United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati, OH 45268
Research and Development
EPA/600/S2-90/043 Feb. 1991
 Project Summary
The  Fate  of  Trace  Metals in  a
Rotary Kiln  Incinerator with  a
Venturi/Packed  Column  Scrubber

D.J. Fournier, Jr., W.E. Whitworth, Jr., J.W. Lee, and L.R. Waterland
   A 5-week series of pilot-scale incin-
eration tests, employing a synthetic
waste feed, was performed at the U.S.
Environmental Protection Agency's
(EPA) Incineration Research Facility (IRF)
to evaluate the fate of trace metals fed to
a rotary kiln incinerator equipped with a
venturi scrubber/packed column scrub-
ber. Eight tests studied the fate of five
hazardous constituent and four
nonhazardous constituent trace metals
as a function of incinerator operating
temperatures and feed chlorine content.
Three tests evaluated the valence state
of chromium in emissions  and dis-
charges as a function of feed valence
state and chlorine content.
   Chromium test results indicated that,
when no chlorine was present in the
feed, 95% of the measured chromium
was discharged in the kiln ash, 1% to 2%
in the scrubber exit flue gas, and 3% in
the scrubber liquor. With chlorine In the
feed, these fractions were 85%, 4%, and
11%, respectively. Kiln ash contained
negligible hexavalent chromium Cr(+6)
for all tests. Scrubber exit flue gas Cr(+6)
as a fraction of total chromium was
nominally 15% with no feed  chlorine,
increasing  to 50% with chlorine-con-
taining feed. The scrubber liquor Cr(+6)
fraction was 20% to 30% with Cr(+3)
feed, increasing to about 60% with Cr(+6)
feed.
   Parametric trace metal test results
confirmed that cadmium, lead, and bis-
muth are relatively volatile, based on
normalized discharge distribution data.
Less than 32% of these metals were
discharged in the kiln ash. Barium, cop-
 per, strontium, chromium, and magne-
 sium are relatively nonvolatile; morethan
 75% of the amounts discharged were
 present in kiln ash. Surprisingly, arsenic
 was found to be relatively nonvolatile.
 Apparent scrubber collection efficien-
 cies generally correlated with observed
 volatilities and were 36% to  45% for
 volatile metals. Except for copper, non-
 volatile metals were collected at 49% to
 88% efficiency, increased feed chlorine
 significantly increased the volatility of
 cadmium,  lead,  and bismuth. It also
 shifted the afterburner exit particulate to
 smaller size distributions.
   The average mass balance closures
 around the kiln ash/scrubber discharges
 ranged from 48% to 96% for individual
 metals. Overall average closure was 71 %.
 From past experience, trace metal mass
 balance closure results for combustion
 sources are typically in the 30% to 200%
 range.
   This Project Summary was developed
 by EPA's Risk Reduction Engineering
 Laboratory, Cincinnati, OH, 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
   The RCRA hazardous waste incinerator
 performance standards promulgated by EPA
 in January 1981 established particulate and
 HCI emission limits and mandated 99.99%
 destruction removal efficiency for principal
 organic hazardous constituents (POHCs).
 Subsequent risk assessments have sug-
 gested that, of the total risk to human health
                                                 Printed on Recycled Paper

