United States
                Environmental Protection
                Agency
 Risk Reduction
 Engineering Laboratory
 Cincinnati OH 45268
                Research and Development
 EPA/600/S2-91/032  Sept. 1991
EPA       Project  Summary
                The Fate  of  Trace  Metals  in  a
                Rotary  Kiln  Incinerator with  a
                Single-Stage  Ionizing  Wet
                Scrubber
                D. J. Fournier, Jr., and L. R. Watertand
                 A 3-week series of pilot-scale incin-
                eration tests was performed at the U.S.
                Environmental  Protection Agency's
                (EPA)Incineration Research Facility (IRF)
                In Jefferson, AR, to evaluate the fate of
                trace metals fed to a rotary kiln Incinera-
                tor equipped with a single-stage Ionizing
                wet scrubber for control of particuiates
                and acid gas. Test variables were kiln
                temperature, ranging from 816° to 927°C
                (1500° to 1700°F); afterburner tempera-
                ture, ranging from 982° to 1205°C (1800°
                to 2200°F);  and  feed chlorine content,
                ranging from 0% to 8%.
                 The test results Indicated that cad-
                mium and bismuth were relatively vola-
                tile, with an average of less than 40%
                discharged  with the kiln ash. Arsenic,
                barium, chromium, copper, lead,  mag-
                nesium, and strontium were relatively
                nonvolatile, with an average of greater
                than 80% discharged with the kiln ash.
                Observed relative metal volatilities gen-
                erally agreed with the volatilities pre-
                dicted based on vapor pressure/tempera-
                ture relationships, with the exception of
                arsenic, which was much less volatile
                than predicted. Cadmium, bismuth, and
                lead were more  volatile at higher kiln
                temperature; the discharge distributions
                of the remaining metals were not signifi-
                cantly affected by kiln temperature.
                 Enrichment of metals in the fine-par-
                ticulate fraction  was observed at the
                afterburner  exit, with an average of
                roughly 50% of the flue-gas particulate
                metal In the less-than-10-p.m size range.
                The distributions of the more-volatile
                metals were shifted to fine particulate
                more so than those of the less-volatile
                metals. Both increased kiln temperature
and the addition of chlorine to the syn-
thetic waste feed caused a shift of met-
als to fine particulate. Apparent scrub-
ber collection efficiencies for the metals
averaged 22% to 71%, and were gener-
ally higher for the less-volatile metals.
The overall average metal collection ef-
ficiency was 43%. It should be noted that
Industrial applications of ionizing wet
scrubbers are typically in multiple stages
and, as such, would be expected to col-
lect metals more efficiently  than the
single-stage scrubber at the IRF.
  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 (sea Project Report
ordering Information at back).

Introduction
  The hazardous waste incinerator perfor-
mance standards, promulgated by EPA in
January 1981 under the Resource Conser-
vation  and Recovery Act, established par-
ticulate and HCI emission  limits and man-
dated 99.99% destruction and removal effi-
ciency for principal organic hazardous con-
st 'rtuents(POHCs). Subsequent risk assess-
ments have suggested that hazardoustrace-
metal emissions may pose the largest com-
ponent of the total risk to human health and
the environment from otherwise properly
operated incinerators. The data base  on
trace-metal emissions from incinerators is
sparse, however; data on the effects of
waste  composition and incinerator opera-
tion on these emissions  are particularly
lacking.


         ^A> Printed on Recycled Paper

-------
   In the response to these data needs and
with support from the Office of Solid Waste,
an extensive series of tests was conducted
at EPA's IRFto investigate the fate of trace
metals  fed to a rotary kiln  incinerator
equipped with a single-stage, ionizing wet
scrubber. This program was a continuation
of a previous IRF test program, conducted
in 1988, that employed a venturi scrubber/
packed-column scrubber as the primary air
pollution control system.
   The primary objective of these test pro-
grams was to investigate the  fate of five
hazardous and four nonhazardous trace
metals fed to a rotary kiln incinerator in a
syntheticsolid-waste matrix. Of interest was
the distribution of the metals as af unction of
incineratoroperatingtemperaturesandfeed
chlorine content. The hazardous trace met-
als investigated were arsenic, barium, cad-
mium, chromium, and  lead.  The nonhaz-
ardous metals were bismuth, copper, mag-
nesium, and strontium.

