EMERGING TECHNOLOGY REPORT:
RECLAMATION OF LEAD FROM SUPERFUND WASTE MATERIAL USING
SECONDARY LEAD SMELTERS
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
NOTICE
This material has been funded wholly or in part by the United States Environmental Protection Agency
under Contract CR-818199-01-0 to the Center for Hazardous Materials Research. It has been subject to
the Agency's review and it has been approved for publication as an EPA document. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
-------
FOREWORD
The Superfund Innovative Technology Evaluation (SITE) Program was authorized in the 1986 Superfund
Amendments. The program is a joint effort between EPA's Office of Research and Development and Office
of Solid Waste and Emergency Response. The purpose of the program is to assist the development of
hazardous waste treatment technologies necessary to implement new cleanup standards which require
greater reliance on permanent remedies. This is accomplished through technology demonstration designed
to provide engineering and cost data on selected technologies.
The Risk Reduction Engineering Laboratory (RREL) is responsible fpr planning, implementing and
managing research, development, and demonstration programs to provide an authoritative, defensible
engineenng basis for support of the policies, programs, and regulations of the EPA. This publication is one
of tne products of that research and provides a vital communication link between the research and the user
community.
This project cpnsisted of demonstration of the use of secondary lead smelters to reclaim lead from a variety
of materials, including materials found at Superfund sites as well as other lead-containing wastes. The
demonstration consisted of feeding materials from five sources - three Superfund sites, and two other
sources, to a secondary lead smelter to determine if the smelter could economically reclaim the lead.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-------
ABSTRACT
There are over 3,000 sites across the United States contaminated wtth lead. Techniques to remediate
these sites include standard stabilization/disposal technologies, reclamation of lead using secondary lead
smelters, soil washing, and biological removal technologies.
Through a Cooperative Agreement with the U.S. Environmental Protection Agency's Risk Reduction
Engineering Laboratory, the Center for Hazardous Materials Research CHMR), in conjunction with a major
secondary lead smelter, has demonstrated that secondary lead smelters may be used economically to
reclaim lead from a wide range of lead-containing materials frequently found at Superfund sites. Such
materials include battery case materials, lead dross, and other debris containing between 3 and 70 percent
During the study, CHMR and the smelter reclaimed lead from five sets of materials, including two
Superfund sites containing primarily battery cases, and one battery breaker/smelter site with a variety of
lead-containing materials. Between 4 and 1600 tons of materials from each of these sites were excavated
and processed at the smelter, while the research team assessed the effects on furnace operation and
performance. Two additional sets of materials, one from the demolition of a house containing lead-based
paint, and the other consisting of blasting abrasive material from work on a bridge coated wifn lead paint,
were also processed in the smelter. The results showed that it was technically feasible to use the
secondary ead smelter to recjaim lead from all of the materials. CHMR considered the use of materials
from a sixth site, which contained approximately 80% soil and 20% waste battery cases. The site was
dropped from consideration for the study, because the smelter could not accept materials with such a high
soils concentration
CHMR also assessed the economics of using secondary lead smelters to reclaim lead from Superfund
sites, and developed a method for estimating the cost of reclaiming lead. This method develops cost as
a function of material excavation, transportation and processing costs combined with cost benefits received
by the smelter (in the form of recovered lead, reduced fuel usage and/or reduced iron usage). The total
remediation costs using secondary lead smelters for the sites and materials studied varied between $35
and $374 per ton, based on January 1994 market prices for lead. The costs were primarily a function of
lead concentration, the market price for lead, distance from the smelter, and the amount of materials which
become incorporated into slag from the process, although other factors affected the economics as well.
Materials with high concentrations of lead were significantly less expensive to remediate than those with
bw concentrations. The cost to remediate materials which left few slag residues in the furnace was lower
than the cost to remediate materials which contained a significant amount of material that remained in the
slag.
-------
TABLE OF CONTENTS
Page
NOTICE ii
FOREWORD iii
ABSTRACT jv
LIST OF TABLES vii
LIST OF FIGURES vii
ACKNOWLEDGEMENTS viii
1 . 0 INTRODUCTION 1
1.1 SOURCES OF LEAD CONTAMINATION 1
1.1.1 Lead Contaminated Soils 2
1.1.2 Battery Breaker Site Materials 2
1.1.3 Integrated Breaker/Smelter Site Materials 2
1.1.4 Other Lead-Containing Materials 2
1.2 AVAILABLE REMEDIATION TECHNOLOGIES 2
1.2.1 Solidification/Stabilization 3
1.2.2 Capping 3
1.2.3 Washing 3
2.0 LEAD RECLAMATION USING SECONDARY LEAD SMELTERS 4
2.1 TYPICAL SMELTING OPERATIONS 4
2.2 USE OF SECONDARY SMELTERS IN LEAD
RECLAMATION FROM WASTE MATERIAL 4
2.3 EVALUATION METHODOLOGY, SITES, AND MATERIALS 5
2.3.1 Acquisition of Materials 5
2.3.2 Smelter Furnaces 9
2.4 DATA AND SAMPLE COLLECTION 10
2.5 QUALITY ASSURANCE/QUALITY CONTROL 10
3.0 RESULTS AND DISCUSSION 13
3.1 PRODUCTION/PARAMETERS 13
3.1.1 Lead Production Ratio 13
3.1.2 Lead Loss Ratio 14
3.1.3 Slag Disposal Ratio 14
3.2 FEED RATIOS OF SUPERFUND MATERIALS TO TOTAL FEED 14
3.3 BENEFICIAL EFFECTS OF CERTAIN MATERIALS ON
FURNACE PERFORMANCE 15
3.4 APPLICABILITY OF LEAD RECOVERY FROM
WASTE MATERIALS 16
3.5 LEAD RECLAMATION EFFICIENCY 17
4.0 LEAD RECLAMATION ECONOMICS 19
4.1 ON-SITE EXCAVATION 19
4.2 TRANSPORTATION 19
-------
5.0
6.0
7.0
4.3
4.4
TABLE OF CONTENTS (continued)
PROCESSING AT THE SMELTER
Page
19
4.3.1
4.3.2
4.3.3
4.3.4
Additional Production Costs C
Base Cost CB 20
' 20
20
20
-PS.' ' '
Additional Disposal Costs Cdjsp
Offsets for the Value of Lead and
Reduction of Other Furnace Feeds.
4.3.5 NET Smelter Processing Cost 21
OVERALL PROCESS ECONOMICS
CONCLUSIONS
RECOMMENDATIONS FOR LEAD RECLAMATION
AS A REMEDIAL APPROACH
REFERENCES
22
23
24
25
APPENDIX A RECLAMATION OF MATERIALS FROM BATTERY CASE PILES AT THE TONOLLI
SUPERFUND SITE IN NESQUEHONING, PA
APPENDIX B RECLAMATION OF LEAD FROM THE NL INDUSTRIES SUPERFUND SITE IN
PEDRICKTOWN, PA
APPENDIX C THE USE OF SECONDARY SMELTING TECHNOLOGY TO RECLAIM LEAD FROM
IRON-SHOT BRIDGE BLASTING MATERIAL
APPENDIX D THE USE OF SECONDARY LEAD SMELTERS TO RECLAIM LEAD FROM BATTERY
CASE MATERIALS FROM THE HEBELKA SITE
APPENDIX E THE USE OF SECONDARY LEAD SMELTERS TO RECLAIM LEAD FROM DEMOLITION
MATERIALS FROM A HUD RENOVATION
VI
-------
LIST OF TABLES
Page
Table Title
1 SUMMARY OF THE EVALUATIONS. 8
2 INPUT, OUTPUT, AND OPERATING PARAMETERS 10
3 METHODS USED TO MEASURE TEST PARAMETERS 11
4 LEAD AND SLAG PRODUCTION COMPARISONS 13
5 REQUIRED CHARACTERIZATION PARAMETERS 17
6 COST OF REMEDIATING SITES 22
7 COMPARISON OF SMELTING AND OTHER TECHNOLOGIES 24
A-1 ANALYTICAL RESULTS FROM TEST MATERIALS A-2
A-2 SUMMARY OF EFFECTS ON BLAST FURNACE PERFORMANCE A-6
A-3 BLAST SLAG ANALYTICAL RESULTS REGULAR FEED ONLY A-10
B-1 DESCRIPTION OF FEED MATERIAL UPON ARRIVAL AT EXIDE B-17
B-2 NL INDUSTRIES FEED MATERIALS ANALYSIS B-18
B-3 REVERBERATORY LEAD FROM TEST MATERIAL VERSUS TYPICAL
FEED LEAD FOR JANUARY 1992 PRELIMINARY INVESTIGATION B-19
B-4 ANOVA ANALYSIS OF BLAST SLAG PRODUCTION B-20
C-1 PENNDOT BLASTING MATERIAL C-2
C-2 TEST MATERIAL BLAST LEAD ANALYSIS C-5
C-3 ANALYSIS OF SLAG SAMPLES C-6
C-4 TEST BLAST FURNACE OPERATING PARAMETERS C-6
LIST OF FIGURES
Figure Title
1 SCHEMATIC OF SMELTING PROCESS 5
2 SCHEMATIC OF RECLAMATION PROCESS. 6
3 PERCENT LEAD IN TEST MATERIAL VERSUS FEED RATIO 15
B-1 TEST FURNACE VS. CONTROL FURNACE SO2 EMISSIONS B-10
B-2 ANALYSIS OF AUGUST REVERBERATORY LEAD PRODUCTION B-11
B-3 ANALYSIS OF AUGUST REVERBERATORY SLAG PRODUCTION B-12
B-4 ANALYSIS OF SEPTEMBER REVERBERATORY LEAD PRODUCTION B-13
B-5 ANALYSIS OF SEPTEMBER REVERBERATORY SLAG PRODUCTION B-14
B-6 ANALYSIS OF OCTOBER BLAST FURNACE LEAD PRODUCTION B-15
B-7 ANALYSIS OF OCTOBER BLAST FURNACE SLAG PRODUCTION B-16
VII
-------
ACKNOWLEDGEMENTS
The Center for Hazardous Materials Research (CHMR) would like to acknowledge the cooperation received
throughout the project from the U.S. Environmental Protection Agency, Office of Research and
Development. In particular, we would like to thank Ms. Laurel Staley, from the EPA's Superfund
Technology Demonstratbn Division who provided helpful guidance on the project. The authors also
acknowledge the contributions Mr. Patrick Augustin in developing and overseeing the early stages of the
project and Mr. Michael Royer, who provided many helpful comments to the original draft manuscript.