-------
and the environment from otherwise prop-
erly operated incinerators, hazardous con-
stituenttrace metal emissions may pose the
largest component. Currently, thedata base
available to support regulations on trace
metal emissions from incinerators is very
sparse.
   In response to these data needs, an
extensive series of tests was conducted at
EPA's IRF, with funding support from the
Office of Solid Waste (OSW), to investigate
the fate of trace metals fed to  a rotary kiln
incinerator equipped with a venturi/packed
column scrubber for particulate and acid
gas control.
   A primary objective was to investigate
the fate of five hazardous constituent trace
metals fed in a synthetic solid waste matrix,
as a function of incinerator operating tem-
peratures and feed chlorine content. These
five metals were arsenic, barium, cadmium,
chromium, and lead.
   As part of the EPA Risk Reduction En-
gineering Laboratory's program to support
the development of regulations concerning
trace  metal  emissions, a separate
OSW-sponsored effort  by another EPA
contractor is developing a numerical model
to aid in predicting the relative distribution of
trace metals in incineratordischarges. Thus,
a second objective was  to supply data to
evaluate the predictive capabilities of this
model and perhaps to guide further model
refinement. To  support this objective, the
tests also included four nonhazardous
constituent trace metals: bismuth, copper,
magnesium, and strontium.
   In the absence of information on the
predominant valence state of chromium in
incinerator discharges, risk assessments
have generally assumed  that the entire
discharge is in  the more toxic hexavalent
form. This  assumption has  resulted in
specifying conservative chromium emission
limits in  the regulatory  process.  Hence,
another objective of this program was to
develop data on the valence state of chro-
mium discharges as a function of chlorine
and chromium valence state in the feed.

Test Program
   The test program consisted of both an
eight-test, parametric, trace metals series
(Tests 4 through 11, in which the test waste
feed contained nine metals), and athree-test
chromium valence state series (Tests 1, 2,
and  3, in which the feed contained either
Cr(+3) or Cr(+6)). All tests were performed
in the rotary kiln incinerator system (RKS) at
the IRF.
   Figure 1 illustrates  a simplified RKS
schematic, and Table 1 summarizes its
design characteristics. The system consists
of a rotary kiln chamber, atransition section,
and a fired afterburner chamber. The pri-
mary air pollution control system  (APCS)
consists of a quench section, a venturi
scrubber, and a packed column scrubber.
In addition, a backup APCS, consisting of a
carbon-bed adsorber and a high efficiency
particulate (HEPA) filter, was in place to
ensure that organic compound and particu-
late  emissions  were negligible during
less-than-optimal test conditions.

Synthetic Test Mixture

   The synthetic waste fired during the test
program was composed of a clay absorbent
containing 30%  (weight) organic liquids.
Trace metals,  in  aqueous solution form,
were metered onto the solid, which was fed
to the rotary kiln via a screw feeder.
   Table 2 lists the metal concentrations of
the syntheticfeed waste for each test. Chro-
mium inhererit in the clay contributed about
50% of the actual feed chromium. Magne-
sium in the clay accounted for virtually all of
the actual feed magnesium.
   The organic liquid,  a mixture of toluene
and varying amounts of tetrachloroethylene
and  chlorobjsnzene,  supplied the  heat
content and POHC. It also introduced chlo-
rine at levels ranging from zero to nominally
8% by weight. The analyzed fractions of
these organics are listed in Table 3.

Test Conditions
            i
   For all tesjts, the clay/organic liquid was
fed at about 63 kg/hr (140 Ib/hr). The aque-
ous metal sc-lution was injected at 1 L/hr.
Estimated solids residence time in the kiln
was  1 hr. The test variables  for the para-
metrictestsw;erefeed chlorine content (0%,
4%, and 8%)J kiln temperature (816°, 871 °,
and  927°C), land  afterburner temperature
(982°, 1093°, and 1204°C). The three levels
for these  variables constituted a factorial
experimentalfmatrix.
   For the thijee chromium tests, chromium
valence state and chlorine content  of the
waste feed were varied. Two tests  had
trivalent chrohiium feed.  One of these had
no chlorine and the other had nominally 8%
chlorine  in the feed.  The third  test  had
hexavalent chromium and no chlorine in the
feed. Incinerator operating conditions dur-
ing these te^ts were  held constant: kiln
temperature | at  about 870°C and  after-
burner temperature at about 1093°C.
   For all tes|ts, excess air was targeted at
11.5% oxygen in the kiln and  7.5% oxygen
in the afterbuj-ner exit flue gas. These were
successfully rnaintained within ±1.5% for all
tests. Figure;2  illustrates the actual  and
target inciner^tortemperatures.Foralltests,
the primary A^PCS operated at design con-
ditions, with Venturi pressure drop at 5.0 to
6.2 kPa (20" to 25" W.C.) and scrubber pH
averaging 7.1.