Test Program
   The test program consisted of nine para-
metric tests in which the waste feed con-
tained the nine metals identified above. All
tests were conducted in the pilot-scale ro-
tary kiln incinerator system at the IRF (Fig-
ure 1).

Synthetic Waste Mixture
   The synthetic waste contained a mixture
of organic liquids added to a clay absorbent
material. Trace metals were incorporated
by spiking an aqueous mixture of the metals
of interest onto the clay/organic-liquid ma-
terial. The waste was fed to the rotary kiln
via a twin-auger screw feeder at a nominal
rate of 63 kg/hr (140 Ib/hr).
   The organic-liquid base consisted of tolu-
ene, with varying amounts of tetrachloro-
ethylene and chlorobenzene  added to pro-
vide a range of chlorine contents. Synthetic
waste chlorine was variedfrom 0%to nomi-
nally 8%. The analyzed organicfractionsfor
the three waste-feed mixtures are given in
Table 1. Table 2 summarizes the average
metal concentrations inthe combined waste
feed over the nine tests.

Test Conditions
   The test variables were kiln temperature,
afterburner temperature, and the chlorine
content of the synthetic waste feed. Seven
specific combinations of these variables
were  selected at test points. Target and
average achieved values for these three
variables are summarized in Table 3. For all
tests, excess air was nominally 11.5% oxy-
gen in the kiln and 8% oxygen in the after-
burner exit flue gas. Estimated solids resi-
dence time within the kiln was 1 hr.
Test Results