CHMR would like to acknowledge the following individuals from the research team who were responsible
for important contributions to this project.
Center For Hazardous Mater/a/s Research
Mr. Stephen W. Paff Project Manager
Mr. Brian Bosibvich Project Engineer
Mr. Mark Ulintz Environmental Specialist
Exide Corporation
Mr. Howard Master Plant Manager
Mr. Steven Timm Engineer
Mr. John Baranski Vice President, Environmental Health and Safety
Work on this project was funded through Cooperative Agreement Number CR-818199-01 established
between the U.S. Environmental Protection Agency, Office of Research & Development, and the Center
for Hazardous Material Research. The organizations which contributed to funding under this cooperative
agreement include the EPA, Exide Corporation, and CHMR. The members of the project research team
appreciate the opportunity to participate in this important project to research and develop a new and
innovative technology for the remediatbn and treatment of contaminated sites. Inquiries concerning this
report or the project may be addressed to:
Mr. Stephen W. Paff, CHMM
Manager, Technology Development
Center for Hazardous Materials Research
320 William Pitt Way
Pittsburgh, PA 15238
(412) 826-5320
VIM
-------
1.0 INTRODUCTION
Lead is used in the production of various consumer and commercial items, from automobile and equipment
batteries to paints to crystal. This widespread use has made it one of the most common contaminants at
sites on the National Priorities List (NPL). The most common current treatment of lead contaminated
wastes at Superfund sites is immobilization, either on-site or in a landfill. Remedial approaches which
involve recovery of lead are preferred over immobilization, which wastes the lead. One such remedial
approach is the use of secondary lead smelters for recovery.
The initial sections of this report provide a brief overview of the sources and types of lead contamination
and available remediation technologies. The remaining sections present results from a study of secondary
lead smelting as a reclamation technology for lead-containing waste material at Superfund sites. The
economics of reclamation are examined to determine if the technology is economically competitive with
other technologies. Finally, a prognosis is given for the use of secondary lead smelters for lead recovery.
1.1 SOURCES OF LEAD CONTAMINATION
The prevalent use of lead in paints, gasoline additives, and other products resulted in wide dispersion
throughout the environment. A review of the literature [1-11] reveals the following sources of lead in the
environment:
• Lead acid battery breaking activities;
• Primary and secondary lead smelting and refining;
• Production of lead acid batteries;
• Production, storage and distribution of gasoline with leaded additives;
Solder use and manufacture;
• Plumbing;
• Ceramics and crystal manufacture:
• Paints (houses, bridges, ships), and paint abrasive blasting material;
• Wire manufacture and coating;
• Automobile demolition (auto fluff);
• Construction demolition (typically in plumbing and paints);
• Production and use of fishing sinkers;
• Pesticide production and use;
• Cathode-ray tube production and use;
• Rifle ranges and munitions dumps, including state game land and military ranges;
Ammunition and explosive manufacturing;
Sewage sludge;
By-products from metal production (e.g., electric arc furnace dust from steel production);
-------
• Radioactive shielding (from x-ray machines to reactors); and,
• Other metals mining, smelting, and alloying (copper, zinc, cadmium, and chromium).
These uses and sources of lead have contributed to many types of lead contamination, including lead
contaminated soils, battery breaker sites, and integrated breaker/smelter sites. These are discussed below.
1.1.1 Lead Contaminated Soils
Lead in soils can be from a variety of sources. The range of lead contamination from 436 Superfund sites
surveyed is between 0.16 and 366,000 mg/kg, compared to up to 200 ppm in typical uncontaminated soils
[12,13] and 51,000 mg/kg in lead ores used at a major primary lead smelter [ 1 ]. Lead has been detected
in soils in urban environments at concentrations of up to 15,000 mg/kg due to a combination of automobile
exhaust, lead paints and batteries [15]. Lead is commonly found in soils at battery breaker sites in
concentrations up to 100,000 ng/kg [8] and at gun clubs, with concentrations up to 10,000 mg/kg [9].
Small amounts of lead are naturally present in the form of galena (PbS) although upon exposure to air it
is converted to carbonates, sulfates, oxides and sulfur salts. Antnropomorpic lead sources in soils
typically include acetates, organometallic compounds, oxides, elemental lead, sulfates, halides, sulfides,
and silicates [16].
The United States Environmental Protection Agency (EPA) has set acceptable lead levels in soils at a
range of between 500 and 1,000 mg/kg. Otherlimits may be used based primarily on the likely uptake of
lead in children or the proximity of the site to residential areas [17].
1.1.2 Battery Breaker Site Materials
The components of lead-acid batteries include the battery case, lead electrodes (typically screens), spacers
that separate the electrodes and prevent shorting, sulfuric acid, and lead battery paste (PbS04). In the
past, the lead in lead-acid batteries was commonly removed by cracking or breaking the battery shell,
draining the sulfuric acid into surface impoundments or lagoons, and pulling out as much of the metallic
and paste lead as possible.
The battery cases, spacers, and some of the paste were sometimes improperly disposed or piled. Battery
cases were often disposed in drainage ditches and pits, and then buried with soil. Sometimes the cases
were burned in pits on site to recover additional lead. These materials, when segregated from the soil,
have lead concentrations ranging from 5,000 to 200,000 mg/kg [81. Battery cases historically were made
from ebonite rubber, which is a hard, black rubber containing coke and coal dusts. In the late 1970's,
battery manufacturers switched to polypropylene cases, which are readily recycled.
Among the materials typically found at battery breaker sites are [1,4,19]: broken or whole battery cases,
jead scrap, battery paste, sulfuric acid, lead debris and, (if the battery parts were burned) partially
incinerated battery parts. There were approximately 20 battery breaker sites on the NPL in 1991 [4].
1.1.3 Integrated Breaker/Smelter Site Materials
Some battery breaking plants also reclaimed lead on-site at smelting facilities. These facilities would sell
the recycled lead to battery manufacturers for use in the production of new lead-acid batteries. Materials
found at these sites include the same materials found at battery breaker sites, as well as alloys, slag,
emissions control dusts, dross, lead oxides, and calcium sulfate sludge from emissions scrubbers.
1.1.4 Other Lead-Containing Materials
Other lead containing materials frequently found at Superfund sites include lead munitions and shot,
automobile fluff, and lead additive residues. Lead is very common, and its forms vary greatly: one site,
for example, contains over 10 million kg of lead-containing plastic wire insulation, and a second site
contains 20 million kg of broken lead-containing crystal, accumulated over several decades.
1.2 AVAILABLE REMEDIATION TECHNOLOGIES
There are several options for treating lead contaminated sites, including soil washing or extraction,
immobilization, and reclamation. There are many variations of each type of remediation technology. A
-------
survey of records of decision (RODs) [18] indicated that over 70 percent of lead-contaminated sites were
treated with some type of immobilization technology, usually stabilization or disposal in a landfill permitted
to receive hazardous waste. Less than 10 percent were treated using washing (size separation or
extraction) technologies, and less than 10 percent were treated using a reclamation process to recover
usable lead. The remainder were treated using thermal technologies such as incineration because the
material was contaminated with organic as well as heavy metal contamination. Incineration destroys the
organic material, but there is typically lead-contaminated ash remaining after treatment. This ash usually
retains the metals and therefore may also be hazardous and require separate treatment and disposal.
Several current treatment technologies are described below.
1.2.1 Solidification/Stabilization
Solidification/stabilization technologies reduce the leachability of contaminants in the soil. Stabilization
refers to mixing of waste with reagents such as flyash to produce an end product that has about the same
consistency as the original waste, but with significantly reduced contaminant leachability. Solidification
involves mixing the waste with ingredients, such as binders or concrete, in cement mixers, pug mills, or
other types of equipment to produce solid blocks of material with considerable strength. The mam goal is
to lock the contaminants into the material so they do not leach out [1,4,21]. The two technologies are often
used together [20,21].
1.2.2 Capping
Capping a waste site reduces the mobilization of the lead by containing it on the site. This prevents direct
contact of the cpntaminated material with the public and the surroundings. Capping is attractive for lead-
contaminated soils, since the lead is often relatively immobile in some soil systems [3]. Capping may be
performed with compacted clay, synthetic liners or both.
1.2.3 Washing
Washing usually involves slurrying the waste with an aqueous solution, and then physically or chemically
removing the contaminants using acid, chelating agents, size separation, or solvents. Some soil washers
separate the heavy metals based on the principle that most contaminants are concentrated in the finer soil
fractions. These technologies remove the finer soils (which must be treated or disposed) and return the
coarse material to the site as clean soil [27].
Although washing is generally used for soils, some companies are developing technologies to re
wash the battery case material. These systems use gravity separation or other techniques to
remove and
segregate
the battery case's from soil, rocks, or other debris, and scrubbers are used to thoroughly wash the cases.
