&EPA
United States
Environmental Protection
Agency
EPA/540/S5-91/005
October 1992
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
Technology Demonstration
Summary
Horsehead Resource
Development Company, Inc.,
Flame Reactor Technology
Under the Superfund Innovative
Technology Evaluation (SITE) program,
the Horsehead Resource Development
Company, Inc., (HRD) Flame Reactor
was evaluated during a series of test
runs. The tests were conducted at the
HRD facility in Monaca, PA, using 72
tons of secondary lead smelter soda
slag (waste feed) from the National
Smelting and Refining Company, Inc.,
(NSR) site in Atlanta, GA. The waste
feed contained lead, zinc, iron, and
many other metals and inorganic com-
pounds. This summary includes an
overview of the demonstration, a tech-
nology description, analytical results,
and conclusions.
The HRD Flame Reactor technology
is a patented high-temperature thermal
process designed to safely treat in-
dustrial residues and wastes contain-
ing metals. The HRD Flame Reactor
processes wastes by subjecting them
to a hot (>2,000°C) reducing gas pro-
duced by the combustion of solid or
gaseous hydrocarbon fuels in oxygen-
enriched air. According to HRD, at these
temperatures, volatile metals in the
waste are vaporized, and any organic
compounds should be destroyed. The
waste materials react rapidly, produc-
ing a nonleachable slag and gases, in-
cluding steam and metal vapors. Metal
vapors further react and cool in the
combustion chamber and cooling sys-
tem, producing metal-enriched oxides
that are collected in the baghouse. The
resulting metal oxides potentially can
be recycled to recover the metals. The
HRD Flame Reactor was evaluated for
effectiveness in treating waste from the
NSR site to form a potentially recy-
clable, metal oxide product and a non-
hazardous, fused effluent slag.
During the demonstration, waste feed
from the NSR site produced a lead-
and zinc-enriched metal oxide product
and an effluent slag, which was deter-
mined to be nonhazardous, based on
extraction by the Toxicity Characteris-
tic Leaching Procedure (TCLP) and
chemical analysis of the extract. Greater
than 77.7% of the 5.41% weight lead
and 80.0% of the 0.416% weight zinc in
the waste feed were recovered in the
recyclable metal oxide product, which
contained 17.4% weight lead and 1.38%
weight zinc. The weight of the oxide
product and effluent slag was 36.6%
less than the weight of the waste feed.
Printed on Recycled Paper
-------
This Summary was developed by
EPA's Risk Reduction Engineering
Laboratory, Cincinnati, OH, to announce
key findings of the SITE Program dem-
onstration that Is fully documented In
two separate reports (see ordering in-
formation at back).
Introduction
In response to the Superfund Amend-
ments and Reauthorization Act of 1986
(SARA), the U.S. Environmental Protection
Agency (EPA) has established a formal
program to accelerate the development,
demonstration, and use of new or innova-
tive technologies that offer permanent,
long-term cleanup solutions at Superfund
sites. This program is called the Super-
fund Innovative Technology Evaluation, or
SITE, program and is administered by the
Office of Research and Development
(ORD).
The SITE program has four main goals:
• Identify and remove impediments
to the development and commer-
cial use of alternative technologies.
• Structure a development program
that nurtures emerging technolo-
gies.
Conduct a demonstration of the
more promising innovative technolo-
gies to establish reliable perfor-
mance and cost information for site
characterization and cleanup deci-
sion-making.
• Develop procedures and policies
that encourage the selection of
available alternative treatment rem-
edies at Superfund sites and other
waste sites and commercial facili-
ties.
Under the SITE program, EPA solicits
proposals from developers of innovative
waste treatment technologies who have
expressed an interest in participating in
the SITE program. Based on these pro-
posals, EPA selects technologies for the
demonstration portion of the SITE pro-
gram. One of the selected technologies
was the Flame Reactor developed by
HRD.