Sampling and Analysis

   Continuous emission monitors were ar-
ranged to monitor O2, CO, CO., NOX, and
total unburned hydrocarbon (TUHC) con-
centrations at the  kiln exit, the afterburner
exit, the scrubber exit, and the stack. Other
samples  collected were feed,  scrubber
blowdown water, kiln ash, and flue gas at
the afterburner and scrubberexits. Samples
were also collected at the stack for evaluat-
ing hazardous waste management permit
compliance. Since the target analytes dif-
fered for  the two test series, the sampling
and analysis matrix was tailored specifically
for each series.
 ,  For the parametric trace metal tests,
composite feed samples were subjected to
ultimate,  volatile organic, and trace metal
analyses. Virgin clay samples were sub-
jected to trace metal analyses to determine
background metal concentrations. Com-
posite kiln ash samples were collected and
analyzed for metals.
   Flue gas sampling at the afterburner exit
consisted of volatile organic sampling train
(VOST)  sampling and  a variation of  a
Reference Method 17*  train that allowed
collection of large particulate mass, HCI,
and any vapor phase metals. At the scrub-
ber exit, a VOST train collected samples for
volatile organic analyses and a Method 5*
train modified forvaporphase metal capture
collected samples for metal  analyses. At
the stack, a Method Strain sampled for HCI
and particulate for permit compliance.
   For the chromium-valence tests, com-
posite feed samples were subjected to ulti-
mate, volatile organic, and total chromium
analyses. Composite grab samples of chro-
mium spike solution, scrubber blowdown,
and kiln ash were collected during each test
for total chromium and Cr(+6) analyses.
  . Flue-gas sampling employed two varia-
tions of a Method 5 train. At the afterburner
exit; one train with impingers containing
0.1 NNapH followed by HNO3/H?O sampled
for particulate load and  total chromium.
Sim ultaneously, a second train with the filter
removed sampled for Cr(+6)  and, during
test 3, for HCI. Impingers in this train con-
tained 0.1 N NaOH which was believed to be
better able to preserve Cr(+6). A similar
arrangement  was duplicated at the scrub-
ber exit.
   The VOST samples were subjected to
purge and trap GC/FID for organic analysis.
The grab composite samples of feed, ash,
  40 CFR, Part 60, Appendix A, July 1,
  1989.

-------
                                                                                                     ATMOSPHERE
     PROPANE
        TRANSFER
        DUCT
                                                                                                   STACK
                                                                                                                     in
                                                                                                                     m
                                                                                                           ID FAN
                                                                                   CARBON BED   HEPA   T(-nDA1M
                                                                                   ADSORBER     FILTER  'u UM/MIN

                                                                                PACKED COLUMN
                                                                                SCRUBBER
        AFTERBURNER
Figure 1. Schematic of the IRF rotary kiln system.

scrubber liquor, and  the  Method 5  and
Method 17 train samples were subjected to
inductively coupled argon plasma spectros-
copy  or atomic absorption s'pectroscopy
methods for metal analysis.