The clean battery cases can then be landfilled as non-hazardous waste [23,28]. It is difficult to remove all
the lead from ebonite battery cases, because the lead permeates the slightly porous ebonite.
Most soils washing technologies produce either a soils fraction with high lead concentration or an hydroxide
sludge, which must be disposed. However, several recently developed technologies purport to remove lead
from contaminated soils and produce metallic lead as a product [35], eliminating the need to dispose of
residuals.
Solidification/stabilization capping, and washing have been proven to be effective in remediating heavy
metals in soils [4,19]. The use of these technologies, however, has not been as successful when there is
large debris, such as battery cases or metallic lead, in the waste stream. Reclamation using secondary
lead smelters provides a more feasible alternative for such materials.
-------
2.0 LEAD RECLAMATION USING SECONDARY LEAD SMELTERS
Secondary lead smelters typically reclaim lead from spent lead-acid batteries. The use of secondary lead
smelters to reclaim lead from Superfund or other waste feeds involves slight rnodifiitions to the normal
smetting process. This section provides a description of typical secondary lead smelting processes, as well
as the modifications required to process the waste material.
2.1 TYPICAL SMELTING OPERATIONS
The evaluation was conducted at Exide/General Battery Corporation's secondary lead smelter location in
Reading, Pennsylvania. A schematic of the process is shown in Figure 1. The operations at this smelter
are typical of those found in the secondary lead smelting industry. As part of normal secondary lead
smelting operations using reverberatory and blast furnace combinations, spent batteries received at Exi de's
Reading smelter are crushed to release the sulfuric acid. Next they are processed through a sink/float
system to separate the battery cases from the heavier lead particles. The plastic battery cases are
recycled on-site for use in the production of new battery cases.
Reverberatory furnaces are charged with lead from the sink/float system as well as other lead containing
material, and are fueled with natural gas and oxygen. These furnaces are tapped for slag, which typically
contains 60 to 70 percent lead, and a pure (soft) lead product.
Blast furnaces are charged with the slag generated from the reverberatory furnaces as well as other
lead-containing materials, and are fueled by coke and air enriched with oxygen. Iron and limestone are
added as fluxing agents to enhance the lead production in the furnace by preventing much of the lead from
remaining in the blast furnace slag. Typical feed concentrations for coke, iron, and limestone are 5.7, 8,
and 0.9 weight percent, respectively [29]. The blast furnaces are tapped continuously to remove lead and
intermittently to remove the slag. The blast slag, which contains primarily silica, iron oxides, and some
lead, is usually not hazardous and is transported to an off-site residual waste landfill for disposal.
Lead produced in the blast and reverberatory furnaces is transferred to the refining process where
additional metals are added to make specific lead alloys. The lead is then sent to the casting operations
where it is molded into ingots ("pigs") for use in the manufacture of new lead-acid batteries. Waste
materials from Superfund and other sites are fed either to the blast or reverberatory furnaces, depending
on the material type, particle size, density, and other parameters. The modifications made to enable
processing of this waste material are discussed in the following section.
2.2 USE OF SECONDARY SMELTERS IN LEAD RECLAMATION
FROM WASTE MATERIAL
Prior work involving the use of secondary lead smelters to reclaim lead focussed on battery cases. In
general, waste material was simply put through a smelter to dispose of it. There were no quantitative
attempts made to analyze the affects on the furnace, or if the reclamation itself was successful [29]. One
goal of this research was to determine how the smelter furnaces react to the addition of various waste
materials.
The first step in the reclamation process, shown schematically in Figure 2, acquiring and transporting
the material to a secondary lead smelter. Generally, this involves excavation or collection, pre-processing,
and transport to the smelter. The lead-containing waste material is typically excavated from lead-add
battery Superfund sites or collected from other sources, such as bridge blasting or demolition operations.
Next, the material may require processing prior to entering the furnace. Preprocessing includes screening
to remove soil, large stones, or noncontaminated debris. Materials larger than 12 inches cannot be fed
into the reverberatory furnaces. These larger materials may cause jams in the belt system that feeds the
reverberatory furnace, or they may remain unburnt in the furnace for too long a period of time. The soil
and debris removed during pre-processing may be treated using one of the technologies described in
section 1.2, since soil cannot be processed through a secondary smelter.
-------
EXCAVATION OR
COLLECTION
TRANSPORT OF MATERIAL
PRE PROCESSING
POCKS, SOILS, DEBRIS
REVERB
FURNACE
LEAD TO
BATTERY PLANT
MIXING WITH
'TYPICAL SMELTER FEED
SLAG TO DISPOSAL
FIGURE 1
SCHEMATIC OF SMELTING PROCESS
The pre-processed material is mixed with regular furnace feed from the crushing and sink/float system.
The feed rate is determined by lead content, size of the material, fuel values, and other parameters.
2.3 EVALUATION METHODOLOGY, SITES, AND MATERIALS
The basic methodology during the project was to acquire quantities of the materials to be tested,
characterize these materials, process them through the secondary smelter, and observe their effects on
the furnaces through sampling and data collection. The primary objective of the tests was to determine
how much and what types of materials could be fed to the furnaces without causing shutdowns or otherwise
adversely affecting smelter furnace performance.
2.3.1 Acquisition of Materials
Materials from three Superfund sites as well as two additional sets of lead-containing materials were
processed during this project. The following sections provide a short description of each of the fiie
evaluations. The feed rates are presented as weight ratios of test material to total furnace feed. Table 1
presents a summary of the materials tested and the evaluations. More information on three of the sites
(Tonolli, NL Industries, and PennDOT) can be found in appendices A, B, and C, respectively.
Tonolli Superfund Site
The Tonolli site is a 30 acre battery breaking and smelting facility located in Nesquehoning, PA. Piles of
ebonite rubber and polypropylene battery case pieces were tested, without any pre-processing. The
material had an average lead concentration of 3.5 percent. Approximately 84 tons of material were fed at
a ration of 10 percent through a reverberatory and blast furnace. The material was too large to be readily
processed In the reverberatory furnace, but was successfully processed in the blast.
-------
SPENT BATTERIES
I
PRETREATMENT
EXPERIMENTAL
WASTE FEED
NATURAL GAS,
OXYGEN —
SOFT LEAD PRODUCT
PLASTIC TO RECYCLE
INTO NEW BATTERY
CASES
LEAD CONTAINING MATERIAL
FURNACE GASES
1
REVER6ERATGRY
FURNACE
> 2000°F
EXPERIMENTAL,
WASTE FEED
COKE, IRON
AIR, OXYGEN
SLAG (APPRGX, 70% Pb)
FURNACE GASES
BUST
FURNACE
2200°F
REFINING
CASTING
f
OTHER
LEAD FEEDS
HARD LEAD
PRODUCT
SLAG TO
RCRA DISPOSAL
LEAD TO BATTERY PLANT
MONITORED
EMISSIONS
SCRUBBER
BAGHOUSE
AFTERBURNER
DEWATERED SLUDGE DUST TO FURNACE
fTO LANDFILL) (FOR Pb RECLAMATION]
FIGURE 2
SCHEMATIC OF RECLAMATION PROCESS
Hebelka Superfund Site
The Hebelka site was a former autompbile junk and salvage yard located in Weisenburg Township, PA.
The site contained battey case debris mixed with soil that had an average lead concentration of 14.7
percent. Approximately 20 cubic yards of material were transported to the smelter. This material was first
-------
percent. Approximately 20 cubic yards of material were transported to the smelter. This material was first
reduced in size to less than 1/4-inch with a hammermill. The material was successfully fed to the
reverberatory furnace at a feed ratio of 17 percent.
Demolition Waste
Demolition waste was obtained from a Housing and Urban Development (HUD) project. The waste
consisted of demolition debris mainly wood) coated with lead based paint and had a lead composition of
between 0.5 and 1 percent. The test material was shredded in a pallet shredder before It was smelted
The demolition debris was processed through both reverberatory furnaces at feed ratio of 10 percent test
material, by weight. At this weight ratio the test material comprised 50 percent of the volume fed to the
furnace. This high volumetric ratio caused malfunctions in the furnaces, so the feed ratio was reduced to
5 percent, at which point the material was successfully fed.
NL Industries Superfund Site
The NL Industries site in Pedricktown, NJ, was an integrated battery breaking, smelting, and refining facility
with its own on-site landfill. There were a wide variety of materials at the site including lead slag, dross,
debris, ingots, hard heads (large chunks of metallic lead), battery case debris, baghouse bags, and
contaminated pallets and iron cans. The evaluation was conducted in two parts: a preliminary investigation
and a full-scale investigation.
-------
Table 1 Summary of the Evaluations
Site
Tonolli
Hebelka
Laurel House
NL Industries
PennDOT
Site Description
Integrated battery
breaker
Auto junk & salvage
yard
Residence
Integrated battery
breaker & smelter
Bridge repainting
site
Site Size
30 acres
20 acres
N/A
11 acres
N/A
Test
Duration
5 days
1 day
1 day
3 months
1 day
Type of Mat' 1
Rubber & plastic
battery cases
Mainly plastic
battery cases
(some rubber)
Wood demolition
debris
lead slag, debris,
dross, ingots,
cases, baghouse
dusts
Iron shot bridge
abrasive
Amt.
Processed
84 tons
12 tons
4 tons
1570 tons
6 tons
Avg. Lead
Cone.
3.5%
14.7%
1%
30 to 60%
3.2%
Feed wt.