Average Trace-Metal Discharge
Distributions
  Figure 2 shows the amounts of metal
found in each discharge stream, as a frac-
tion  of the total in the three discharge
streams: kiln ash, scrubber-exit flue gas,
and scrubber liquor. In Figure 2, the bar for
each metal  represents the range in the
fraction accounted for by each discharge
stream over all nine tests, with the average
fraction from all tests noted by the midrange
tick mark. Metal discharge distribution data
in Figure 2 are plotted versus the volatility
temperature of each metal,  which is the
temperature at which the  effective vapor
pressure of the metal is 1O* atm. The effec-
tive vaporpressure is the sum of the equilib-
rium vapor pressures of all species contain-
ing the metal. It reflects the quantity of metal
that  would vaporize under a given set of
conditions. A vapor pressure of 10* atm is
selected  because it represents a measur-
able amount of vaporization. The lower the
volatility temperature, the more volatile the
metal is expected to be.
  Figure 2 indicates a correlation between
observed volatility and volatility tempera-
ture for all the metals tested, except arsenic.
With the exception of arsenic, average nor-
malized  kiln-ash fractions generally in-
creased with increasing volatility tempera-
ture. Cadmium and bismuth were relatively
volatile and were less prevalent in the kiln
ash than were the more-refractory metals.
Kiln-ash fractions accounted for the major-
ity of arsenic, lead, barium, copper, stron-
tium, magnesium, and chromium.
  Based on volatility temperature, arsenic
is expected to be the most volatile element.
The  data, however,  show arsenic to be
apparently refractory, remaining largely with
the kiln ash. The volatility temperature for
arsenic is based on the vapor pressure of
As2O3. The fact that arsenic is significantly
less volatile than expected (were As2O3 the
predominant arsenic species) suggests that
either some other, less-volatile arsenic com-
pound (perhaps an arsenate) was preferred
or that some other chemical interaction,
such as strong adsorption to the clay, oc-
curred.
Effects of Incinerator Operating
Conditions on Metal
Distributions
  Increased kiln temperature caused a no-
ticeable increase  in the volatility of cad-
mium, bismuth, and lead. Figure 3 shows
that as the kiln temperature increased there
was a significant decrease in the kiln-ash
fraction of these metals, with corresponding
increases in the scrubber-exit flue-gas and
scrubber-liquor fractions. Although the vola-
tility of lead increased with higher kiln tem-
perature, lead still remained relatively re-
fractory and was found primarily in the kiln
ash. Kiln temperature within the tested range
had no significant effect on the discharge
distributions of any of the remaining metals.
  Afterburnertemperatureswithinthe tested
range did not clearly affect the distributions
of any of the metals among the scrubber-
exit flue-gas and scrubber-liquor discharge
streams. Data on the effect of feed chlorine
content are inconclusive pending investiga-
tion of an apparent relationship between
feed chlorine and the efficiency of the ana-
lytical procedure for metals in kiln ash.
Apparent Scrubber Collection
Efficiencies
  The apparent scrubber collection  effi-
ciency for flue-gas metals was determined
for each test. The apparent efficiency repre-
sents the ratio of the normalized metal frac-
tion measured in the scrubber liquor to the
sum of the normalized metal fractions mea-
sured in the scrubber liquor and scrubber-
exit flue gas. Figure 4 summarizes the effi-
ciency data. The bar for each metal repre-
sents the range of scrubber efficiencies
over the nine tests, with the overall average
for the nine tests noted by the midrange tick
mark. Average metal collection efficiencies
ranged from 22% to 71%; the overall aver-
age for all metals was 43%. It should be
noted that industrial applications of ionizing
wet scrubbers are typically in multiple stages
and, as such, would be expected to collect
metals more efficiently than the single-stage
scrubber at the IRF.  Figure 4 shows  that
there were significant variations in the effi-
ciencies for each metal. Average efficien-
cies, however, were generally higherforthe
less-volatile metals.
  Within the limits of data variability, none
of the test variables affected scrubber col-
lection efficienciesfor arsenic, barium, stron-
tium, magnesium, and chromium. Efficien-
cies for cadmium, bismuth, lead and copper
increased with increased  kiln temperature
and waste feed chlorine content. Increased
efficiency might be expected with increased
feed chlorine content if the  presence of
chlorine leads to the formation of more-
soluble metal chlorides. It is unclear, how-
ever, why increased kiln temperature would
directly lead to  increased collection  effi-
ciency. Apparent scrubber collection  effi-
ciencies for metals did not vary with after-
burner exit temperature.

-------
                              Single-stage Ionizing
                                  Wet Scrubber
                                 Scrubber Liquor
                                  Recirculation
         Quench
                _L
    Afterburner   U
     Air
Natural'
Gas,
Liquid
Feed
Transfer
Duct
    Ash
                                                         Demister
                                                    \Carbon Bed
                                                    Adsorber  He'pa
                                                              Filter
                                      Atmosphere