Test Results

Trace Metal Discharge
Distributions

   When subjected to incineration condi-
tions, the metals are expected to vaporize to
varying degrees, depending on their volatili-
ties. Figure 3 shows the amounts of metal
found in each discharge stream, normal-
ized to the total found in the three discharge
streams: kiln ash, scrubber exit flue gas,
and scrubber liquor. In Figure 3, the bar for
each  metal represents the range in the
fraction accounted for by each discharge
stream over all eight parametric tests, with
the average fraction from all tests noted by
the midrange tick mark. Metal discharge
distribution data in Figure 3 are plotted ver-
sus the metal's "volatility" temperature. A
metal's volatility temperature is the tem-
                                                   RECIRCULAT10N   RECIRCULATION
                                                   PUMP           TANK
perature at which a principal vapor species
of the metal hasavaporpressureof 10'6 atm.
Figure 3 shows that the more volatile met-
als, i.e., those having lower volatility tem-
peratures (Cd, Bi, and Pb), tend to be less
prevalent in the kiln ash. The more refrac-
tory metals, i.e., those with higher volatility
temperatures (Ba, Cu, Sr, Cr, and Mg) tend
to be found predominantly in the  kiln ash.
This observation isconsistent with expecta-
tion.
   The notable exception, arsenic, exhib-
ited  unexpected  refractory behavior and
remained  predominantly  in the kiln ash.
Arsenic's plotted  volatility temperature  is
that for As2O . The fact that arsenic was
observed to be  significantly less volatile
than expected, based on the As O3 volatility
temperature, suggests that As2O3 was not a
predominant arsenic species in the incin-
erator and that some other, less volatile
species (perhaps an arsenate) was the fa-
vored arsenic compound. An  alternative
explanation is that strong chemical inter-
action occurred between arsenic and the
clay feed matrix.
                                                                                   SLOWDOWN
                                                                                   COLLECTION OR
                                                                                   DISPOSAL
   It is interesting to note the sharp break in
observed volatility between lead and bis-
muth (average kiln ash fractions of 20% to
32%) and barium (average kiln ash fraction
of 77%). Six of the eight  parametric tests
were performed with kiln temperature of
870°C, one at 825°C, and one at 930°C.
These temperatures are right in the range of
thevolatilitytemperaturechangefrom621"C
(Bi) and 627°C (Pb) to 849°C (Ba).
Metal Distributions in Scrubber
System

   The above observations suggest that
substantial portions of the volatile metals
escaped the incinerator. For effective con-
trol, these need to be captured by the ven-
turi and packed column scrubber system.
Figure 4 shows that this  may not be the
case. The apparent scrubber collection ef-
ficiency averaged only 36% to 45% for the
volatile metals (Cd, Bi, and Pb). Collection
efficiencies for the most refractory metals
(Sr,  Mg, and Cr) averaged greater than
65%. This behavior is consistent with ex-

-------
Table 1.    Design Characteristics of the IRF Rotary Kiln System;

Chacioristics of the Kiln Main Chamber
   Length, outsicte
   Diameter, outside
   Length, inside
   Diameter, inside
   Chamber volume
   Construction
   Refractory

   Rotation
   Solids retention
     tf/ne
   Burner

   Primary fuel
   Feed system
     Liquids
     Sludges
     Solids
   Temperature (max)
2.61 m (8 ft 7 in)
1.22m (4 ft)
2.13m (7 ft)
0.95 m (3 ft 1-1/2 in)
1.74m3 (61.36 ft?
0.63 cm (0.25 in) thick cold rolled steel    ;
12.7cm (5in) thick high alumina castable refractory, variable depth
  to produce a frustroconical effect for moving solids
Clockwise or counterclockwise 0.2 to 1.5 rp'm
1 hr(at 0.2 rpm)                       i

American Combustion Burner, rated at 880,kW (3.0 MMBtu/hr) with
  dynamic O2 enhancement capability
Propane

Positive displacement pump via water-cooled lance
Moyno pump via front face, water-cooled lance
Metered twin-auger screw feeder or fiber pack ram feeder
900 °C (1650 °C)
Characteristics of the Afterburner Chamber
   Length, outside    3.05m (10 ft)
   Diameter, outside  1.22 m (4 ft)
   Length, inside     2.74m (9 ft)                          \
   Diameter, Inside   0.91 m(3ft)                          l
   Chamber volume  1.80m3 (63.6 ft3)
   Construction      0.63 cm (0.25 in) thick cold rolled steel    ',
   Refractory        15.24 cm (6 in) thick high alumina casfaWe^refracto/y
   Gas residence time 1.2 to 2.5 sec, depending on temperature and excess air
   Burner           American Combustion Burner rated at 440 kW (1.5 MMBtu/hr) with
                       dynamic Ot enhancement capability    '.
   Primary fuel       Propane
   Temperature (max) 1200 °C (2200 °F)                      [