Ratio
20%
17%
5%
20 to 50%
13%
09
-------
During the preliminary investigation, approximately 370 tons of all types of the above materials were
processed. Analyses revealed an average lead concentration of 57 percent. The larger pieces of debris
were removed and processed through a blast furnace, while the bulk of the material was fed into a
reverberatory furnace at feed ratios of up to 100 percent. The feed was sufficiently dense to cause
breakdowns in the reverberatory furnace conveyor feed system, so the feed ratio was reduced to 50
percent test material, by weight.
During the full scale operation, approximately 1200 tons of material were transported to the smelter over
a three month period. For the first two months, the test material, which contained approximately 50 percent
by weight lead, was processed in the reverberatory furnace, with a feed ratio of 20 to 30 percent. The ratio
was limited due to high amounts of calcium in the NL material. The high calcium concentration slowed the
operation of the furnaces.
The test material for the last month of the investigation consisted mainly of larger pieces of slag and debris
with an average lead concentration of 30 percent. This material was charged directly to one of the blast
furnaces, at a feed ratio of approximately 30 percent.
Pennsylvania Department of Transportation
The Pennsylvania Department of Transportation PennDOT) used an iron-shot abrasive blasting material
to remove old lead-based paint from a bridge in Belle Vernon, PA. Sixteen 55 gallon steel drums of this
material, containing an average of 3.2 percent lead, were processed at the smelter.
The bridge blasting material was primarily iron (60 percent) with 5 percent calcium and 5 to 10 percent
moisture content. The material contained too much moisture to be incorporated into the reverberatory feed.
The test material, including the drum, was fed to one of the blast furnaces at a feed ratio of approximatety
13 percent, by weight.
Bypass 601 Site Material
Material was obtained from the Bypass 601 Superfund site in Concord, NC. The material consisted of
battery cases, mixed with soil. Overall, it contained approximately 1 percent lead. The material contained
approximately 80 percent soil and 20 percent battery cases. Based on this, the use of the reclamation
technology was deemed not to be feasible, unless a suitable separation method for the cases and soil was
developed. Although relatively simple size separation methods for separating the soil from the battery
cases exist, they did not appear likely to be economically viable, because the soil contained significant
rocks and other debris which would remain with the battery cases, and because a majority of the lead
would stay in the soil. Therefore, the site was dropped from consideration by this study.
2.3.2 Smelter Furnaces
Exide's secondary smelter has two reverberatory furnaces and two blast furnaces. Each reverberatory/blast
combination Is similar in design and construction, as well as production potential. During the evaluation,
one furnace (either the reverberatory or blast, depending on the waste material) was charged with regular
smelter feed and the other was charged with regular feed mixed with certain amounts of the waste material.
This provided "control" and "test" furnaces, and comparisons could be made between the two. Because
the furnaces normally undergo wide fluctuations in production, the comparisons were only valid when the
furnaces were fed materials over four to six shifts.
Reverberatory furnaces are often used to remove metallic lead, and to produce a fused slag feed for the
blast furnaces. The reverberatory furnace has a two to three hour residence time. The most serious
problem encountered by introducing new feeds into the reverberatory furnaces was the stacking of unburnt
material inside the furnace. This can cause impaired performance or even damage to the furnace. This
condition takes from one to several hours to occur, so the furnaces were still monitored after the test feed
was stopped.
Blast furnaces are used to reduce lead oxides to metallic lead, and remove them as product. The blast
furnace has a residence time of several hours. Therefore the monitoring of the blast furnace charged with
test material was continued for several hours after the waste feed is discontinued. This allowed time to
more accurately determine if the test feed had any positive or negative affects on the blast furnace
performance.
-------
2.4 DATA AND SAMPLE COLLECTION
Project personnel collected samples and data to assess the furnace performance, characterize the input
material, and characterize the furnace outputs. Table 2 shows the parameters that were measured and
how they were obtained. The input material parameters were characterized to provide information related
to the feed, so that comparisons of the effects of different feeds could be made. The furnace performance
parameters, such as air flow, oxygen usage, fuel usage, and furnace feed rates provided measurements
of the furnace performance while the experiments were conducted, principally indicating when production
levels were falling or materials were clogging in the furnace. The output parameters were the most
important ultimate measurements of furnace performance - including both production rates and quality, as
well as residuals generation.
The data generated from measurements and sample analyses were used to compare the performances
of the test and control furnaces. The amount of lead in each product is useful in making a mass balance
for the lead. The other parameters, such as oxygen, air, and fuel, are useful in determining the cost for
processing the test material.
The procedures for obtaining the samples (marked "S") in Table 2 varied depending on the materials to
be sampled. For samples of incoming waste materials, representative samples were obtained by
compositing two to four individual grab samples obtained from various locations and depths within the pile.
Generally, a total of one to two Kilograms of such materials were obtained. Duplicate samples were
obtained for approximately 25% of the samples. Samples of lead and slag produced in the furnaces were
obtained through the use of a small crucible placed into the lead or slag stream.
TABLE 2
INPUT, OUTPUT, AND OPERATING PARAMETERS
Input Material
Characterization
total lead (S)
sulfur (S)
silica (S)
calcium (S)
moisture content (S)
density (M)
particle size distribution (M)
BTU value (S)
Furnace Performance
Parameters
test material in the feed (M)
air flow (M)
% oxygen enrichment (M)
fuel usage (M)
lead inputs (S)
iron inputs (S)
% test material in feed (M)
Furnace Output
Parameters
lead production rates (M)
slag production rates (M)
slag viscosity (0)
% lead in the slag (S)
% sulfur in the slag(S)
back pressure (M)
sulfur aioxide emissions (M)
calcium sulfate sludge (S)
(S)-Sample (M)=Measurement (0)=0perator Observation
Furnace performance and output parameters were monitored through a combination of review of operators
notes and logs, and direct observatipn by CHMR personnel. In general, production data were obtained
exclusively through operator logs, since that was the only form available for the data. These logs are
accurately kept and monitored by plant management (who use the data for accounting purposes, and
therefore require accuracy). CHMR obtained copies of the logs when such data were to be used.
2.5 QUALITY ASSURANCE/QUALITY CONTROL
The project was conducted in conformance with the approved Quality Assurance Projed Plan (QAPP),
dated October 28, 1991.
The QAPP procedures related to several sets of parameters, including those listed in Table 3. T h e
analytical or measurement methodology is also listed.
10
-------
Some of the total parameters, including total lead, sulfur, iron, silica, calcium, were also determined using
semiquantitative emissions spectrographic methods, which determined an approximate concentration of
a wide range of compounds. These methods were employed when approximate measurements (±30%)
were sufficiently accurate, or as screening methods to determine if a wide range of elements were present
in a sample.
TABLE 3
Methods Used to Measure Test Parameters
TYPE OF PARAMETER
Test Material Parameters
Furnace Perf. Parameters
Furnace Output Parameters
PARAMETER
Total lead
Total sulfur
Total Silica
Total Calcium
Moisture
Density
Percentage Test Material in
Furnace Feed (weight)
Air Flow to Furnaces
Oxygen Enrichment
Total furnace feed
Fuel Usage
Lead quality
Lead production rates
Slag production rates
Percent tead in slag
Percent sulfur in slag
SO2 Concentration in
Stacks
ANALYTICAL OR
MEASUREMENT
TECHNIQUE
SW 846 6010 (digestion 8
CP
ASTMD129-64 (digestion
&ICP
Spectrography
SW 846 71 40
Drying/weighing
Weigh known volume
Ratio of charges, density of
materials
Plant rotameters
Plant setting
Plant shift records/
observations
Plant ftowmeters
Spectrography (measure
impurity concentrations)
Plant shift records
Plant shift records
SW 846-60 10
ASTM D 129-64 or
equivalent
Plant CEM (calibrated
weekly)
Quality control procedures for the laboratory analyses included quality control both within and without the
analytical laboratory. For quality control within the analytical laboratory, CHMR reviewed the laboratory
11
-------
standard procedures for conformance with the standard analytical methods. These procedures called for
the use of periodic duplicates, spiked samples, and blanks.
Quality control procedures outside the analytical laboratory related mainly to in-plant measurements. Many
of these measurements were obtained from plant shift data, which are presumably accurate. The
production data was corroborated where possible by CHMR personnel observations - how many charges
of material were fed to the furnaces during a given shift? When was the furnace operating? How long did
they pour product or slag? These observations showed consistency between the numbers obtained and
the practices observed. For example, over two shifts when CHMR personnel noted a 30 percent reduction
in the number of charges fed to the furnaces, there was a corresponding 28 percent decrease in the weight
of material fed to the furnace noted on the shift records, and a decrease in lead production of over 20
percent. Based on these observations, CHMR believes that the production data were accurate to within
±10 percent.
Other observations, sutfur dioxide emissions data, for example, were obtained from instruments which are
required by operating permits to be maintained and calibrated. Although it was not the focus of the study,
CHMR personnel noted no obvious deviations from those requirements while on site.
Material density was obtained by weighing a known volume --typically a carboy - of material, and/or
weighing an entire truckload of material as it was dumped at the plant site. This latter method was believed
to be more accurate, because the volume used could be determined based on truck bed size before the
material left the waste site and settled in the truck. From analyses of the feed samples from the Tonolli
study see Table A-1), CHMR found a range of bulk densities of 23.2 to 28.4 Ib/ft3 (average 25.3 Ib/ft3)
using the carboy method, and 23.7 to 31.4 Ib/ft3 (average 27.6 Ib/ft3) using the truck weight measurement.
The overall average of the densities was 26.7 Ib/ft3. The relative percent difference (RPD) (calculated as
the difference between the duplicate analyses divided by the average of the two values), between the
methods of calculating was 8.7 %.