The HRD Flame Reactor SITE Demon-
stration took place at the HRD facility in
Monaca, PA, using secondary lead smelter
soda slag from the NSR site in Atlanta,
GA, as the waste feed. The HRD technol-
ogy involves a high-temperature metals
recovery process that produces a poten-
tially recyclable metal oxide product and a
nonhazardous, based on the RCRA Tox-
icity Characteristic (TC) rule, effluent slag.
The primary objectives of the HRD
Rame Reactor SITE Demonstration in-
cluded the following:
• Evaluate the technology's ability to
treat waste materials to form a re-
cyclable metal oxide product and
a nonhazardous, fused effluent slag
Evaluate the system's reliability
Develop overall economic data on
the technology
Secondary objectives were also defined.
These were objectives that would be of
interest to potential technology users but
concerned testing auxiliary systems rather
than the actual Flame Reactor. Second-
ary objectives included the following:
• Assess airborne emissions from the
process
• Verify the predictions of the HRD
thermodynamic operating model so
that it can be used to predict costs
for other projects
Overview of the HRD Flame
Reactor SITE Demonstration
The HRD Flame Reactor SITE Demon-
stration took place in February and March
1991. The waste material from the NSR
site consisted of granular slag containing
arsenic, cadmium, iron, lead, sodium, zinc,
and other metals, plus carbon, chlorine,
silicon, sulfur, oxygen, and other inorganic
compounds. This waste material was
chosen as it was readily available, it con-
tained high concentrations of several re-
coverable metals (lead and zinc), it con-
tained no organic compounds (which could
not be treated under HRD's state per-
mits), and it was representative of a waste
type available in large quantities through-
out the country. The waste material was
dried and passed through a hammermill
before treatment in the HRD Flame Reac-
tor. The demonstration test runs included
(1) a series of shakedown runs to estab-
lish optimal operating conditions, (2) a
blank run with no waste feed, (3) four test
runs (including one that was not used for
interpretation of results because of opera-
tional problems), and (4) a series of addi-
tional runs to process remaining waste
material and to try to improve the struc-
tural integrity of the effluent slag.
Process operating data and analytical
samples were collected. The operating
data included (1) waste feed consumption
rate, (2) oxide product and effluent slag
production rates, (3) natural gas and oxy-
gen consumption, (4) electrical consump-
tion, (5) temperatures throughout the sys-
tem, and (6) flow rates throughout the
system. Contaminant concentrations were
measured in the waste feed, oxide prod-
uct, the effluent slag, and the stack gases.
The waste feed was analyzed for energy
content, ash content, moisture, metals,
sulfur, chloride, fluoride, carbon, and total
organic carbon content. The oxide prod-
uct and effluent slag were analyzed for
metals. The waste feed and effluent slag
were also extracted by the TCLP, and the
extracts were analyzed for metals. Stack
gases were analyzed for carbon monox-
ide, carbon dioxide, oxygen, nitrogen ox-
ides, sulfur dioxide, total hydrocarbons,
hydrogen chloride, particulate, and met-
als.
Technology Description
Figure 1 presents the process flow dia-
gram for the HRD Flame Reactor pro-
cess. The process consists of a (1) feed
system, (2) Flame Reactor, (3) slag sepa-
rator, (4) combustion chamber, and (5)
oxide product recovery system. The feed
system operations include fuel and waste
feed storage and handling, metering and
injecting of waste, fuel, oxygen, and air
into the Flame Reactor.
The Flame Reactor is a two-stage sys-
tem. The first stage consists of a fuel
burner system composed of two separate
burners; the second stage consists of a
metallurgical reactor. Carbon-based com-
bustion and gasification and metal oxide
smelting reactions occur in the two-stage
reactor system. The Flame Reactor is 15
ft tall, positioned vertically, with an internal
diameter of 23 in.