Characteristics of the Air Pollution Control System
   System capacity
     Inlet gas flow   107rrf>/min (3773 acfm) at 1200 °C (2200 °,F; and 101 kPa (14.7psia)
   Pressure drop
     Venturi scrubber 7.5 kPa (30 in WC)
     Packed column  1.0kPa(4inWC)
   Liquid flow
     Venturi scrubber 77.2 L/min (20.4 gpm) at 69 kPa (10 psig) \
     Packed column  116 L/min (30.6 gpm) at 69 kPa (10 psig) ,
   pH control        Feedback control by NaOH solution addition


 Table 2.  Integrated Feed Metal Concentrations for the Parametric Trace Metals Test Series

                                                     Metal Concentration (ppm)
pectation. Most metal vaporized at some
point in the incinerator will ultimately con-
dense when the flue gas is cooled to its
scrubber exit temperature of about 75°C.
Condensation occurs via fume formation or
condensation onto available flue gas par-
ticulate. Fume formation results in very fine
paniculate.  Condensation onto available
particulate results inconcentratingthe metal
in fine particulate, since condensation is a
per  unit of surface area event and the
surface-area-to-mass ratio is increased in
fine  particulate. Both mechanisms tend to
concentrate volatilized metal in fine particu-
late. Since venturi scrubbers collect coarse
particulate more efficiently than they do fine
particulate, a poorer efficiency in collecting
volatile metals is expected.
   This same mechanism may explain the
poor collection efficiency observed for ar-
senic  and,  perhaps, copper.  The above
discussion noted that arsenic was observed
to be relatively refractory based on its high
kiln  ash fraction. However, any amount of
arsenic carried in entrained flyash out of the
kiln  and into the afterburner would experi-
ence the  higher afterburner temperatures
(980° to 1200°C in tests performed). Thus,
further opportunity to volatilize in the after-
burner exists.  If the predominant arsenic
species had a volatility temperature above
test  kiln temperature (825° to 930°C) but
belowafterburnertemperature, then volatil-
ization in the  afterburner  may have oc-
curred, leading to the poor scrubber collec-
tion  efficiency for the small amount of ar-
senic carried into the afterburner. Similarly,
entrained copper (volatility temperatures of
1116°C) may have volatilized  in the after-
burner, leading to the poor scrubber collec-
tion  efficiency that was  observed. If this
were the case, the good collection efficien-
cies for barium could remain unexplained,
however.
                 Indigenous  Testl    Test2   Tests    Tesf4    Tests   Test6    Test7   Tests   Test9   Test 10   Test 11
 Metal            (in clay) (9/29/89) (9/28/89) (9/26/89)  (9/14/88) (8/25/88) (9/16/88) (8/30/88)  (9/7/88)  (9/9/88) (9/20/88) (9/22/88)
Arsenic
Barium
Bismuth
Cadmium
Chromium
(total)
(+6)
Copper
Load
Magnesium
Strontium


12


53


3
22,000
34
25
36'
98
5 .

79 84 85 68
41 1.6 1.6 ;
260
33 >
77,700'
790
23
33
75
4

64

230
26
17,700
140
23
33
95
4

61

240
26
16,700
140
24
39
84
5

68

310
34
17,000
170
19
36
85
4

66

260
28
17,700
150
19
36
99
4

66

260
28
17,900
160
32
36
110
6

66

270
33
16,100
200
33
37
120
5

67

310
32
16,500
200

-------
Table 3.
                                                                                                                    KILN ASH
Weight % in mixture
Test Test Date

Mixture 1
Target
Composition
Measured Composition
1 9/29/88
2 9/28/88
4 9/14/88
Mixture 2
Target
Composition
Measured Composition
5 8/25/88
6 9/16/88
7 8/30/88
8 9/07/88
9 9/09/88
10 9/20/88

Mixture 3
Target
Composition
Measured Composition
3 9/26/88
11 9/22/88
Toluene



28.6
27.9
27.9
23.2


21.7

16.7
20.5
19.7
17.1
16.5
22.5



14.9

15.9
14.6
Tetrachloro-
ethylene



0
0
0
0


3.4

3.0
3.6
3.2
3.1
2.9
3.9



6.9

7.5
7.1
Chloro-
benzene



0
0
0
0


3.4

3.6
3.5
3.2
3.0
2.9
3.8



6.9

6.7
6.9
Chlorine
Content'