The most pressing quality control issues related to the sampling. Battery case piles are notoriously non-
homogeneous, because the particles within them segregate both according to density and particle size.
In one early experiment, for example, CHMR obtained total lead concentrations ranging between 3 and 16
percent.
To reduce errors introduced by the non-homogeneity, CHMR took composite samples from several areas
within a pile (3 to 4). In addition, the samples were often taken of the material as it was loaded onto the
truck for transportation to the smelter. This provided a more representative sample of material actually sent
to the smelter.
During the NL Industries study (Appendix B), CHMR obtained duplicate analyses of eight input material an
slag samples between two different analytical laboratories see especially Tables B-1 and B-5). The RPD
for ead concentration ranged between 0.9% and 42%, with an average of 17.9%.
Overall, CHMR's estimates of the total lead concentrations and other parameters, where they are based
on averages from several samples, are estimated to be within 20 percent of the actual value.
12
-------
3.0 RESULTS AND DISCUSSION
In general, the study demonstrated that various materials may be processed in secondary lead smelters
witn relatively few effects on overall furnace performance. The mpst significant effects were caused by
processing materials in a furnace without property pre-processing it, or by processing too much material
at one time. For example, the battery case pieces in the Tonolli feed were so large that they burned too
slowly in the reverberatory furnace, causing a reduction in the furnace production rate. Later, battery case
material from the Hebelka site was successfully processed in the reverberatory furnace after it had been
shredded in a hammermill to a particle size of less than 1/4-inch. The NL Industries material was initially
unsuccessfully processed at 100 percent feed ratio because it was too dense for the feed system. When
the ratio of test material to total feed was lowered to 50 percent, the material was processed with few
problems.
More extensive summaries of the evaluations from these sites (NL Industries, Tonolli, PennDOT, Hebelka
and HUD) may be found in Appendices A through E, respectively.
3.1
PRODUCTION PARAMETERS
The lead and slag production from the furnaces varied with each evaluation. Table 4 shows three lead and
slag production parameters derived from the daily production of lead and slag at the smelter. The
demolition matenal evaluation is not included in Table 4 because this test was performed primarily to
qualitatively determine if the furnaces could process the light, wooden material.
TABLE 4
LEAD AND SLAG PRODUCTION COMPARISONS
Furnace
Reverb
Blast
Parameter
Wt % Test Feed
% Pb in Test Feed
Pb Prod. Ratio
WL % Test Feed
% Pb in Test Feed
Pb Prod. Ratio
Pb Loss Ratio. /.„
Slag Ratio
Evaluation
Tonolfi
N/F
N/F
N/F
10%
3.5%
1.3
0.6
0.97
'Composition of bias
N/F means that the test mat
Hebelka
17%
14.7%
1.20
N/F
N/F
0.9
ft
1.00
NL Industries
20to50%
57%
0.90
30%
30%
1.25
1
0.97
PennDOT
N/F
N/F
N/F
13%
3.2%
1.1
2.6
0.97
slag was not analyzed
3rial was not fed to that furnace.
3.1.1 Lead Production Ratio
The Lead Production Ratio is the ratio of the amount of lead produced by the furnace processing test
material to that produced by the furnace processing regular feed. This parameter can be determined for
the reverberatory furnaces only when the furnaces are fed waste material, but can be determined for the
blast furnaces in all cases, since the blast furnaces indirectly receive the test feed through the reverberatory
slag. A value greater than 1 indicates that the furnace processing the mix of test and regular feeds
produced more lead than the control furnace.
13
-------
The Ratios for all the other evaluations were different than 1, but not significantly so given the relatively
short durations of each of the tests as compared to the three month ML Industries evaluation.
For the NL Industries material, the lead production ratio for the reverberatory furnace is consistently below
1, averaging 0.9 over the three month period in which the materials were added to the furnace. However,
during the same time period, the blast furnace associated with the "test' reverberatory furnace produced
significantly more lead than its control counterpart. It appeared that the use of the test material shifted the
production of lead from the reverberatory tp the blast furnace. This shift is attributed to the relatively high
concentration of antimony and lead oxides in the waste feed material. Antimonial lead and lead oxides melt
at higher temperatures than elemental lead, and are therefore more likely to be removed in the blast
furnaces, which operates at a higher temperature than the reverberatory furnaces.
3.1.2 Lead Loss Ratio
The secondary lead smelter has two major outlets for the lead input: production and blast furnace slag
(waste). The Lead Loss Ratio(Lp) Is the ratio of the loss of lead in the slag from the test furnace (total slag
produced in the test furnace multiplied by lead concentration in that slag) to the loss of lead in the control
furnace. This ratio can only be calculated for the blast furnaces because the blast slag is a waste product
from the process. An Lf, value greater than 1 means that more lead is lost in the blast slag produced by
the furnace(s) processing test material than in the control furnace(s).
The Tonolli evaluation produced a much lower LRthan 1, while the value for the PennDOT evaluation was
2.6. The composition of the blast slag produced in the Hebelka evaluation was not determined. The
apparent low value during the Tonolli experiment is most likely due to normal variatipns in the furnace
operations and sampling procedures, and it is not attributable to a significant increase in efficiency.
Over the course of tne experiment with the PennDOT materials, the lead concentrations in the slag were
high and appeared to be climbing in both furnaces, even though the furnaces were not fed the same feed.
Therefore it is difficult to determine how significant the apparent increase in lead loss was. However, there
is reason to believe that some of the increased LR was due to the feed, since approximately twice as much
iron than is usually fed was added during the course of the experiment, and additional iron binds lead to
the slag. Lead concentration in the slag and quantity produced when the NL Industries site material was
processed were not different than when typical feed was processed. Since this was the longest evaluation
in terms of time, this may show that most material will have no effect on furnace performance in the long
run.
3.1.3 Slag Disposal Ratio
The Slag Disposal Ratio (Sd is the amount of blast slag produced by the test furnace divided by the
amount of slag produced by the control furnace. This is another term that is exclusive to the blast furnaces
because reverberatory slag is fed to the blast furnace, while the blast slag is disposed. In all of the
evaluations, the S/^was effectively 1, which means the test and control furnaces produced nearly equal
amounts of blast slag.
3.2 FEED RATIOS OF SUPERFUND MATERIALS TO TOTAL FEED
Secondary lead smelters normally accept feed materials derived principally from the breaking and
demolition of lead-acid batteries. These normal feedstocks consist primarily of lead grates, terminals, and
sludge from the batteries. This material typically contains 60 to 70 percent lead, by weight with the
remaining materials consisting of plastic, rubber, wood scrap, sulfur, dirt and residues from the batteries,
and moisture and oxides combined with the lead. Superfund wastes and other materials, which contain
lower concentrations of lead and higher concentrations of other constituents that may be harmful to the
furnaces, must be blended with "normal" feed prior to being led to the furnaces.
The mix ratio of normal feed to waste materials was one of the parameters tested during this study. Based
on furnace performance and operations results, it was possible to determine whether the furnace performed
successfully or unsuccessfully at any given feed ratio. A test is considered to be unsuccessful if the feed
ratio of test material has to be lowered, or discontinued altogether.
The tests show that the successful mix ratios may be correlated with the percentage of lead in the waste
material. Figure 3 shows the feed ratios which were successful, plotted against the percentage of lead in
the waste material (i.e., prior to blending with normal feed). The results are nearly linear from streams
containing 3 percent lead to those containing 60 percent lead. This delineates the region marked
14
-------
100% ,-
80% -
2
Q
UJ
UJ
Sb
o
ffi 60%
40%
• 20%
i UNSUCCESSFUL
I UNTESTED
SUCCESSFUL
0%
0% 10% 20% 30% 40% 50% 60%
% LEAD W TEST MATERIAL
a SUCCESSFUL EXPEFUMENT A UNSUCCESSFUL EXPEWMENT
FIGURE 3
PERCENT LEAD IN TEST MATERIAL VERSUS FEED RATIO
the waste material (i.e., prior to blending with normal feed). The resujts are nearly linear from streams
containing 3 percent lead to those containing 60 percent lead. This delineates the region marked
"successful" on the figure. When the feed ratios versus lead concentration for unsuccessful runs are
plotted, a second region (the "unsuccessful" area) emerges. Finally, a third region, in which no tests were
performed is also marked on the figure. No test feed with more than 60 percent lead was fed to the
furnaces, so the regions in the range above 60 percent lead cannot be determined from the experimental
results.
It is important to note that the test results represent an empirical observation based on the tests performed.
Some of the correlation can be easily accounted for- lead concentration affects density for example, and
material which is too dense is not easily conveyed into the furnaces, while material which is too tight will
fill up the furnaces and cramp out other feeds. Much of the correlation, however, is not easily accounted
for, and is simply based on furnace operator observations. CHMR believes that further study may help
delineate the "untested" region, but that it may be more practical to simply perform a treatability study on
a given feed to more accurately determine the acceptable feed ratio.
3.3
BENEFICIAL EFFECTS OF CERTAIN MATERIALS ON FURNACE PERFORMANCE
Some materials, particularly ebonite battery cases (which contain coal or coke dusts), represent a
potentially significant source of energy to the furnaces. The battery case materials had average BTU-
values of over 11,060 BTU per pound. Coke is a typical feed to a blast furnace. Ebonite rubber cases
were successfully substituted for a portion of the coke in the blast furnace. Battery cases will not substitute
for all of the coke since the large chunks of coke provide structure inside the furnace, which battery cases
cannot provide.