Materials passing through the reactor
discharge continuously into the slag
separator, which separates molten effluent
slag from reactor off-gases. The slag
separator is positioned horizontally, with a
slight upward angle between the reactor
and the combustion chamber. The gases,
particulate, and metal vapors flow toward
the combustion chamber, countercurrent
to the effluent slag. The molten effluent
slag runs out through a tap hole on the
discharge end of the unit.
The reactor off-gases are reacted with
air in a refractory-lined combustion cham-
ber. The temperature of the combustion
chamber gases is typically between 600
and 800 °C.
The oxide product recovery system is
designed to cool the gas stream and cap-
ture the metal oxides formed in the com-
bustion chamber. The gas is cooled by a
shell-and-tube heat exchanger and by the
addition of ambient air. The oxide product
recovery system includes a jet-pulsed
baghouse designed to recover oxide
product from the gas stream. The bag-
house emits off-gas through a stack and,
when pulsed, discharges the oxide product
into enclosed bags for recovery. A fan
between the baghouse and the stack pro-
vides an induced draft for the system.
-------
Waste Feed
Combustion Air
Compressor
Oxygen
Stack
Gas
Exhaust
Effluent
Slag
Oxide Product
Figure 1. HRD Flame Reactor process flow schematic.
The HRD Flame Reactor is designed to
thermally treat granular solids, soil, flue
dust, slag, and sludge containing metals.
After entering the reactor, the waste feed
reacts in less than 0.5 sec, allowing high
waste throughput. The treatment process
yields two products: a metal oxide that
may potentially be recycled and a
nonleachable, nonhazardous effluent slag
that potentially may be used as aggre-
gate.
Volatile metals in the waste material,
such as cadmium, lead, and zinc, are va-
porized and oxidized, then captured down-
stream in an oxide product collection sys-
tem. Nonvolatile metals are predominantly
encapsulated in the effluent slag. Accord-
ing to HRD, for optimal reaction condi-
tions, feed particles should contain less
than 5% weight total moisture, and at least
80% weight of the feed should be sized
smaller than 200 mesh (0.0029 in. or 75
microns). Waste material may be pre-
treated by drying and by physical size re-
duction to meet these specifications. The
fusion temperature of feed materials should
not exceed 1,400 °C. Deviations from these
specifications are acceptable but tend to
decrease throughput and reduce the re-
covery of metals in the oxide product.
The waste feed, after drying and size
reduction, is transferred to portable stor-
age bins. It is then transferred by a tubu-
lar drag conveyor system to the day bins.
From the day bins the waste feed is me-
tered and pneumatically transferred to the
HRD Flame Reactor, where it is heated to
high temperature by the fuel-rich combus-
tion of natural gas or coal and oxygen-
enriched air. Volatile metals, water, car-
bonates, sulfates, and other volatile inor-
ganic compounds are vaporized; organic
compounds and carbon are burned. Non-
volatile and noncombustible materials are
fused into slag by the high temperatures
and fall through the reactor into the hori-
zontal slag separator. Fused, effluent slag
exits through the slag tap and then cools.
Volatilized matter is drawn by reduced
pressure into a combustion chamber,
where air is. introduced and oxidation oc-
curs. The oxidized gases are further copied
in a heat exchanger, and the metal oxides
are collected in an oxide product collec-
tion system. The metal oxide product from
this collection system is periodically re-
moved by a jet-pulse and transferred by
auger to enclosed oxide product storage
bags for recycling.
Demonstration Results
During the HRD Flame Reactor SITE
Demonstration, a comprehensive sampling
and analysis program was undertaken to
characterize the waste feed, the oxide
product, the effluent slag, and stack gas
emissions. A mass balance was performed
to account for distribution of the waste
feed into the Flame Reactor products and
to determine the percent recovery of met-
als and the weight reduction of the waste
-------
feed. The HRD Rame Reactor Demon-
stration also assessed the Flame Reactor's
operational reliability and the costs of
waste treatment. The following discussion
Is broken down into three sections: (1)
sampling and analytical results, (2) mass
balance analysis, and (3) operational reli-
ability and costs.
