Cementitious Stabilization
of
Chromium, Arsenic, and Selenium
in
a
Cooling
Tower Sludge
Roger
D.
Spence
T.
Michael
Gilliam
Alan Bleier
Chemical Technology Division
Oak
Ridge National Laboratory*
P.O.
Box
2008
Oak
Ridge,
TN
3783 1-6202
9
8
3
To
be presented at the
88th
Annual
Meeting
&
Exhibition
of
the
UWh4.A
San
Antonio, Texas
June 18-23,
1995
*Managed
by
Martin
Marietta
Energy
Systems,
Inc.,
under
contract
DE-ACO5-84OR21400
wtih
the
U.S.
Departmaent
of
Energy
1
DISCLAIMER
Portions
of
this document
may
be illegible
in
electronic image products. Images
are
produced from the best available original
document.
95-RP
130.03
INTRODUCTION
The
U.S.
Department of Energy
Oak
Ridge Operations Office
has
signed
a Federal Facility Compliance
Agreement
(FFCA)
with the
U.S.
Environmental Protection Agency Region
IV
regarding Oak Ridge
Reservation
(ORR)
mixed (radioactive and hazardous) wastes that
are
subject
to
the land disposai
restrictions
(LDR)
of
the Resource Conservation and Recovery Act
(RCM).
The
FFCA establishes an
aggressive schedule for conducting studies and treatment method development under the treatability
exclusion
of
RCRA (treatability studies) for those mixed wastes for which treatment methods and
capabilities have yet to be defined. One
of
these wastes
is
a radioactive cooling
tower
sludge with the
waste codes DO04 (arsenic),
DO05
(barium), DO07 (chromium),
DO08
(lead), and DO10 (selenium).
This
paper presents some results of a treatability study of the stabilization
of
this
cooling tower sludge
in
cementitious waste forms.
The sample
of
the cooling tower sludge obtained for this study
was
found to be not characteristically
hazardous [see the toxicity characteristic leach procedure (TCLP) results for the unspiked sludge
in
the
results section]
in
regard to arsenic, barium, chromium, lead, and selenium, despite the waste codes
associated with
this
waste. However, the scope
of
this study included spiking three RCRA metals to
two
orders
of
magnitude above the initial concentration to test the
limits
of
cementitious stabilization.
Since the sample appeared to contain little of the
RCRA
metals that
was
extractable, the decision was
made to spike the sludge to an initial level and then at
two
more levels, each an order of magnitude
above the preceding level. Based on prior characterization data
of
the coolig tower sludge, the metals
selected for spiking were chromium, arsenic, and selenium. Chromium and arsenic were spiked at
concentrations of
200,
2,000,
and
20,000
mg/kg, and selenium was spiked at 100,
1,000,
and
10,000
mgkg (concentrations based on the metal in the sludge solids).
The stabilization
of
arsenic, chromium, and selenium,
as
well
as
other metals, can be achieved with
grout.' The stabilization of inorganic arsenic can be achieved, even at
high
concentrations.
Stabilization of arsenic concentrations
as
high
as
20,000
mgkg in arsenic sludge has been reported
using Portland cement.* However, "very high levels" can be troublesome.' Specifically, arsenic sulfide
species are soluble under high pH conditions, uniike other metal sulfides with solubilities that are
usually much lower than those of the corresponding metal hydroxides. Hence, this behavior for arsenic
could make ground granulated blast fbrnace slag (referred to hereafter
as
slag) an undesirable stabilizing
agent for arsenic because the iron sulfide in slag can supply the problematic sulfide species. On the
other hand, the use of iron salts
is
promising in the stabilization
of
arsenic,'
so
slag may be
an
ideal
stabilizing agent
if
the detrimental role of iron sulfide can be controlled.
Chromium stabilization usually involves the reduction
of
the hexavalent state to
the
trivalent state
and
the precipitation
of
the chromium(m) hydroxide. The implication of this approach
is
that a
straightforward high-pH stabilization, such
as
that provided by Portland cement alone, may not succeed.
