the
 base. Bearing capacities of the  crust
 varied from 410 kPa (60 psi) near  the
 peak to 3100 kPa (450 psi) near the
 base. Several tests were made with a D-
 8 Caterpillar tractor at Site J to deter-
 mine the maximum angle at which  the
 vehicle could negotiate the slope of this
 material. During testing, the cleats of
 the tractor broke through the crust, and
 the vehicle lost traction at an angle of 17
 degrees to the horizontal. After approx-
 imately six passes over the same spot,
 the material became so moist that it
 could not support personnel.
  The crust also absorbed water and,
 during periods of continued rainfall,
 reverted to a material characteristic of
 the original FGD gypsum filter cake.
When water collected, the gypsum
 increased  in moisture content and
 approached a slurried condition.
Tabto 6.    Physical Characteristics of Impounded, Chemically Treated FGD Waste Core Samples

         Characteristics                     Site B                         Site C
                                                                                 Site E
   Solid Content (% by Weight)

   Unconfined Compressive
   Strength, Wet kPa (psi)

   Density (g/cm3):

    Wet

    Dry

    Void Fraction

   Permeability Coefficient (cm/s)
                         45

                     190 to 580
                     (28 to 84)
                        1.37

                        0.63

                        0.75

                2.1 x10~*to3.7x10~*
                                                                           61

                                                                      280 to 6900
                                                                       (40 to 996)
                                                                          1.52

                                                                          0.92

                                                                          0.63

                                                                  5.2 x 10~* to 3.2 x 10'7
                      52

                  190 to 1800
                  (24 to 260)
                      1.36

                      0.70

                      0.71

             1.1 x 10'* to 6.9 x 10-7
                                 10

-------
  Two sets of FGD gypsum samples that
were produced at the pi lot FG 0 scrubber
at the EPA Industrial  Environmental
Research Laboratory at Research Tri-
angle Park were obtained. After labora-
tory filtration, one set of gypsum samples
exhibited substantially higher uncon-
fined compress! ve strength, 410 kPa (60
psi), than the other set, 140 to 170 kPa
(20 to 25 psi), which contained about 5%
sulfite. The unconfined compressive
strength of  the samples with the 5%
calcium sulfite was approximately equal
to that of an FGD waste that is predomi-
nantly calcium sulfite.
  Typical wet bulk densities for gypsum
filter cake with 80% solids content were
determined to be in the range of 1.3 to
1.4 g/cm3. Wet bulk density of Shawnee
gypsum clarifier underflow that settled
in the impoundment beneath  the Site H
filter cake pile and was saturated with
moisture was approximately 1.7 g/cm3
  The oxidation of FGD wastes to gypsum
results in a waste material that is readily
dewatered by vacuum filtration to a
solids  content in the  range of 75 to 85
wt%. When stacked above grade, the
filter cake may crack under freeze-thaw
or wet-dry conditions, thereby allowing
rainwater to enter  into the material.
Additionally, it may erode when exposed
to rainfall and produce a runoff contain-
ing concentrations of dissolved solids of
approximately 2500 mg/l. These obser-
vations indicate that  special site main-
tenance may be required on  an opera-
tional scale to reconfigure the disposal
pile after weathering and to control the
runoff.  On the  basis of experience in
phospho-gypsum waste management
in the fertilizer industry, this is not
unexpected.