Sampling and Analytical
Results
The HRD Flame Reactor SITE Demon-
stration included a comprehensive sam-
pling and analysis program that determined
the following:
• Constituents and their concentra-
tions In the waste feed, oxide prod-
uct, and effluent slag
• TCLP results
• Stack monitoring and emissions
sampling
Each of these items Is discussed below.
Constituent Analyses
Constituent analyses were performed
on the waste feed, oxide product, and
effluent slag to determine if the technology
produced a potentially recyclable oxide
product enriched in lead and a
nonhazardous effluent slag product.
Table 1 presents the constituent analy-
ses data for the waste feed, the oxide
product, and the effluent slag. The data
show clearly that volatile metals, such as
lead, cadmium, and zinc, are concentrated
in the oxide product, while nonvolatile
metals, such as aluminum, calcium, iron,
are concentrated in the effluent slag. The
oxide product contains some nonvolatile
species, because some unreacted feed
and effluent slag particles are entrained
with the off-gas stream.
The HRD Flame Reactor technology
produced an oxide product enriched in
lead, cadmium, and zinc. Table 1 com-
pares the waste feed, effluent slag, and
oxide product compositions. Comparison
of the waste feed concentration to the
oxide product concentration for lead
(5.41% to 17.4% weight), cadmium
(0.0411% to 0.128% weight), and zinc
(0.416% to 1.38% weight) indicates a sig-
nificant partitioning of these volatile metals
to the oxide product.
The main constituents of the effluent
slag are iron (20.4% weight), sodium
(15.5% weight), aluminum (1.53% weight),
and calcium (1.30% weight). HRD reports
that silicon is present in the effluent slag
at an average concentration of 10.2%
weight. In general, the effluent slag is
composed of the oxides of nonvolatile
metals such as iron, calcium, and alumi-
num. Silicon and sodium appear in both
the oxide product and the effluent slag.
TCLP Results
TCLP tests were performed on the
waste feed and on the effluent slag. Table
2 presents the mean concentrations and
ranges for all TCLP results and the ap-
propriate Resource Conservation and Re-
covery Act (RCRA) regulatory limits for
each waste code.
The waste feed processed by the HRD
Flame Reactor was a RCRA characteris-
tic hazardous waste because lead (RCRA
waste code D008) and cadmium (RCRA
waste code D006), when extracted by the
TCLP procedure, leach above the RCRA
TC rule limit. Lead leached at an average
of 5.75 mg/L, compared with the RCRA
TC rule limit of 5.0 mg/L. Cadmium was
well above the RCRA TC rule limit of 1.0
mg/L, leaching at an average of 12.8 mg/
L. The other 'metal extracts were well be-
low the RCRA TC rule limits for character-
istic wastes. No organic compounds were
present in the waste feed. Table 2 presents
the TCLP results and the RCRA TC rule
limits for comparison.
TCLP extraction and analysis was not
performed on the oxide product because
this product is intended for recycling.