15
-------
Other materials represent substitutes for other furnace feeds. Iron is typically fed to a blast furnace to help
separate lead from the slag. The PennDOT bridge blasting material represented a source of iron and
calcium in the blast furnace. The material contained so much iron (over 60 percent) and so little lead in
comparison (approximately 3 percent), that it may be considered an iron source which happens to be
contaminated with lead, rather than a lead source to the furnace. In this capacity, it represents a pptentially
beneficial reuse for a material which most state and municipal authorities have found difficult to dispose.
This material would have to be added at a slower rate than during the test run to be used as an iron
source
3.4 APPLICABILITY OF LEAD RECOVERY FROM WASTE MATERIALS
Broadly speaking, lead reclamation in secondary smelters is an applicable technology whenever sufficient
lead exists in a form which can be economical retrieved from the Superfund material. However, there
are certain materials, such as materials containing large amounts of soil, In which the lead may not be
easily reclaimed using the technology. This section describes the procedures to be used to determine
applicability and the range of applicability determined by the experiments.
Generally, to determine if a material can be processed through a secondary lead smelter, a two step
procedure is required: material characterization followed by a pilot or treatability study at the smelter.
Table 5 lists the typical parameters which must be included in a material characterization. Not all these
parameters were obtained or measured during the study: many are based on requirements obtained from
the smelter. Note that the smelter operators are not interested in the TCLP leachability of the metals in
the waste, but the total metal content. The leachability only determines the regulatory status (hazardous,
non-hazardous) of the material, and does not affect its properties in a smelter.
The table is intended to provide an approximate gauge of the materials that can be processed through the
smelter. Because there is such a wide range of lead-bearing materials found in the environment, it is
probably necessary to discuss a particular waste feed with smelter personnel before deciding whether the
lead is suitable for reclamation. Some wastes may be blended with other materials io make suitable
feedstock. A high calcium waste, for example, may be blended with other materials with little or no calcium
and fed to the furnace. Of course, this blending may significantly increase the cost of storage and
processing. Other materials, such as metallic aluminum, are prohibited in furnaces.
After characterizing the waste material to determine its suitability in a smelter, a pilot-scale study may be
necessary to better determine the effects on the furnaces and the appropriate operating parameters. These
will likely have significant bearing on the economics of treating the material. The demolition material tested
during this study, for example, had to be shredded and blended relatively slowly with normal furnace feed.
This increased the handling and storage requirements by more than 50 percent. A more detailed cost
analysis is presented later in this paper.
16
-------
TABLE 5
REQUIRED CHARACTERIZATION
PARAMETERS
Constituent
Total lead
Total antimony
Total arsenic
Total copper
Total halogens
Total tin
Total iron
TCLP - metals
Physical state
Total sulfur
Total silicate
Total soil content
Fuel value (BTU/lb)
Total calcium
Total aluminum
Particle size
Oil and Grease
Moisture
Density
Ash content
(excluding lead)
Lower limit (If
applicable)
1%
NA
NA
NA
NA
NA
NA
NA
No liquids or
gases
NA
none
none
none
none
none
5 microns
none
none
none
none
Upper limit (If
applicable)
100%
NA
2%
2%
<1%
2%
none
NA
No liquids or
gases
10%
20%
25%
NA
10%
1%
I m
<2%
15%
none
70% (est.)
What does parameter affect?
Economics (based on lead
recovery), feed rates
Needed to determine refining
requirements
Lead quality, refining
Lead quality, refining
Damage flue gas system, permits
Lead quality
Feed rates (iron is a furnace feed)
Overall need to process
Conveyance, furnace performance,
permits
CaS04 sludge production, sulfur
emissbns
Furnace performance
Furnace performance, economics,
slag production
Economics
Furnace performance, feed ratios
May cause explosions in the
furnaces
Loading, conveyance
Permits, feed system, material
handling
Heat requirements, may cause
steam explosions
Loading, conveyance, economics
Slag production, furnace
performance, economics
3.5
LEAD RECLAMATION EFFICIENCY
In determining whether the process is suitable for Superfund activities, it is important to determine if the
secondary lead smelting process actually reclaims lead from the various materials. Unfortunately, there
was no way to precisely measure the extent to which the lead from the Superfund materials is reclaimed
after these materials are mixed with regular furnace feeds. Therefore, a methodology was developed to
estimate the minimum reclamation efficiency based on conservative assumptions regarding the partitioning
of lead inside the furnaces.
17
-------
Generally, lead inputs to the smelter include feeds to the reverberatory and blast furnaces. During normal
smelting operations, outputs from the smelter include [29]:
Lead production from both furnaces (usually between 99.5 and 99.7 percent of the lead
feed);
Blast furnace slag (which contains 0.3 to 0.5 percent of the lead feed);
Calcium sulfate sludge from the emissions control system (which contains approximately
0.004 percent of the lead feed); and,
Fugitive and stack air emissions (which contain less than 1CT5 percent of the lead feed).
The furnaces reclaim over 99 percent of the lead fed to them. The only significant outlet for lead besides
lead product is the slag, which must be landfilled. For the purposes of estimating reclamation efficiency,
the slag may be estimated to contain 1.5 to 2 percent lead, which the literature suggests can be maintained
over a variety of operating conditions [30].
The minimum reclamation efficiency can be estimated noting that the lead partitions between the lead
product, which is essentially pure lead, and slag, which contains 1.5 percent (by weight) lead. The
percentage of any feed material partitioning to the slag is approximately equal to the concentration of non-
lead, noncpmbustible materials in the feed (which is approximately equal to the ash content minus the lead
concentration in the feed). Thus, for test material with a measured ash content of 50 percent, and 30
percent lead, it Is reasonable to assume that the remaining 20 percent of the feedstock will become part
of the slag. If we further assume that the slag produced by Superfund materials is indistinguishable from
that produced by normal feeds operating under the same furnace conditions, and therefore will contain the
same weight fraction of lead as the slag normally contains, then we can estimate the amount of lead from
the Superfund material lost with the slag. This can be compared with the original amount of lead in the
Superfund material to determine a reclamation efficiency:
e = 1 - [PbJ ([A] - [PbMPb]
where e is the minimum reclamation efficiency; [Pb?] is the lead concentration in the slag; [A] is the ash
content of the test material (as a weight fraction); and [Pb] is the weight fraction of lead in the feed. Based
on this estimate, the minimum reclamation efficiencies for the various feedstocks used during the
experimental program ranged from a high of 99.5 percent for the Pedricktown material with 45 percent lead
and 60 percent ash content, to a low of 10 percent for the PennDOT material, with an ash content of 70
percent, a lead content of approximately 3 percent, and a lead concentration in the slag of approximately
4%. (Note that the lead concentration in the slag was especially high dunng the time that the PennDOT
material was run for reasons which appeared to be independent of the PennDOT feedstream. The lead
reclamation efficiency would normally be considerably higher.) The results show that lead reclamation
always occurs, even when minimum reclamation conditions are assumed.
18
-------
4.0 LEAD RECLAMATION ECONOMICS
The cost of using secondary lead smelters for the recovery of lead from Superfund sites is determined by
the cost of the three basic operations:
On-site excavation and preprocessing of the material (CE)
Transportation (CT)
Processing at the smelter (CP)
These costs are discussed in further detail below.
4.1 ON-SITE EXCAVATION
Excavation costs for Superfund materials vary depending on the health and safety requirements for given
sites, contaminants and contaminant concentration. For battery cases, or lead debris, dross, etc., which
are relatively accessible on the site, and require little more than dust control during excavatipn, the costs
range between $5 and $15 per cubic yard. On-site costs, denoted as CE, will Increase if additional on-site
processing is required.
4.2 TRANSPORTATION
Transportation costs for hazardous wastes are dependent on the amount of material transported, and the
distance transported. Typical costs range between $0.20 and $0.35 per ton-mile (transport 1 ton 1 mile).
For most material, the cost of transportation, CT, is estimated as:
Cy = $0.3 D/mile
where D is the distance from the site to the smelter in miles. Note that for longer distances (over 300
miles), alternate means of transportation (i.e. tandem trucks, rail shipment, etc.) may be more economical.
The transportation economics shift slightly for some materials, particularly those with specific gravities
significantly less than 1. These materials may include whole (i.e., uncrushed) battery cases which, because
of their structure, have relatively large void volumes and therefore a low density. Unless these materials
are crushed on-site, the transportation costs will be based not on their weight but on their volume. This
will necessitate a correction factor on the transportation cost equation w'nich is equal to the inverse of
the material bulk density in ton peryd^, and a maximum of 1. Thus, the overall transportation cost can be
estimated as:
CT = 0.3 is equal to max (1/p,1), in which p is expressed in dimensions
of ton per yd.3
4.3 PROCESSING AT THE SMELTER
Based on the results of the study, processing costs for the materials vary significantly depending on the
concentration of lead, the market price for lead, and the percentage of the feed which becomes slag. As
the market price for lead or the concentration of lead in the feed material decreases, the cost of processing
Superfund materials will increase, because the lead represents a salable commpdity generated during the
reclamation. If the material contains a greater fraction of constituents which exit the furnace in the slag
fraction, then the cost increases commensurate with the disposal costs of slag. Most of the other
parameters (for example, a slight increase in oxygen usage in the furnace) have little overall affect on the
cost of processing.
CHMR has developed a model for estimating the costs associated with processing Superfund materials at
a smelter. The model breaks the cost into four major categories: a base cost (per ton) to cover a portion
19
-------
of normal smelting costs, additional production costs specific to the material, additional disposal costs, and
offsets of the value of the recovered lead and reductions in other feeds (if any) to the furnace. These are
described further below.