StabiIization of chromium concentrations as high
as
16,300 mgkg in electroplating sludge has been
reported using lime-s~lfide.~*~ Furthermore, whereas potassium silicate can stabilize similar chromium
concentrations
in
wood-preserving waste, cement-silicate and Portland cement do not.' Consequently,
high slag compositions were tested
in
the present study in order to take advantage
of
the reducing
potential of the slag.
Selenium is rarely found in industrial wastes,
it
does not leach above the RCRA limit
of
1
ma,
and
it
2
95-RP130.03
usually does not detectably leach fiom most cementitiously
stabiied
wastes. Stabition
of
selenium
concentrations
as
high
as
1275
mgkg in a
mix
of wastes
has
been reported
using
cement-fly ash.’
Based on the preceding considerations, Portland cement, Class
F
fly
ash,
and
slag
were selected
as
stabiiing agents
in
the present study. Perlite, a fine, porous volcanic rock commonly
used
as
a filter
aid, was used
as
a
water-sorptive agent in
this
study
in
order to control bleed water
for
high
water
contents. The highly porous perlite dust absorbs large amounts
of
water by capillary action
and
does not
present the handling
and
processing problems exhibited by clays used for bleed water control.
EXPERIMENTAL
The study scope included controlling the sludge water content and varying
this
water content over
a
wide range. For
this
reason, the cooling tower sludge was
first
oven dried at
105°C.
The dried sludge
was then sieved
through
4.75-mm
sieve openings and homogenized to provide
the
feed sludge
solids
for
the experimental design (see Table
1).
Homogeneity was tested by standard total analysis
(EPA
Method
305
1)
of a marker element (chromium) in five subsamples
of
the dried-sieved homogenized sludge. The
percentage relative standard deviation
(%
RSD,
Le.,
standard deviation divided by the mean times
100)
for chromium was
12%.
Grout Preparation
The grout preparation consisted
of
fist mixing the sludge solids
with
water and
the
spike compounds
and then mixing with the stabilizing agents. The treated sludge (grout)
was
cured
in
a humid
environment at room temperature for
28
d to make the cementitious waste
form.
The spiking procedure
consisted
of
mixing the spike compounds with some
of
the water overnight, adding
this
slurry
to the
sludge solids, rinsing the
slurry
container with the remainder of the water and adding
this
rinsate to the
sludge solids, and mixing this concoction for
20
min
with
a
model
N-50
Hobart
mixer
using a wire whip
on
low
speed. The spiked, wet sludge was then mixed with a
dry
blend
of
the Wig additives
for
4
min
in
the
Hobart
mixer. The compounds used for spiking were Na$r20,,
A@,,
and SeO,.
The dry blend consisted
of
as
many
as
four additives blended for
2
h
in
an
8-qt twin-shell blender
(Patterson-Kelley
Co.).
The four dry blend additives consisted
of
(1)
Type
I-II
Portland cement
(cement)
fiom
the Dixie Cement Co.,
(2)
Class
F
fly
ash
(fly
ash) fiom the American Fly
Ash
Co.,
(3)
ground granulated
blast
hrnace slag (slag) (Blaine fineness of 6220 cm2/g)
fiom
the Koch Minerals Co.,
and
(4)
perlite (Grade
H-200)
fiom the Harborlite
Corp.
The composition
of
each
of
these dry additives
was varied over a wide range, and only cement was present
in
every
dry
blend. Slag is a cement
substitute, but requires activation by a base.
Thus,
when slag was used
as
the
main
binder,
a
smd
amount
of
cement was
also
added to activate the slag. The grout compositions
were
chosen
in
a
statistical design (mixture experiment), but this statistical approach is not discussed
in
this
paper.