TCLP testing determined that the efflu-
ent slag was not a RCRA characteristic
waste. Cadmium, chromium, lead, mer-
cury, and silver concentrations in the TCLP
Table 1. Composition of the Waste Feed, Effluent Slag, and Oxide Product
Waste Feed '
Analyte (% Weight)
Aluminum 0.596 (0.490-0.787)
Antimony 0.0373 (0.0278-0.0455)
Arsenic 0.0515(0.0428-0.104)
Barium 0.0861 (0.0804-0.0940)
Beryllium <0.00011
Cadmium 0.0411(0.0356-0.0512)
Calcium 0.653 (0.552-0.835)
Chromium" 0.00877(0.00631-0.0113)
Copper 0. 185 (0. 146-0.259)
Iron 10.3 (9.56-13.0)
Lead 5.41(4.82-6.17)
Magnesium 0.228(0.163-0.346)
Manganese 0.0753 (0.0672-0.0903)
Mercury 0.000068
(0.000054-0.000087)
Potassium 0.244 (0.204-0.284)
Selenium 0.00727 (0.00400-0.0175)
Silicon 3 0.276 (0. 1 76-0.444)
Silver 0.000339
(0.000160-0.000540)
Sodium 12.2(1 1.5- 13.2)
Thallium 0.0253 (0.0181-0.0317)
Tin 0.282 (0.261-0.314)
Zinc 0.416 (0.321-0.681)
Carbon 15.0 (9.56-19.6)
Chloride 2.46(2.12-2.89)
Fluorine as Fluoride 0.0130
(0.0106-0.0166)
Sulfur 5.25 (4.77-6.44)
Moisture 3.35 (2.26-4.07)
Ash 81.6 (80.6-82.4)
Effluent Slag '
(% Weight)
1.53 (1.33-1.85)
0.0357(0.0100-0.190)
0.0262 (0.00921-0. 134)
0.165(0.139-0.183)
0.000101
(<0.000087-0.000110)
0.000373
(<0.00023-0.00135)
1.30 (1.06-1.45)
0.00890 (0.00339-0.0385)
0.344 (0.273-0.389)
20.4 (16.7-22.8)
0.552 (0.156-1.14)
0.543 (0.441-0.761)
0.175(0.132-0.231)
<0.000010
0.238(0.199-0.269)
0.00344 (<0.00226-0.0176)
0.327 (0. 183-0.525)
0.000394
(0.000250-0.000510)
15.5 (12.8-16.8)
0.0689 (0.0535-0.0852)
0.0796(0.0544-0.111)
0.113(0.0709-0.168)
NA
NA
NA
NA
NA
NA
Oxide Product *
(% Weight)
0.0562 (0.0459-0.0623)
0.125(0.122-0.131)
0.110 (0.101-0.117)
0.0282 (0.0248-0.0323)
<0.00010
0.128(0.108-0.138)
0.202(0.155-0.234)
0.0300 (0.0278-0.0312)
0.161 (0.138-0.178)
3.22 (2.91-3.56)
17.4 (15.9-18.4)
0.0327 (0.0266-0.0368)
0.0265 (0.0214-0.0300)
0.000013
(<0.000010-0.000014)
0.707(0.630-0.751)
0.00520 (0.00415-0.00659)
0.127(0.113-0.137)
0.00269
(0.00190-0.00342)
15.7(13.7-16.8)
0.00746 (0.00714-0.00773)
0.660 (0.612-0.687)
1.38 (1.00-1.62)
NA
NA
NA
NA
NA
NA
'Average of 18 values; the range is shown in parentheses.
'Average of 3 values; the range is shown in parentheses.
3Due to matrix interferences, analytical results are known to be lower than actual concentrations
for the waste feed and effluent slag. When analyzed by HRD, chromium levels were, on average,
0.024% in the waste feed and 0.040% in the effluent slag. Silicon levels detected by HRD were,
on average, 8.10% in the waste feed and 10.2% in the effluent slag.
NA = Not analyzed. < = less than.
When an analyte was not detected, the detection limit was used in calculating the average value.
-------
Table 2. TCLP Results of Waste Feed and Effluent Slag
Analyte
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Waste Feed '
(mg/L)
0.213
(<0.210-0.264)
0.0347
(0.0177-0.0675)
12.8
(7.61-15.8)
0.184
(0.140-0.283)
5.75
(4.35-6.80)
<0.010
(<0.010)
0.0716
(<0.030-0.160)
<0.050
(<0.050)
Effluent Slag '
(mg/L)
0.474
(<0.210-0.930)
0.175
0.109-0.281)
<0.050
(<0.050)
<0.060
(<0.060)
<0.330
(<0.330)
<0.010
(<0.010)
0.0326
(<0.030-0.073)
<0.050
(<0.050)
RCRA TC Rule
Limits
(mg/L)
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
RCRA
Waste Code
D004
D005
D006
D007
D008
D009
D0 10
D011
1 Average of 18 values; the range is shown in parentheses.
mg/L = milligrams per liter.