Underdrain Disposal of
Untreated Wastes
  Sites F and G and the  lower  portion of
gypsum Site H (Figure 13) were evalu-
ated for disposal of untreated FGD
waste  deposited  in an underdrained
impoundment. The underdrainage
system consisted of a 0.3-m (1-ft) thick
sand layer at the base of each  impound-
ment,  under which perforated plastic
pipes, imbedded in pea gravel, collected
and  fed leachate that seeped through
the waste to a sump, through a gravity
drain system. Water was pumped from
the sump to  remove all  seepage.
  In a  properly designed operational
site, the underdrainage (and all rainfall
on the disposal site) could be recycled to
the scrubber. This would effectively
Figure 13.    Filling lower portion of Site H with scrubber clarifier underflow that
               has been force-oxidized to gypsum.
reduce the normal amount of fresh
makeup water. However, it would in-
crease the concentration of soluble
salts (e.g., chloride) in the scrubber loop.
The resultant chloride concentration
could  be as high as 9000 mg/l for
lime/limestone scrubber waste and
25,000 mg/l for gypsum waste. These
concentrations are not anticipated to be
detrimental to scrubber operation;
however, steps to prevent scrubber
corrosion would be necessary.
  The  underdrainage enhances de-
watering of the material such that the
additional settling that results from the
removal of the occluded water results in
an  untreated material with very high
bearing capacities when  contained in
an impoundment (Table 5); these ponds
drain rapidly because of their highly
pocous base. In the Shawnee testing,
the underdrained material could support
personnel during the pouring operation.
Sites G (untreated time scrubbing waste)
and H  (gypsum) supported wheeled
vehicles within a  day after placement
(Figure 14). Site F  (untreated limestone
scrubbing waste), the weakest of the
three, did not drain as rapidly as the
others but still attained high bearing
strength within 1 week after filling.
  Use of an underdrainage system at
the disposal site allows for the collection
of all seepage for return to the scrubber
system, thus, maintaining control of
leachat* during the full  period. Recla-
mation of this type of disposal method
would require an earthen cap that is
contoured and  maintained to shed
water after disposal site closure. After
Site F had been in service for 9 months,
all water was removed from the under-
drain  sump, and the site was covered
with a layer of indigenous soil, using a
rubber-tired earth mover. This cap,
which was contoured and planted with
grass so that its surface would shed
rainwater, was approximately 0.9 m (3
ft) thick at the center line and 0.6 m (2ft)
thick at the edges. In the spring of 1979,
the soil cap was recontoured and re-
duced to a constant thickness of 0.6 m (2
ft) for a separate EPA/TVA project that
included an experimental planting of
young trees. Recently, Site F (and
chemically treated  Site E) has shown
increased seepage. This may be  due,
however, to the recontouring of the soil
covering and puncturing of the soil cap
for the trees. Underdrained Site G was
covered in  April of 1980, but (because of
low rainfall since construction) the
evaluation of this type of reclamation is
inconclusive.
  Regarding seepage into the subsoil
from  an underdrain system, a low
hydrostatic pressure exists at the site
and subsoil interface because the hy-

-------
draulic head is interrupted at the porous
base and the system is vented to the
atmosphere. If, for example, the subsoil
coefficient of permeability were 10~6
cm/sec, the penetration with under-
drainage would be 0.36 m/yr (14 in./yr)
only in the vicinity of the trenches
holding the drainage pipes, which may
be approximately 0.25 x 0.25 m (10 in. x
10 in.) and as. much as 30  m (100  ft)
apart. Penetration of the water into the
soil between the drainage trenches
would be negligible because the water
would be only a  film on the base of the
site. By comparison, the subsoil seepage
from a similar pond without underdrains
would be about  1.8 m/yr (6 ft/yr) from
the entire pond bottom. Considering the
depth of  seepage and the pond base
area contributing to the seepage, the
underdrained site would release about
0.4% as much water as the nondrained
pond.

Cost Estimates of FGD Waste
Disposal
   Costs were estimated for the various
FGD waste disposal methods related to
the type of disposal evaluation conducted
at Shawnee. The engineering cost
estimates for the disposal methods
summarized in  Table  7  are based  on
Figure 14.    Untreated underdrained Site G showing physcial stability within
              days after filling.
mid-1980 dollars and on other conditions
summarized in Table 8. Cost-effective
and environmentally sound disposal
methods—namely, chemical treatment,
untreated with underdrainage, and
gypsum with a low permeability indige-
nous clay liner and surface drainage-
cost between 0.79 and 1.46 mills/kWh
The disposal cost of chemically treatec
waste of 1.46 mills/kWh is based or
costs derived from data provided by the
three treatment contractors who parti
cipated in the Shawnee project and has
been updated to mid-1980 dollars.
Table 7.    Cost Comparison of Disposal Alternatives (Mid-1980 Dollars?
Annual Disposal Costs
Disposal
Untreated Waste Pond
Indigenous Liner
Synthetic Liner
(Polymeric)
Bentonite Clay
Liner
VolclaySS-100
Liner
Mills/kWh
0.51
1.23
2.06
0.94
$/ 'Tonne of
Dry Waste
5.01
11.96
20.01
9.13
c
$/Tonne
of Coal
1.51
3.62
6.04
2.76
Capital Investment Cost
$/kW Significant Cost Items"
14.80 Land (780 acres)
38.70 Land and liner
66.20 Land, clay liner, and clay transportation
29.00 Land and liner
  Underdrained                         0.86         8.33
   •
 Force-Oxidized Waste with Surface
 Drainage