0
0
0
0


4

3.7
4.2
3.8
3.6
3.4
4.6



8

8.5
8.3
100
£ 80
SS"
LLJ
zl 60

z 3
O u_ 40
ss
1^
u. 20
o
I"


—

- }



Bi

:. ;
k i r
: I- 1 "J **•



.Pb
C.
^ I I I I I
0 200 400 600 800 1000 1200 1400 1600




45
*>
EC
as-35
m O
!±: lit
g- £ 30

wgg 25

2 § S 20
F °, 15
S "~ 10
(^
5
0
f

_

^
_

—

~
-
-
r AS
* i
t on
VOLATILITY TEMPERATURE ('C)


SCRUBBER LIQUOR




- Cd


Bi






Pb



- -Ba


Cr
u^-r-
n Ann enn o/vi -t nnn •! onn 1A(Y\ 1 Ann
'Based on measured tetrachloroethylene and chlorobenzene concentrations.
                                                                                                             VOLATILITY TEMPERATURE {'C)
                                                                                                              SCRUBBER FLUE GAS
   1,200 -
1,150 —
   1,100 —
                        ®
                                              © TARGET CONDITION

                                              n ACHIEVED CONDITION.
                                             -1  INTERSECTION IS MEAN
                                                CONDITION; BARS DENOTE
                                                RANGE OVER TEST.
                                                                                   ff
80
70
60
SO
40

30
20

10
0

- _,
-
—
_

.Cd


Bi

_


Pb






_ — Ba
i*8 i i i iS 4



-Cu
Sr u Cr
1 IX "£*
                                                                                      0     200    400     600    800    1000   1200   1400
                                                                                                          VOLATILITY TEMPERATURE CC)


                                                                        Figure 3. Distribution of metals in discharge streams.
                                                                                                                                                1600
   1,050 —
   1,000 -
                              I-
                                 I
                                     ^
                                               H
                                                  I
                  800            850              BOO

                           KILN TEMPERATURE ('C)
   Figure 2. Actual versus target operating temperatures
            for parametric trace metal tests.
10U
g 90
E 0 80
§i 70
oca
8£ 60
L. UJ
Ez 50
if:
8 20
10
?Ba
_ 4- Cr
w
j-

• - As "1
i
-

BI '
• Cd f



tPb



(,



a





Cu



.Cr




i i i i i i i i
                                                                                     0      200    400     600    800    1000   1200   1400   1600
                                                                                                          VOLATILITY TEMPERATURE ('C)

                                                                         Figure 4. Apparent venturi/packed column scrubber efficiencies for trace
                                                                                   metals.

-------
Effect of Incinerator Operating
Conditions

   Within the tested range, kiln tempera-
ture  had a minor effect on the discharge
distributions of arsenic, cadmium, and lead.
Of the three metals, the  effect was most
pronounced, though still minor, for cad-
mium.
   Within the tested range,  afterburner
temperature had little or  no effect on dis-
charge distributions of the nonvolatile met-
als.  Increasing  afterburner temperature
caused a slightly increased fraction of the
three volatile metals (cadmium, lead, and
bismuth) in the flue gas and a correspond-
ing decreased fraction in the  scrubber li-
quor. The most pronounced effect was for
lead.
   Varying the chlorine in the waste had
little to no effect on the distribution of ar-
senic, barium, chromium, strontium, or
magnesium. However, the increase in vola-
til'rtyforthe volatile metals cadmium, bismuth,
and lead, in addition to that for copper, was
substantial. The most pronounced effect
was for lead. This observation is consistent
with  expectations  since metal chlorides
(possible with chlorine in the feed) are typi-
cally more volatile  than are the base metal
or metal oxides. This is especially the case
for lead. With  no chlorine present, lead
metal is the most volatile lead species, with
a volatility temperature of 627°C. With
chlorine present, the formation of the signifi-
cantly more volatile PbClx, with a volatility
temperature of -15°C, is possible.