< = less than.
extracts of the effluent slag were below
the method detection limits of 0.050, 0.060,
0.330, 0.010, and 0.050 mg/L, respec-
tively. Selenium was below the detection
limit (0.030 mg/L) for all but two samples,
where the TCLP extract concentrations
were 0.0338 and 0.0730 mg/L. Arsenic
and barium were consistently above the
detection limit, with concentrations rang-
ing from 0.210 to 0.930 mg/L for arsenic
and from 0.109 to 0.281 mg/L for barium.
Concentrations for all metals were well
below the RCRA TC rule limits. Addition-
ally, the waste feed was not a listed waste.
Consequently, the effluent slag from the
demonstration can be disposed of in a
nonhazardous waste (Subtitle D) landfill.
Stack Monitoring and Emissions
Sampling
During the HRD Flame Reactor SITE
Demonstration, stack gases were sampled
for metals, hydrogen chloride gas, and
particulate emissions, and were continu-
ously monitored for sulfur dioxide, nitrogen
oxides, oxygen, carbon dioxide, carbon
monoxide, and total hydrocarbons. The
metals and particulate emissions were
sampled using an EPA Modified Method
5, isokinetic, multiple metals sampling train.
Hydrogen chloride gas emissions were
determined by a single point EPA Method
26 sampling train. The continuous emis-
sion monitors used the following: EPA
Method 6C for sulfur dioxide, EPA Method
7E for nitrogen oxides, EPA Method 3A
for oxygen and carbon dioxide, EPA
Method 10 for carbon monoxide, and EPA
Method 25A for total hydrocarbons. All
the standard EPA methods can be found
in 40 Code of Federal Regulations (CFR)
60, Appendix A, and the multiple metals
train is discussed in the Methods Manual
for Compliance with the Boiler and Indus-
trial Furnace (BIF) Regulations [40 CFR
266, Appendix IX]. Emission results for
metals, hydrogen chloride gas, particu-
late, and continuous emissions monitoring
are discussed below.
Hydrogen chloride gas emissions dur-
ing the HRD Flame Reactor SITE Demon-
stration were between 38.5 and 46.4 Ib/
hr. This high emission rate occurred be-
cause the Flame Reactor had no acid gas
control system, and the waste feed was,
on average, 2.46% chloride by weight.
The BIF rule has promulgated risk-based
emission limits on hydrogen chloride gas.
The addition of a wet scrubber should
control hydrogen chloride gas emissions
to below the applicable standards.
Because the HRD Flame Reactor pro-
cess uses a baghouse to capture the metal
oxide product, particulate emissions from
the Flame Reactor are low when the
baghouse is maintained and operated
properly. During analysis of the demon-
stration samples, problems occurred with
the gravimetric analysis, preventing accu-
rate determination of the particulate emis-
sions. Thus, although no particulate data
were obtained .during the demonstration,
the Flame Reactor, when equipped with a
state-of-the-art baghouse (emission con-
trol system), should have more effective
particulate emission control.
Emissions of sulfur dioxide, nitrogen ox-
ides, oxygen, carbon dioxide, and total
hydrocarbons were continuously monitored
for the blank run and for each test run.
The HRD Flame Reactor currently has an
air quality permit issued by Pennsylvania
Department of Environmental Resources
that limits sulfur dioxide emissions to less
than 500 ppm for commercial operations.
During the HRD Flame Reactor SITE
Demonstration test, the sulfur dioxide
emissions were below 500 ppm except for
a 2-minute period during one run immedi-
ately following system startup, after a
shutdown was required to cool the off-gas
systems. During this 2-minute period the
maximum sulfur dioxide emission was 514
ppm.