                                           Stack


                                          to Fan
           Rotary
           Kiln
                 Natural
                 Gas, Liquid
                 Feed
          Rotary Kiln
          Incinerator
       Scrubber Liquor
       Recirculation
  Modular Primary Air
Pollution Control Devices
                                                      Redundant Air
                                                     Pollution Control
                                                         System
  Figure 1. Schematic of the IRF rotary kiln incineration system.
Metal Distributions In Flue-Gas
Partlculate by Particle Size
  The participate samples from the after-
burner-exit flue-gas  sampling train were
size-fractionated, and trace-metal distribu-
tions as  a function of particle size were
determined. Rgure 5 shows the metal distri-
butions in the particle-size range of less
than 10 u,m, averaged over all nine tests.
The average of the  nine total participate
samples  is also shown. The data show a
relationship between the relative volatility of
each metal (as indicated by its volatility
temperature noted on the horizontal axis)
and its propensity for redistribution to finer
paniculate. This is indicated  by the higher
fractions  of the metals with lower volatility
temperatures, in the  less-than-10-u.m par-
ticle-size fractions.
  This behavior is consistent with expecta-
tion. Most metal vaporized at some point in
the incinerator will ultimately condense when
the flue gas cools. Condensation occurs via
fume formation or condensation onto avail-
able flue-gas particulate. Fume formation
results in very fine particulate. Condensa-
tion onto available particulate results in con-
centrating the metal in fine particulate, be-
cause 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  particulate. Interestingly, ar-
senic behaves as a volatile metal with re-
spect to enrichment in fine particulate.
  The effects of kiln temperature,  after-
burner temperature, and waste feed chlo-
rine content are shown in Figure 6. The size
distributions of the metals most nearly re-
                                         flect the overall sample particle-size distri-
                                         bution for Test 2 (lowest kiln temperature),
                                         Test  5 (highest afterburner temperature),
                                         and Test 1 (no chlorine in the waste feed);
                                         very little redistribution among the particu-
                                         late was observed. Forthese three tests, an
                                         average of about 20% to 25% of each m etal
                                         and the total particulate sample were in the
                                         less-than-10-u.m particulate.
                                           With increased kiln temperature, the size
                                         distributions of all metals except chromium
                                         shifted to about 60% less than 10 urn.
                                         Increased afterburner temperature caused
                                         a shift in the  overall  sample to coarser
                                         particulate, most likely because of fine par-
                                         ticles melting or softening and coalescing
                                         into larger particles. A corresponding shift in
                                         metal-specific distributions to coarse par-
                                         ticulate was observed.
                                           The addition of chlorinated compounds
                                         to the synthetic waste feed mainly affected
                                         cadmium, lead, copper, and chromium dis-
                                         tributions. With chlorine content increased
                                         from  0% to 4%, the fraction of cadmium,
                                         lead, and copper accounted for by the less-
                                         than-10-um particulate increased from about
                                         20% to roughly 55%. No further redistribu-
                                         tions of these metals were observed with
                                         increased chlorine from 4% to 8%.  For
                                         chromium, increasing chlorine content!rom
                                         0% to 4% to 8% caused a corresponding
                                         shift of 2% to 20% to 50% in particulate of
                                         less than 10u.m.

                                         Conclusions
                                           Test conclusions include the following:
                                           • Cadmium and bismuth  were relatively
                                         volatile, with an average of less than 40% of
                                         the discharged metal accounted for by the
                                                            3
kiln ash. Arsenic, barium, chromium, cop-
per, lead, magnesium, and strontium were
relatively nonvolatile, with an average of
greater than 80% of the discharged metal
accounted for by the kiln ash
   • Observed metal volatilities generally
agreed with the order predicted by metal
volatility temperatures, with the notable ex-
ception of arsenic. Arsenic has the bwest
volatility temperature of metals tested but
was observed to be one of the least-volatile
of the metals. This suggests that As2O3 was
not the predominant arsenic species in the
incineratororthat the arsenic was adsorbed
by the clay/ash matrix.
   • Kiln temperature affected the relative
volatility of  cadmium, bismuth, and lead.
The fractions of these metals discharged in
the kiln ash decreased with increasing kiln
temperature.
   • Afterburner exit temperature did not
clearly affect metal partitioning among the
scrubber-exit flue-gas and scrubber-liquor
discharge streams.
   • Enrichment of metals in the f ine-particu-
late fraction of the afterburner-exit particu-
late was observed; an average of roughly
50% of the flue-gas particulate metal was in
the less-than-10-u.m size range compared
with an average of about 30% for the total
particulate sample. The distributions of the
more-volatile metals were shifted  to fine
particulate more so than those for the less-
volatile metals. Arsenic behaved as a vola-
tile metal with respect to its distributions
among the  afterburner-exit  flue-gas par-
ticle-size ranges.
   • Each test variable affected the distribu-
tions of at least some of the metals among
the flue-gas particulate particle-size ranges.
Size distributions of the metals most nearly
reflected  the overall sample particle-size
distribution  for Test 2 (lowest kiln tem-
perature), Test 5 (highest afterburner tem-
perature), and Test  1 (no chlorine in the
waste feed); very little redistribution among
the particulate was  observed. For these
three tests, about 20% to 25% of each metal
and the total particulate sample were found
in the less-than-10-u.m particulate
   •Increasing kiln temperature from 816° to
927°C (1500° to 1700°F) caused the aver-
age particle-size distributions to shift from
roughly 20% less than 10 urn  to an average
of 60% less than 10 urn for all test metals
except chromium.  For cadmium, copper,
and lead, an increase in waste feed chlorine
content from 0% to 4% caused their distri-
butions to shift from roughly 20% less than
10 jim to 55% less than 10 u.m. No further
effects with feed chlorine increased to 8%
were observed for these metals. For chro-
mium,  increasing chlorine content from 0%
to 4% to 8% caused a corresponding shift of
2% to 20% to 50% in particulate less than 10
u.m.