Panicle Size Distribution

    Figure 5 shows that the afterburner exit
particle size distributions for all tests were
roughly log normal. The absence of chlorine
in the waste appears to shift the particles to
a larger size distribution. This is consistent
with the expectation that presence of chlo-
rine increases the volatility of the inorganic
constituents in the feed. In such a case,
condensation of volatilized inorganic com-
poundswouldtend to produce finerparticles.

 Metal Particle Size Distribution

    Trace metal contents in four size frac-
tions in afterburner exit particulate (nomi-
 nally <2,2 to 4,4 to 10, and >10 u.m) were
 analyzed.  Figure  6a shows  the  average
fractions of arsenic, barium, cadmium,
chromium, and lead that were distributed
over the four size ranges. Similar results for
the four nonhazardous constituent metals
 (bismuth, copper, magnesium, and stron-
 tium) are shown in Figure 6b.
    It appears that the more volatile metals
 bismuth and barium tend to concentrate on
                                              10            to

                               Particl* D!am«t»r (mieronii)

Figure 5.    Afterburner exit flue gas size distributions for the parametric trace metal tests.
                                                                               to
o

I
£   »4
o
v
••*

i   6°
a.
 £
 3
     0.2
            •  ;  Total Sampl*

            +    A«

            O  '  B«

            A  '  Cd

            x  ;  cr

            V    Pb
                                                                                10
                                  Particl* Di«m«t«r (microns)
Figure 6a.  Hazardous constituent trace metal size distributions in afterburner exit flue gas
           particulate for the parametric trace metal tests.
the finer particlejs. The relatively nonvolatile
metals chromium, copper, magnesium, arid
strontium tend to concentrate on the larger
particles. This observation is consistent with
the volatilization!/concentration mechanism
discussed previously.
                                          Chromium Valence State
                                          Distributions
                                             Chromium discharge distributionsforthe
                                          focused chromium valence state tests, ex-

-------
pressed asfractionsof total measured chro-
mium, are tabulated in Table 4. In these
tests  as well, chromium clearly remained
predominantly in the kiln ash (greater than
80% of total). The scrubber exit flue gas and
blowdown liquor streams  combined.ac-
counted for less than 5% of the discharged
chromium with no feed chlorine. With chlo-
rine in the feed, the flue gas chromium
fraction doubled to about 4%. Similarly, the
scrubber blowdown liquor chromium frac-
tion increased to about 11%.  .
    Table 5 summarizes the fraction of the
total  chromium in feed  and discharge
streams  analyzed  as  being Cr(+6). As
shown, the feed was 52% Cr(+6) for the test
with Cr(+6) spiked into the feed and was
analyzed as 2% Cr(+6) for the tests with
Cr(+3) spiked into the feed. For all tests, the
kiln ash chromium was composed of negli-
gible amounts of Cr(+6).
    With no chlorine in the feed, the scrub-
ber exit flue gas Cr(+6) fraction was the
same (12% to 16%) regardless of whether
Cr(+6) was present in the feed. In contrast,
when  the feed contained chlorine, roughly
half the scrubber exit flue gas chromium
was Cr(+6). This would be expected if some
of the entrained particulate chromium from
the kiln vaporized in the hotter afterburner
and reacted with the flue  gas chlorine to
form chromyl chloride (CrO2CI2), a relatively
stable compound with chromium as Cr(+6).
   The scrubber liquor Cr(+6) fraction for
the test with  Cr(+6) in the feed was signifi-
cantly higher than for the  other two tests
with only  Cr(+3) in the feed. This is as
expected  if some of the chromium in the
scrubber inlet flue gas was present as soluble
Cr(+6) species (CrO~ and  Cr2O7=).