Mass Balance Analysis
The HRD Flame Reactor SITE Demon-
stration included a mass balance analysis,
which calculated (1) weight reduction of
'the waste feed and (2) percent recovery
of metals. Each of these items is dis-
cussed below.
The HRD Flame Reactor reduced the
weight of the waste feed by 36.6% (that
is, the effluent slag and oxide product
weighed 36.6% less than the amount of
waste feed). The weight of the waste feed
was reduced because carbon was essen-
tially completely converted to carbon di-
oxide, moisture was converted to steam,
chloride was converted to hydrogen chlo-
ride gas, and sulfur was partially converted
to sulfur dioxide.
The metal recoveries, when calculated
based on concentrations in the waste feed
and oxide product, were less than 100%.
The mass balance closure for the demon-
stration was also less than 100%. These
values were low because of residual ma-
terial buildup in the combustion chamber
and heat exchanger. For lead, zinc, and
cadmium, these percent recoveries are
77.7, 80.0, and 75.0, respectively. The
actual percent recoveries of lead, zinc,
and cadmium are expected to be higher
and may range from 90% to 99% weight.
The particle size distribution (PSD) of
the waste feed and the brief residence
time in the reactor (between 0.1 and 0.5
sec) affect the kinetics of the treatment
reactions. For the demonstration, 66.6%
by weight of the waste feed particles were
smaller than 200 mesh. This PSD yielded
a 77.7% weight recovery of lead.
HRD Flame Reactor Operational
Reliability and Treatment Costs
The HRD Flame Reactor SITE Demon-
stration included an analysis of the Flame
Reactor's operational reliability and treat-
ment costs. Both of these items are dis-
cussed below.
•U.S. Government Printing Office: 1992 — 648-080/60130
-------
Information collected on the reliability of
tha HRD Flame Reactor during the dem-
onstration revealed that the HRD Flame
Reactor had no major operational prob-
lems; however, auxiliary systems, such as
the oxide product collection system, cool-
ing water system, and feed system, expe-
rienced problems that did not affect the
operation of the Flame Reactor but im-
pacted the overall system performance.
The oxide product collection system,
consisting of a shell-and-tube heat ex-
changer, a baghouse, an induced draft
fan, and a stack, was undersized for the
demonstration. The Flame Reactor was
sized to handle 20,000 tons/yr of electric
arc furnace (EAF) dust, but the off-gas
handling system was put together from
surplus zinc smelter parts. Due to deterio-
ration of those used parts, the off-gas
handling system presently cannot handle
the volume of gas generated from pro-
cessing 20,000 tons/yr of EAF dust. The
operating conditions required for the
demonstration produced high off-gas vol-
umes, and the Flame Reactor system was
typfeally shut down after about 4 hr of
operation because the oxide product col-
lection system was undersized. For a
commercial operation, the oxide product
collection system would include a larger
baghouse and a higher capacity induced
draft fan to introduce a large volume of
cooling air. Because of this addition, the
existing heat exchanger would not be re-
quired.
The cooling water system also devel-
oped problems. The supply line to the
shall-and-tube heat exchanger developed
an underground leak, and makeup water
was added to the cooling tower. This
problem did not affect the operation of the
reactor and would not occur during com-
mercial operation because the heat ex-
changer would not be used.
During a test run, one of the surge
hopper screw feeders in the feed system
jammed. For approximately 30 min, the
other day bin was used at twice the nor-
mal capacity to keep the waste feed rate
constant. The operation was not adversely
affected.
The estimated cost per ton for treating
secondary lead smelter soda slag ranged
from $208 to $932. A 50,000 tons/yr waste
treatment scenario cost $208 per ton and
included a more efficient waste' pretreat-
ment system than presently exists at the
HRD facility; the SITE Demonstration test
scenario cost $932 per ton. The estimated
costs of the HRD Flame Reactor system
are highly site-specific and rather difficult
to identify without accurate data from a
site remedial investigation report or waste
profile. Variability in the waste character-
istics and the costs of transporting waste
to the HRD Flame Reactor, as well as the
costs of transporting, shipping, and han-
dling residuals, could significantly affect
costs presented in this economic analysis.