4.3.1 Base Cost CB
According to the literature [12, 17, 27, 29] the cost of processing material in a smelter is approximately
$0.10±0.02 per pound or $200± 40 perton of material processed. Note that this is not the cost of
producing a pound of lead which is typically estimated at between $0.12 and 0.17 per pound, but is the
cost of processing a pound of feed. The two are related, since furnace feeds [including coke, iron and
limestone] contain 55 to 65 percent lead. This cost includes the cost of handling and loading the material,
maintaining the smelter, normal furnace inputs (coke, natural gas, oxygen, etc.), labor costs, environmental
compliance costs, etc.
4.3.2 Additional Production Costs
These costs include additional labor and equipment costs associated with handling an atypical feed at the
smelter. Included in this category are costs associated with maintaining a staff member on-site during
loading to ensure that the proper materials are loaded onto the trucks (estimated at $6 per ton); labor costs
asspciated with inspecting, unloading, and mixing the material at the smelter ($10 perton); costs for
additional management time and attention for the material ($3 per ton); additional capital costs at the
smelter, including loading equipment, storage areas, and piloting costs (estimated at $18 perton); legal and
contractual costs associated with handling Superfund materials ($10 per ton); and miscellaneous costs,
such as analytical costs (estimated at $10 per ton). These additional production costs remain relatively
fixed independent of the amount or types of matenals processed, at $60110 per ton.
For light materials, the additional processing costs will be based on their volume, rather than weight. More
than likely, however, the smelter will choose to crush or grind them before feeding them to the furnaces,
at a cost of between $10 and $15 per cubic yard. To denote this, a correction factor T = $15 per ton is
applied when the bulk density of the matenal is below 0.8 tons per yd3. If the density is above 0.8 tons per
yd3, «F= 0.
4.3.3 Additional Disposal Costs Cdisp
Additional disposal costs associated with processing Superfund materials include the disposal of slag and
calcium sulfate sludge. Based on measurements taken at the smelter, CHMR determined that typical
smelter feeds produce in the range of 200 Ibs of slag per ton of feed 10 percent slag in the feed), and 60
Ibs of sludge per ton of feed (based on 3 percent sulfur in the feed). This material is typically disposed of
at a cost of $150 perton. The cost to dispose of 200 Ibs of slag and 60 Ibs of sludge is included above
as part of the overall cost to process the material. However, some Superfund and waste materials may
contain significantly higher percentage of compounds which will end up in the slag, or sulfur which will
produce sludge. Therefore, the overall processing cost must be adjusted to include the cost to dispose of
the waste materials produced by the feed. This adjustment is calculated as:
Cdisp= $150 ([N] - 0.1) + $100 ([S] - 0.03)
where [N represents the percentage of non-combustible, non-lead, and non-volatile material in the
superfund material, and [S] is the sulfur content of the feed. [N] can then be calculated by:
[N] = [A] -1 .1 [Pb]
where [A] is the ash content of the waste material, measured at or above 1500°F, and [Pb] is the
percentage lead in the material. The correction factor to the lead concentration corrects for the appearance
of lead oxides in the lead. The overall disposal cost differential is therefore:
Cdisp = 150[A] - 165 [Pb] + 100 [S] - 18.
4.3.4 Offsets for the Value of Lead and Reduction of Other Furnace Feeds
The value of the recovered lead in a waste material is given by:
CPb= - e, P [Pb]
20
-------
where Er is the reclamation efficiency for the feed in question (assumed to be approximately 1), P is the
market price for lead in dollars per ton and [Pb] is the weight fraction of lead in the material. The negative
sign on the cost indicates that this is not a "cost" per se, but in reality a credit for the lead content to the
smelter. The January 1994 market price for lead was approximately $700 per ton. This is down from over
$750 per ton in 1991, but up from $450 per ton in 1992. The volatility in the price of lead is due to large
sales of lead from the former Soviet Union and the price is anticipated to stabilize during 1994 [32].
A second benefit from Superfund materials is decreased coke usage. For materials such as battery cases,
which can be essentially a supply of coke to the furnace, the change in required coke is related to the heat
value of the Superfund materials (which in this discussion will be denoted by [BTU] and given in units of
BTU per pound). If coke were used only as a fuel source inside a furnace, and if the battery case materials
burn similarly to coke, then the reduction in coke usage based on addiiion of a Superfund materials would
be equal to the ratio of the BTU value of the Superfund material to that of coke (i.e., if you feed 1 BTU of
energy to the furnace from the Superfund material, you would save 1 BTU's worm of coke in the furnace).
However, coke is not used only as a fuel source, but also to provide structure to the materials inside the
furnace. In addition, battery cases, with different sizes and shapes, do not burn at the same relativity
homogenous rate as coke input to a furnace. Therefore, a one-to-one reduction in energy supply between
Superfund materials and coke is not possible. Based on the reduction in coke requirements for the Tonolli
feed, in which a 10 percent feed of a material with a heating value of 12,000 BTU per pound allowed for
a reduction of 30 percent in the coke usage, CHMR estimates that approximately 25 percent of the heat-
value of a Superfund waste can be supplied to offset the blast furnace requirements for coke. Based on
this, the BTU-value of coke (13,000 BTU per pound) and the market price or the coke used in the furnace
(approximately $150 per ton), the cost savings from the BTU-value of the Superfund material can be
estimated as:
Ccoke= - 0.007 ([BTU] - 2000)
where Ccoke is in terms of dollars per ton of feed material. The cost savings include savings from not having
to add the coke as part of the base feedstock to the furnace. The 2,000 BTU per pound differential in the
equation is included because below a minimum level, the material probably provides no realizable benefit
to the furnace, and may even increase fuel requirements because of its thermal capacity.
Likewise, the content of a Superfund material may provide significant benefit to the furnace. The
extent of reduction allowed by the addition of Superfind materials containing iron could not be estimated
from the experiments conducted, because the smelter fed the material at a rate significantly higher than
they usually feed iron. However, assuming a 75 percent tradeoff from metallic iron input to the furnace,
and based on the market price of scrap iron ($130 per ton), the potential cost avoidance savings are:
CFe=-360 max ([Fe]-0.1, 0)
where [Fe is the weight fraction iron in the feed. Note that the equation predicts that there is no benefit
if the iron fraction is below 0.1. The formula includes a correction to avoid double counting the iron which
needs to be disposed.
Thus, overall cost benefits and avoidance CA based on the lead in the Superfund material and reductions
of coke and iron usage is:
CA= - P [Pb] - 0.007 ([BTU] - 2,000) - 360 max([Fe] - 0.1, 0)
with [BTU] in BTU per pound and [Fe] as weight percentage metallic iron in the feed, and the lead
reclamation efficiency assumed to be 1.
4.3.5 Net Smelter Processing Cost
The net cost to process the waste materials in a secondary lead smelter, is therefore:
Cp = 256 + 150 [A] -^65 1 P) JPbl -t 100 [S] - 0.007 [BTU]
360 max([Fe]- 0.1,0)
where Cp is in units of dollars per ton of material.
21
-------
4.4 OVERALL PROCESS ECONOMICS
The overall process economics can be determined by combining the various components (excavation
transportation, and processing the materials):
where Cyrt is the overall cost of processing in dollars per tpn. Fixing the cost of excavation and required
pre-treatment at $15 per ton, inserting a function of distance (D) from the smelter (in miles) for
transportation costs, and combining the processing cost for the material, C^ becomes:
CTot= 271 + 0.34>D + ¥ + 150 [A] + 100 [SI- (165 + P) [Pb]
- 0.007 [BTU] - 360 max([Fe] - 0.1, 0)
with Cjot in dollars per ton of material. [A] is the percent ash content of the waste material; [S], [Pb], and
[Fe are the percent sulfur, lead, and iron concentrations, respectively. [BTU] is the heating vaue in BTU:
an P is the price for lead in dollars per ton.
Correction factors $ and ¥ are functions of the bulk density of the material (in ton per yd3) where * = max
[1/ p, 1) and ¥ = 0 if p> 0.8 ton per yd3 and $15 per ton if p 5 0.8 ton per yd3. Neither correction factor
will likely apply if the material is crushed on-site to increase its density and decrease handling costs before
being snipped to the smelter, but, of course, the cost of this crushing must then be added to pre-processing
and excavation costs.
Based on this cost model, the overall cost of processing the materials from the sites studied during this
research, have been calculated, and are presented in Table 6. The table includes two costs, the first based
on a conservative market price for lead ( 650 per ton) and the second based on a more plausible long-term
cost for lead ($750 per ton). Npte that the overall cost of using secondary lead smelters as a treatment
technology for Superfund sites is dependent on the lead content and market price for lead and total ash
content (i.e., slag generation potential) of the material.
TABLE 6
COST OF REMEDIATING SITES
Site
Tonolli
Hebelka
Demolition Material
NL Industries
PennDOT
P=$650/t
Cost/ton
$228
$174
$374
$80
$231
P=$750/t
Cost/ton
$224
$160
$373
$35
$228
Distance
D (miles)
40
75
100
200
250
%Ash
[A]
20
30
4
65
70
% Lead
[Pb]
3.5
14.7
1
45
3.2
22
-------
5.0 CONCLUSIONS
Lead is a common contaminant at Superfund sites, present at between 30 and 50 percent of the sites, and
reclamation is one viable option for remediating the sites. The following conclusions may be drawn from
the study of reclamation of lead from Superfund sites using secondary lead smelters.
Lead was successfully reclaimed in a secondary lead sme|ter from a variety of materials, including
battery case pieces, dross, lead debris, spent abrasive materials, and demolition material
contaminated with lead paint. The lead concentration in these materials ranged form 1 to 45
percent.
The economics of reclaiming lead from Superfund sites are dependent on lead concentration, the
market price for lead, distance from the smelter, the amount of materials which become
incorporated into slag from the process, iron content, BTU-value of the wastes, and to a lesser
extent, sulfur concentration.