Conclusions
     The observed metal volatilities, based
     upon normalized discharge distribu-
     tion data, generally agreed with theo-
     retical predictions based on volatility
     temperature. A notable exception was
     arsenic, which is'theoreticallythe most
     volatile metal; it exhibited one of the
     lowest  observed volatilities.
     Cadmium,  bismuth,  and lead were
     relatively volatile, with an average of
     less than 32% of the discharge metal
     accounted for by the kiln ash. Barium,
     copper, strontium, chromium,  mag-
     nesium, and arsenic were more re-
     fractory, with greater than 75% of the
     metal discharge accounted for by the
     kiln ash.
     Apparent venturi/packed  column
    scrubber collection  efficiencies de-
    creased as metal volatility increased.
                                               SS.8
  .S  84  -\
  a
  *r
  «*
  C


  a

  _>
  7  ««
  a  le
  3
     0.2
             •   To

             •f   81
             O   Cu
             A   M«

             X   sr
                                  Partial* Oiam*t«r (microns)
Figure 6b.  Nonhazardous trace metal size Distributions in afterburner exit flue gas paniculate for
           the parametric trace metal tests.
 Table 4.  Total Chromium Discharge Distributions for the Chromium Valence State Test Series
                                   Total Cr Fraction (% of measured)
                         Testl
                    Cr{+6) feed, no Cl
     Test 2
Cr(+3) feed, no Cl
        Test3
Cr(+3) feed, 8.5% feedCI
Kiln ash                   95.5
Scrubber exit flue gas         1.3
Scrubber liquor              3.2
Total                     100.0
Apparent scrubber Cr        71
   removal efficiency
     95.2
       1.8
       ao
     100.0
     63
        84.4
         4.2
        100.0
         73
Tables.
           Hexavalent Chromium Fractions for the Chromium Valence State Test Series

                                         Cr(+6)/total Cr(%)
                         Testl
                    Cr(+6) feed, no Cl
     Test 2
Cr(+3) feed, no Cl
        Test 3
Cr(+3)rfeed, 8.5% feedCI
Composite feed            52
Kiln ash                    0.3
Scrubber exit flue gas       12
Scrubber liquor            57
      2
      0.2
     16
     21
         2
         0.1
        48
        28

-------
     Increasing feed chlorine content
     measurably increased the volatility of
     the volatile metals  (cadmium, lead,
     and bismuth) and  of copper (less
     volatile).
     Consistent with expectation, increas-
     ing kiln temperature tended to slightly
     decrease the fraction of volatile metal
     accounted for by the kiln ash.
     Discharge distributions of the non-
     volatile metals were generally unaf-
fected by variations in the test variables
over the range tested.
Particle size distributions for the after-
burner exit flue  gas were roughly
log-normal- the presence of chlorine
appeared to shift the particles to a finer
distribution.
In the scrubber exit flue gas, chromium
existed predominantly in the trivalent
form regardless of valence state in the
feed, provided the feed contained no
     chlorine; its hexavalent fraction in-
     creased in the test where chlorine
     was present in the feed.

   The full report was submitted in fuKill-
mentof Contract No. 68-03-3267by Acurex
Corporation under the sponsorship of the-
U.S. Environmental Protection Agency.
   D. J. Foumler, Jr., W. E. Wh'rtworth, Jr., J. W. Lee, and L. Hi Water/and are with Acurex Corporation, Jefferson, AR 72079.
   R. C. Thumau is the EPA Project Officer (see below).
   The complete report, consists of two volumes entitled "The Fate of Trace Metals in a Rotary Kiln Incinerator with a Ventun/
     Packed Column Scrubber:"                           \
   Volume I (Order No. PB90-263864 /AS; Cost: $23.00 subject to change) discusses the results of the technical data.
   Volume II Otder No. PB90-263872/AS; Cost: $31.00 subject to change) consists of the appendices.
   Both volumes of this report will be available only from:     j
                   National Technical Information Service
                   5285 Port Royal Road                 •
                   Springfield, VA 22161                 '
                   Telephone: 703-487-4650
   The EPA Project Officer can be contacted at:
                   Risk Reduction Engineering Laboratory  '
                   U.S. Environmental Protection Agency  \
                   Cincinnati, OH 45268
United States
Environmental Protection
Agency
  Center for Environmental Research
  Information
  Cincinnati, OH 45268
               BULK RATE
         POSTAGE & FEES PAID
          EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300

EPA/600/S2-90/043

-------