Costs presented are order-of-magnitude
estimates. A more detailed discussion of
the economics of this technology is pre-
sented in the HRD Applications Analysis
Report.
Quality Assurance Procedures
The primary quality assurance objective
of this and all SITE demonstrations is to
produce well-documented sampling and
analytical data of known quality. To ac-
complish this goal, a detailed and com-
prehensive Quality Assurance Project Plan
(QAPP) was developed before the dem-
onstration. This QAPP contained specific
quality assurance targets for precision,
accuracy, -completeness, representative-
ness, and comparability. It also specified
the (1) analytical methods to be used, (2)
holding times, (3) number and type of
blanks, (4) matrix spikes and matrix spike
duplicates, (5) laboratory duplicate
samples, (6) reference standards, and (7)
method detection limits.
The waste feed and effluent slag from
the reactor are both nonhomogeneous and
composed of a matrix that is difficult to
digest and analyze for metals. Therefore,
a study was undertaken to select the best
digestion method for determining metals
in these matrices. Based on this study, a
modification of EPA SW-846 Method 3050
using a reduced sample size was chosen.
However, when this method is used, the
results are known to be poor for the di-
gestion of silicon and for the digestion of
chromium in a high silicon content matrix.
Conclusions
Based on the results of the HRD Flame
Reactor SITE Demonstration, the follow-
ing conclusions can be made concerning
the performance of HRD Flame Reactor
technology:
The HRD Flame Reactor technol-
ogy processed secondary lead
smelter soda slag from the NSR
site and produced both a poten-
tially recyclable metal oxide product
and an effluent slag meeting RCRA
TC rule criteria.
Although the Flame Reactor stack
emissions were monitored, a site-
specific risk analysis is required to
assess the impact of these stack
emissions. Such an analysis was
outside of the scope of this report.
The atmospheric emissions of met-
als from the Flame Reactor could
be a concern, however, due to data
limitations, no conclusions could be
reached on metal emissions.
The HRD Flame Reactor achieved
a net weight reduction of 36.6% of
the waste feed when processed into
oxide product and effluent slag.
During the demonstration, the HRD
Flame Reactor had no major op-
erational problems; however, auxil-
iary systems such as the oxide
product collection system, cooling
water system, and feed system ex-
perienced problems that did not af-
fect the operation of the Flame Re-
actor. HRD agrees with EPA that
these systems require refinement.
The HRD thermodynamic model
can be used to set preliminary op-
erating conditions and to determine
order of magnitude estimates for
parameters used in a cost estimate,
such as fuel and oxygen flow rates.
The HRD Flame Reactor system pro-
cessed secondary lead smelter soda
slag from the NSR site at a cost of
$932/ton for the demonstration. This
cost included extensive testing. Data
from HRD for similar applications show
that the HRD Flame Reactor can pro-
cess this waste for $208/ton in com-
mercial operation.
-------
-------
The EPA Project Managers, Maria K. Richards and Donald A. Oberacker,
are with the Risk Reduction Engineering Laboratory, Cincinnati, OH 45268
(see below).
The complete report entitled "Technology Evaluation Report: Horsehead
Resource Development Company, Inc., Flame Reactor Technology,"
(Order No. PB92-205 855/AS; Cost: $26.00, subject to change) discusses
the results of the SITE demonstration. This report will be available only
from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
A related report entitled "Application Analysis Report: Horsehead Resource
Development Company, Inc., Flame Reactor Technology," (EPA/540/
A5-91/005), discusses the applications of the demonstrated technology.
The EPA Project Managers 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
Official Business
Penalty for Private Use
$300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/540/S5-91/005
------- |