  Indigenous Liner                      0.79         7.10

  Synthetic Liner
  (Polymeric)                           1.28        12.22

 Chemically Treated Waste Landfill        1.46        14.16
                        2.51     26.00 Land and underdrainage system
                        2.19    21.20  Oxidation equipment
                        3.78    39.20  Oxidation equipment and synthetic linei

                        4.78    17.78  Dewatering equipment and chemical
                                       reagents
 'Cost comparisons are for conditions summarized in Table 8.
 ''Relative to the various alternatives presented.
 "Includes disosal of ash in the waste.
                                 12

-------
  'able 8.    Summary of Base Conditions for Cost Estimation

        Item
                                               Base Condition
  Dollar Base

  Electric Utility Plant Size

  Coal Burned


  Heat Rate

  Annual Operating Hours


  Disposal Site Lifetime

  SOz Removal

  FGD Waste Generated
  Limestone Utilization


  Annual Capital Charges,
   3O-yr Average

  Cost of Land Used for Disposal

  Disposal Site Location

  Total Disposal Area Require-
   ments, (Including Berm) for
   a 9-m (30-ft) Waste Depth

    Chemically Treated Waste
    Force-Oxidized Waste
    Untreated Waste
    Lined
     Underdrained
                                 Mid-1980

                                 Two 500-MW units

                                 3.5% sulfur: 2.8 x 107 J/kg (12,000 Btu/lb)
                                 14% fly ash

                                 9.5 x 10* J/kWh (9000 Btu/kWh)

                                 5694 hr (30-yr average for a 65% lifetime
                                 capacity factor)

                                 30 yr

                                 90%

                                 5.8 x JO5 tonnes (6.4 x 10s tons)/yr, dry
                                 (general case)

                                 6.0 x 10s tonnes (6.6 x W5 tonsj/yr, dry
                                 (gypsum case)

                                 80% for limestone scrubbing; 100% for forced
                                 oxidation to gypsum


                                 17%

                                 $4940/ha ($2000/acre)

                                 Within 1 mile of steam plant
                                 263 ha (650 acres)
                                 236 ha (582 acres)

                                 295 ha (730 acres)
                                 316 ha (780 acres)
  These cost estimates differ slightly
from those presented in previous reports
on this project as a result of a revision
upward in average unit capacity factors
for the power stations and a reduction in
disposal land costs. Based on a sampling
of U.S. power companies and govern-
ment agencies, the estimated cost of
land was changed from $12,400 to
$5000/ha ($5000 to $2000/acre).
Also, on the basis of recent data, the 30-
year average unit capacity factor for
new coal-fired utility power plants is
now projected at 65% rather than 48.5%.
Both of these factors caused reductions
of about 0.17 mills/kWh for untreated
disposal and about 0.44 mills/kWh for
the gypsum disposal cases and are
reflected in Table 7.
  The cost of chemically treated waste
disposal increased by 0.21  to  1.46
mills/kWh (Table 7). Although costs
declined  as a result of the  land and
:apacity factor revisions indicated  here.
the decrease was offset by the inclusion
of capital investment contingency,
startup, and modification costs and
interest during construction (Table 9),
which had been previously applied only
to the untreated and gypsum disposal
costs.
  For the indigenous clay liner case, a
soil permeability coefficient of 10~7
cm/sec was taken to be representative
of clay to be used for pond  lining, and
consequently no cost was associated
with liner material. The cost of pond
construction for disposing of untreated
and force-oxidized wastes increases to
1.28 mills/kWh if a synthetic liner must
be  added compared to a  cost of 0.79
mills/kWh without a liner. The estimated
installed cost  of a  synthetic liner is
$6.46/m2 ($5.40/yd2).
  For the underdrained site, a design for
a minimal sand bed using a 0.3 m (1 ft)
thick sand layer and pipe spacing of 41
m  (133 ft) is the lowest cost design for
this type of disposal. Ten 20-ha (50-
acre) sites are required for this disposal
mode according to the baseline condi-
tions, and the cost is 0.86 mills/kWh.
The cost of the underdrainage  compo-
nents, including sand, gravel, pipes, and
fittings,  is approximately 20% of the
total capital cost.
  Disposal of force-oxidized FGD waste
includes  a  15% slurry pumped to the
site, where it settles, on the basis of test
site data, to approximately 65% solids
and the supernate water is recycled to
the scrubber. The cost of the oxidation
equipment is included as part of the
disposal costs.
  Differences in  capital investment
costs for the various alternatives shown
in Table 7 are attributable to differing
requirements for each type of disposal;
e.g., land requirements, whether a
synthetic linar or underdrainage system
is installed, and the  necessity  for de-
watering, oxidation,  or special slurry
                                                                                13