-------
  Table 1. POHC Concentrations In Clay/Organic-Uquid Feed

                                        Weight % in mixture
Test
1
2 through 8
(average)
9
Toluene
23.1
17.8
11.6
Tetrachloroethylene
0
3.1
6.0
Chlorobenzene
0
3.0
5.6
Chlorine Content*
0
3.6
6.9
  •Calculated based on measured tetrachloroethylene and chlorobenzane concentrations.
   Tablo2. Average Integrated Feed Metal Concentrations
Concentration,
Metal mg/kg
Arsenic
Barium
Bismuth
Cadmium
Chromium
48
390
330
10
40
Metal
Copper
Lead
Magnesium
Strontium
Concentration,
mg/kg
380
45
18,800
410
  Teblo 3. Target and Average Achieved Test Conditions
Feed Mixture Cl
Content, %
Test
1
2
3
4
5
6
7*
8*
9
Date
8/17/89
8/2/89
8/4/89
8/1/89
8/16/89
8/15/89
8/9/89
8/11/89
7/28/89
Target
0
4
4
4
4
4
4
4
8
Actual
0
3.5
3.5
3.5
3.7
3.6
3.5
3.8
6.9
Kiln Exit
Temperature,0C(°F)
Target
871 (1600)
815 (1500)
927 (1700)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
Average
900 (1652)
819 (1507)
929 (1704)
877 (1610)
885 (1625)
887 (1629)
881 (1618)
879 (1615)
881 (1617)
Afterburner Exit
Temperature, °C(°F)
Target
1093 (2000)
1093 (2000)
1093 (2000)
1093 (2000)
1204 (2200)
982 (1800)
1093 (2000)
1093 (2000)
1093 (2000)
Average
1088 (1990)
1095 (2002)
1092 (1998)
1096 (2006)
1163 (2125)
1017 (1863)
1103 (2018)
1098 (2008)
1087 (1988)
  'Test points 7 and 8 are replicates of test point 4; together the three tests provided the components of an IRF trial bum.
  • The 9-test averages of apparent scrub-
ber collection efficiencies for each of the
metals ranged from 22%to 71%; they were
generally higherforthe less-volatile metals.
The overall average collection efficiency for
all metals was 43%. Note, however, that the
IRF ionizing wet scrubber is a single-stage
unit; industrial applications of ionizing wet
scrubbers are typically in multiple stages
and, as such, would be expected to collect
metals more efficiently.
  • Apparent scrubber collection efficien-
ciesfor cadmium, bismuth, lead, and copper
increased with increased kiln temperature
and waste feed chlorine content. Afterburner
temperature had no discernible effect on
apparent scrubber collection efficiencies
for any of the metals.
  The full report was submitted in fulfill-
ment of Contract 68-C9-0038 by Acurex
Corporation under the sponsorship of the
U.S. Environmental Protection Agency.

-------

1
.5
1

!
I
1
Fraction in
f
1
•p
1
i
100
80
1 *°
*Q 40
*
20
0
C
Kin Ash
r
___

-

Cd
f* I 1 1 ^ 1
J ^ +
4- Cu


ft
t i i i i
) 200 400 600 800 1000 1200 1400 1600
Volatility Temperature fC)
100
80
1
1 *°
5 *
20
0
0
100 ,
80
H 60
I
5? 40
20
0
Scrubber-Exit Flue Gas
Cd
.