Modified
TCLP
Measurements
Both the sludge solids (unspiked and spiked at the three levels) and cementitious waste
forms
were
extracted using a modified TCLP. The modifications to the TCLP consisted
of
(1)
extracting
a
10-g
sample in
200
id
of
extraction fluid;
(2)
size reduction to
<4.75-mm
particles; and
(3)
analysis
of
the
extract
for
arsenic, selenium, and mercury by an inductively
coupled
argon plasma spectrometer (model
ICAP
61E
Trace Analyzer
fiom
Thermo Jarred Ash).
TCLP
extraction fluid
#2
(an
aqueous solution of
acetic acid at a
pH
of
about
2.8)
was required for
all
the TCLP extractions in this study.
3
95-RP
13
0.03
Density
Measurements
The bulk density of the as-received sludge and dried-sieved homogenized sludge was determined
by
weighmg a
known
volume (using a graduated cylinder) of the
granular
sludge and calculating the bulk
density. The bulk density of the cementitious waste forms was measured by packing a 2-in. cube mold
with the freshly made grout, determining the net weight of the
grout,
and calculating the bulk density.
The corresponding volumes of the as-received and dried-sieved homogenized sludge were calculated
along with each grout volume. The ratio of each grout volume to the sludge volumes gives
an
estimate
of
the volume increase that can be expected
fiom
cementitious stabilization
of
the cooling tower sludge.
RESULTS
Table
1
lists the compositions (including the spike levels of chromium, arsenic, and selenium in the
dried-sieved homogenized sludge solids) of the cementitious waste forms made
fiom
the cooling tower
sludge. Table
2
lists the RCRA metal concentrations
of
the TCLP
extracts
for
the dried-sieved
homogenized sludge (unspiked, low spike, medium spike, and
high
spike). Table
3
lists the TCLP
extract concentrations of the cementitious waste forms made from the cooling tower sludge according to
the compositions listed
in
Table
1.
The TCLP performance of the grouts proved to
be
sensitive
to
the
find
extract
pH.
For example, the
fraction of the spiked arsenic and selenium extracted
in
the TCLP test showed a definite correlation with
the final extract
pH,
as
illustrated in Figure
I.
The chromium concentrations had no obvious correlation
with
pH.
Figure
2
illustrates the ratio of the TCLP extract concentrations
of
chromium, arsenic,
and
selenium before and after treatment as a function of the
final
extract
pH.
These plots are quite 'similar
to
those of Figure
1.
(The results for the untreated spiked sludges are plotted
in
Figure
1
along with the
results for the grouts, but not in Figure
2.)
Table
4
lists
two
ratios for each grout composition:
(1)
the
grout
volume to the volume of the
as-
received sludge and
(2)
the grout volume to the volume
of
dried-sieved homogenized sludge. The
measured
bulk
densities (standard deviations
of
0.02
kgK,
for both)
were
0.81
and
1.07
kg/z
for the as-
received sludge and the dried-sieved homogenized sludge, respectively. The as-received sludge
experienced a
24.9
wt%
mass
loss
on drying, implying
an
initial water content of 24.9
wt%
in
the
as-
received sludge. Typically, the waste consists of the solids
plus
water and the volume increase over
this
combination would be reported. The bulk volume of
this
combination was not measured, and the air
voids in the granular as-received and dried sludge made estimation
of
this
volume inaccurate. Hence,
the volume increases over the combination of sludge solids plus water were not obtainable. The volume
ratios listed in Table
4
overestimate the volume increase for
high
water
contents because the water
volume is included in the grout and not the waste solids, but the ratios also underestimate for low water
contents because
of
the air voids present in both the as-received and dried sludge.
DISCUSSION
The compositions
of
the grouts in this study were intentionally varied over a wide range, including
variations
fiom
high
waste loadings to low waste loadings and
fiom
high
water contents to low water
contents. Most
of
the grouts listed
in
Table
1
formed relatively
weak
waste forms.
Only
a few
with
higher binder content formed strong cementitious monoliths. The grouts with
high
waste loadings
resulted
in
wet,
soft
products that flowed under their
own
weight, even after a 28-d cure. The higher
4