-------
Table 9.    Contingency and Miscellaneous Capital Investment Costs

                 Item                                 Cost
 Engineering

 Miscellaneous Services

 Contingency


 Startup and Modification Allowance


 Interest During Construction'
10% of capital equipment

0.5% of capital equipment

12% of capital equipment, plus
engineering, and miscellaneous services

6.7% of capital equipment, plus all of
the above

1.6% of capital equipment, plus all of
the above
 'Estimated as 4 months.

handling  equipment. The least capital
intensive method at $14.80/kW is for
the untreated case and involves waste
disposal in ponds with an indigenous
clay base. The most capital intensive
technique is for forced oxidation of the
waste prior to gypsum and disposal in a
pond lined with polymeric material; the
cost associated with this option is
$39.20/kW. As shown in Table 7, the
use of a liner has a marked effect on the
cost of disposal. Use of treated bentonite
clay liner (Volclay SS-100), a polymeric
liner, and a pure bentonite clay liner
increased the untreated waste disposal
cost from $14.80/kW to $29.00, $38.70,
and -$66.20/kW,  respectively.

Findings
  The results obtained from the FGD
waste disposal evaluation project indi-
cate several cost effective methods that
physically stabilize the wastes, prevent
water pollution, and permit site recla-
mation. These methods include chemi-
cal treatment, underdrainage of untreat-
ed waste, and surface  drainage  of
ponded gypsum.
  The results  show that:
  1. Chemical treatment reduces the
    coefficient of permeability by at
     least one order of magnitude and
     reduces the initial concentration
     of major species in leach ate  by
     about 50%.  Chemical treatment
    converts the waste into a structural
     material, is amenable to subgrade
     or abovegrade landfilling, and can
     be contoured to promote the runoff
     of rainwater. Typically, chemical
    treatment did not appreciably re-
     duce the concentration  of trace
     elements in the leachate. However,
     of the more than 600 leachate
     analyses conducted, only two
     showed  concentrations in excess
     of 10 times the National Interim
     Primary Drinking Water Regula-
       tions for selenium (13 and 20
       times the standard).
     2. Typically, after approximately 2
       years, the concentrations of the
       IDS in the leachates of the chemi-
       cally treated and untreated ponded
       wastes have approached  levels
       that approximate saturation with
       gypsum (approximately 2500
       mg/l).
     3. The ground waters associated
       with  all sites show no effects
       attributable to the waste disposal
       operations.
     4. Runoff from waste that was force-
       oxidized to gypsum shows  a  TDS
       concentration of between 2500
       and 3200 mg/l, and TSS  in the
       same runoff ranges between 4
       and 300 mg/l. When stacked
       above grade, uncompacted gypsum
       filter cake exhibits cracking under
       exposure to freeze/thaw or wet/dry
       conditions, thereby allowing rain-
       water to enter into the material.
       For structural strength of the  gyp-
       sum filter cake to be maintained,
       the site should be managed to
       reconfigure the disposal pile after
       weathering to shed water so that
       it is not allowed to reslurry by
       wetting.
     5. The underdraining  of untreated
       waste results in  a structurally
       sound material in an impoundment
       with all seepage and rainfall being
       controlled. Operationally the water
       would be returned to the scrubber
       and reused.
     6. Site closure by draining, covering,
       and contouring with indigenous
       clay of one chemically treated and
       two untreated underdrained sites
       has been shown to be adequate to
       support construction vehicles.