B
" m.

'AS
I ,
f
;
i* c.
4- t Cr
T f* , i , ,± M?,^
200 400 600 800 1000 1200 1400 1600
Volatility Temperature fC)
Scrubber Liquor


Bi
Cd
'As


Efe Cu _
4- a. Sr ,. Cr
JL T ^ Mp ±
              200      400      600      800     1000     1200     1400     1600
                                   Volatility Temperature fC)
Figure 2. Normalized distributions of metals in the discharge streams.

-------
                  Cadmium Dischargg Distributions
    70
            Kin Ash        Scrubber-Exit
                           RUB Gas
                   Bismuth Discharge Distributions
                         Liquor
            Kin Ash
      Scrubber-Exit
       Rue Gas
Lead Discharge Distributions
Liquor
            KinAsh        Scrubber-Exit         Liquor
                            Rus Gas
IUU
I so
.1
rj
I «
-Q
a 40
2
§
g ^
0
A
-_


—
-
• As






"" "

Bi .

-Cd

\


_Ba



• Pb



\
Sr .






19





-Co

-Cr


•



\
                                          500              1000              1500
                                                 Volatility Temperature (°C)

                     Figure 4 Apparent scrubber collection efficiencies lor  metals.
                                                                                                                                            2000

a-
'v 40
S
Cumulative Perct
%
0 n
As &
n g Cu %
; D
Sample
D
Cr
1 1 1
                                                                                               600             1000            1500
                                                                                                       Volatility Temperature fC)
                                                                                                                           2000
                                                                           Figures. Average distribution of metals of less than 10\un in afterburner-exit
                                                                                     flue-gas particle-size fractions.
Figure 3 Effects of Kin temperature on the discharge distributions
          of cadmium, bismuth, and lead.

-------
        100
        80
    s
     v
Kiln Temperature:
UAs ,.
3t j
•


0 *871°C (160CPF)
t nw, °927°C (170CPF)
m g

a

c
( t
i

<
pBa j
f

<
JCu t
• Sr<


1 1
p :

^
"Wp
f
^Sample
T *
., $Cr
                            500              1000              1500
                                    Volatility Temperature fC)
  2000
    o
    v
        100
        80
        60
        40
        20
Afterburner Temperature:
As A 9S2°C (1800PF)
, * 7093°C (2000PF)




Cd A ^ D 1204°C (2200»F)

j
a A
ii
* **
f j
IBa
.
'fti ^
t j
a-Jf ,
I
DD D a i
a
I ' i
'%
t
i
* Sample
« |
L
                           500              7000              7500
                                    Volatility Temperature fC)
 2000
       ao
  I"  *°
   V
       20
Feed Chlorine Content:
*As i

5 i
< •
fCd * 0%
*4%
1 ° 8%
*1
<
r (
* fia '
Q
<
(OU ^9 :
aj ,

li
i
j
	 1 	 i i J
i Cr
M Sample
f|

                          500               rooo              rsoo
                                   Volatility Temperature fC)
2000
Figure 6. Effects of kiln temperature, afterburner temperature, and waste feed chlorine content on
         the particle-size distribution of metals in the afterburner exit flue gas.
                                                              £u.S. GOVERNMENT PRINTING OFFICE: 1991 - 548-028/400)11

-------
    D. J. Fournier, Jr., andL R. Waterland are with Acurex Corp., Jefferson, AR 72079.
    /?. C. Thurnau 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 Single-stage Ionizing Wet Scrubber."
      " Volume I: Technical Results," (Order No. PB91-223 388; Cost: $23.00, subject
        to change).
      "Volume II: Appendices," (Order No. PB91'-223 396; Cost $31.00, subject to
        change).
    The above reports 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:
            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/S2-91/032

-------