     7. Ponding  of untreated material is
       generally the least costly method
     of FGD waste disposal, but if the
     pond does not have a base material
     considered to be impermeable a
     liner must be added to reduce
     seepage, thereby significantly
     increasing the disposal cost. Since
     FGD wastes are thixotropic in
     nature, ponds of untreated wastes
     are considered nonstructural sites
     and are generally difficult to re-
     claim,  except possibly in areas of
     low rainfall and high evaporation.
  8.  Chemical treatment, ponding with
     underdrainage, gypsum with  an
     indigenous liner and surface drain-
     age, and chemical treatment with
     landfilling cost between 0.79 and
     1.46mills/kWh($7.10to$14.15/
     tonne waste, dry  basis).

Recommendations
  It is recommended that the following
FGD  waste disposal, site closure, and
reclamation techniques be further eval-
uated:
   1 .Long-term monitoring of covered
     disposal sites to evaluate  the
     effectiveness of site closure, rec-
     lamation, and repair. The two sites
     (one underdrained and one chemi-
     cally treated) which were covered
     with soil and sloped to facilitate
     runoff of rainwater have experi-
     enced seepage. This may be due to
     the  recontouring  and puncturing
     of the soil cap for the planting of
     trees and appears amenable to
     repair. Another underdrained site,
     which  was covered in April of
     1980, has not been effectively
     evaluated because of the low
     rainfall since construction.
  2.  Investigation of  the disposal of
     slurried force-oxidized waste.  It
     has been noted  that stacking of
     gypsum filter cake is subject to
     slumping and erosion. Although
     hard surface layers (crusts) form
     during'dry periods, these are not
     permanent and revert to the origi-
     nal condition during wet periods.
     Above-ground disposal and costs
     of slurried FGD gypsum similar to
     the phosphate gypsum disposal in
     central Florida  have not been
     assessed.
  3.  Long-term evaluation of the effects
     of the time and  weather on the
     physical and structural character-
     istics  of the  chemically treated
     materials. Recent tests of treated
     waste core samples have shown
     that although permeabilities are
     within the range of previous results^
                                 14

-------
values are typically on the lower
side. Weathering of a chemically
treated site that was converted to
a runoff configuration has resulted
in the formation of approximately
10 cm (4 in.) of granular waste on
the surface. In addition, the top 10
to 15 cm (4 to 6 in.) of waste in Site
B, which is generally underwater,
has deteriorated into a soft silty
material with  minimum physical
strength.
                                      P. R. Hurt, P. P. Leo, J. Rossoff, andJ. R. Witzare with The Aerospace Corpora-
                                        tion, Energy and Resources Division. P.O. Box92957, Los Angeles, CA 90009.
                                      Julian W. Jones is the EPA Project Officer (see below).
                                     • The complete report, entitled "Disposal of Flue GasDesulfurization Wastes: EPA
                                        Shawnee  Field Evaluation, Final Report," (Order  No. PB  81-212  482;
                                        Cost: $21.50, subject to change) will be available only from:
                                             National Technical Information Service
                                             5285 Port Royal Road
                                             Springfield, VA 22161
                                             Telephone: 703-487-4650
                                      The EPA Project Officer can be contacted at:
                                             Industrial Environmental Research Laboratory
                                             U.S. Environmental Protection Agency
                                             Research Triangle Park,  NC 27711
                                                                            1B
                                                                                                     -757-OU/729*

-------
                                              i
                                                           .
                                                          i s
                                                          a i
                                                          o =

                                                          I
                                                          N>
    ;   Cfl
O
>C
O
CO  «CO
0^00250
CX5CC
  •30
  H
     C*J
     'Sf
     n
                                                    m
01
                                                        3.S
                                                        S|s
                                                          I  ,

-------