The cost for recovering lead from the five sites selected for this project, based on a conservative
price for lead ($650/ton), ranged between $80 and $374 per ton of materials.
The Superfund material must generally be mixed with regular furnace feed prior to being
processed. The acceptable weight ratios of the Superfund material to overall furnace feed
Superfund plus regular materials) were found empirically to be a function of percentage lead in
' ie Superfund materials, with the approximate linear function: Mix ratio = 0.1 + 0.5[Pbfeed], where
[Pbfeed] is the concentration of lead in the Superfund material.
£
th
Overall, CHMR concludes that secondary lead smelters provide a viable alternative to stabilization and
disposal for the treatment of wastes found at battery breaker and secondary lead smelter Superfund sites,
as well as for other commonly found lead-containing waste streams.
Factors in selecting reclamation using secondary lead smelters for the materials found at Superfund sites
include:
Lead concentration;
Material type (dross, debris, slag, soil, etc.);
Density of the material;
Sulfur, calcium, iron, antimony, soil, aluminum, and silica concentrations;
Moisture content (smelters generally cannot accept wet materials);
Presence of organic wastes, haiogenated materials, and gaseous materials (which typically are
prohibited in smelters); and,
BTU-value.
In all cases, a thorough analysis of the site materials is anticipated to be necessary before the material is
accepted. In some cases, a treatability study to determine the allowable mix ratios and effects of the
material on the furnaces may be necessary before a stream is accepted by the smelter.
23
-------
6.0 RECOMMENDATIONS FOR LEAD
RECLAMATION AS A REMEDIAL APPROACH
Table 7 provides a comparison of the use of secondary lead smelting to other remediation technologies
described in the paper for a variety of feeds. The table includes all costs associated with processing each
feed in the technology indicated. The table shows that secondary smelting is the most economical
alternative for lead debris, and competitive for battery cases.
Secondary lead smelting has other advantages over the alternative technologies, including:
It is a permanent solution - The lead is reclaimed and put back to its original use. The
long term effects of other technologies, such as landfilling or stabilization, are still
uncertain.
It reduces liability - Once the material is processed in the smelter, it is no tonger a waste,
but a product permanently ending the long term CERCLA liability of disposing the waste.
It uses existing technology - This eliminates the need to develop mix ratios for solidiiition
or stabilization
Smelter receptivity - In a limited informal survey, CHMR found that virtually all secondary
lead smelters are permitted to accept lead-containing hazardous wastes, and that
approximately half were amenable to doing so.
CHMR estimates that the total amount of battery case material and other debris which could be processed
annually in the U. S. secondary smelters is approximately 300,000 tons. This is based on 25 tons per day
per furnace, and 280 production days per year. It appears at this point that several U. S. smelters are
currently planning to accept Superfund materials.
TABLE 7
COMPARISON OF SMELTING AND OTHER TECHNOLOGIES
Technology
Stabilize/Capping
Stabilize/Disposal
Physical Separation
Extractive Washing
Secondary Lead
Smelting
Media Price Range/ton
Soils
Battery Cases
Soils
Soils
Battery Cases
Soils
Battery Cases
Battery Cases
Debris
Other Materials
$30-100
$50-150
$50-200
$60-200
$60 250
$60-250
$75 - 300
$100-250
$60-200
$100-300
Comments
Limited application
Process + dispose
Metal Recovery
Metal Recovery
24
-------
7.0 REFERENCES
1. E. Earth and R. Soundararajan, "Solidification/Stabilization Treatment of Lead Battery Site Soils,"
HMCRI Superfund '90 Conference, Hazardous Materials Control Research Institute, Greenbelt,
MD, 1990, p. 665.
2. P. Watts, C. Pryately and J. Gormley, 'Characterization of Soil-Lead Contamination at an NPL Site
Using a Combination of Field Screening and Laboratory Analytical Techniques," HMCRI Superfund
'91 Conference Hazardous Materials Control Research Institute, Greenbelt, MD, 1991, p. 498.
3. J. Swartzbaugh, et al., "Remediating Sites Contaminated With Heavy Metals, Parts I and II,"
Hazardous Materials Control, November/December 1992, p. 36, and March/April 1993 p. 51.
4 . M. Royer, A. Selvakumar, and R. Gaire, "Control Technologies for Remediation of Contaminated
Soil and Waste Deposits at Superfund Lead Battery Recycling Sites," J. Air and Waste
Management, Vol 42, No. 7, 1992, p. 970.
5. A.Y. Lee, A. Wethington, and M. Gonnan, "Treatment of Lead Wastes from Lead-Acid Battery
Recycling Plants," EPD Congress 1993, J. Hager (Ed.), The Minerals, Metals and Materials
Society, Warrendale PA, p. 927.
6. W.E. Fristad, "Terramet Lead Leaching and Recovery at an Arm Ammunition Facility," Emerging
Technologies In Hazardous Waste Management V, Tedder Ed.), American Chemical Society
Industrial & Engineering Chemical Division, 1993, GA, p. 513.
7. R.J. Schmitt, "Automobile Shredder Residue - the Problem and Potential Solutions," Recycling
of Metals and Engineered Materials, J. Linden et al (Eds.), The Minerals, Metals and Material
Society, 1990, Warrendale PA, p. 315.
8. J.L. Hessling et al., "Results of Bench Scale Research Efforts to Wash Contaminated Soils at
Battery Recycling Facilities," 2nd International Symposium on Metals Speciation, Separation
and Recovery, Rome Italy, 1989, p. 183.
9. A. Saracino and C. Parent, "Soil Washing of Lead-Contaminated Soil at a Former Gun Club Site,"
Wallace-Kuhl & Associates, West Sacramento, CA, 1992.
10. E.R. Krishnan and R.J. Turner, "Overview of Metals Recovery Technologies for Hazardous Waste,"
National Research & Development Conference on Control of Hazardous Materials HMCRI,
1991.
11. S. Paff, "Using Secondary Lead Smelters for the Recovery of Lead from Rubber Battery Cases and
Other Materials," Center for Hazardous Materials Research, Pittsburgh, PA, 1992.
12. W.F. Kemner and E.R. Krishnan, "Electromembrane Process for Recovery of Lead from
Contaminated Soils," Innovative Waste Treatment Technology Series, 1990, p. 87.
13. R.S. Simms and K. Wagner, "Treatment Technologies Applicable to Large Quantities of Hazardous
Waste Contaminated Soils," USEPA Municipal Environmental Research Laboratory Cincinnati, OH,
1991.
14. T. Scherer and D. Gillette, "Affordability Analysis of Lead Emission Controls for a Smelter-Refinery",
EPA Control Technology Center, Research Triangle Park NC, 1989.
15. W. Urban and S. Krishnamurthy, "Remediation of Lead Contaminated Soil," Foster Wheeler
Enviresponse, Edison, NJ, 1993.
25
-------
16. A. Davis, M. Ruby and P. Bergstrom, "Geochemical Controls on the Bioavailability of Lead from
Mine Waste Impacted Soils," HMCRI Superfund '91 Conference, HMCRI, Greenbelt, MD, 1991,
p. 564.
17. S. Griffin, "Application of USEPA's Uptake Biokinetic Model for Establishing Soil Lead Cleanup
Levels at Superfund Sites," HMCRI Super-fund '91 Conference, HMCRI,Greenbelt, MD, 1991
p. 495.
18. Center for Hazardous Materials Research, 'Site Characteristic Profile Study and Competitive
Technology Analysis," May, 1991.
19. M. Royer and T. Basu, "Selection of Control Technologies for Remediation of Lead Battery
Recycling Sites,' EPA Risk Reduction Engineering Laboratory, Edison, NJ, 1991.
20. T. deGrood, "Appeal of stabilization, solidification grows for treating wastes, HazMat World,"
February, 1991, p. 60.
21. USEPA, "Solidification/stabilization of organics and inorganics," Engineering Bulletin, EPA/540/S-
92/015, May, 1993.
22. D.S. Schleck, "Treatment, When Does It Apply?." HMCRI Superfund '90 Conference, HMCRI,
1990, p. 677.
23 M Royer "Selection of Control Technologies for Remediation of Lead Battery Recycling Site
Wastes,' USEPA Engineering Bulletin, EPA/540/S-92/011, 1992.
24. D.R. McCombs, "Operating a Hazardous Waste Landfill, HazMat World, September, 1991, p. 73.
25. R. Gager, "Hazwaste Landfills Struggle for Growth and Safety," HazMat World June, 1991, p. 46.
26. Center for Hazardous Materials Research, 'Acid Extraction Treatment System for Treatment of
Metal Contaminated Soils," September, 1993.
27. H. Masters and B. Rubin, "EPA's Mobile Volume Reduction Unit for Soil Washing," USEPA's RREL,
Proceedings from the 7th Annual RREL Hazardous Waste Research Symposium, Cincinnati,
OH, 1991, p. 89.
28. Brice Environmental Services Corporation, "Qualifications for Soil Washing Services," 1992.
29. R. Coleman and R. Vanndervort, "Evaluation of Paul Bergsoe and Son Secondary Lead Smelter,"
USEPA Office of Research and Development, EPA-60072-80-022, 1980.
30. P. Queneau, D. Cryar, and D. Mickey, "Optimizing Matte and Slag Composition in Rotary Furnace
Smelting of Lead Residues," Primary and Secondary Lead Processing, Jaeck (Ed.), The
Metallurgical Society/Pergammon Press, New York, 1989.
31. R. Isherwood et al., "The Impact of Existing and Proposed Regulations Upon the Domestic Lead
Industry," U.S. Bureau of Mines, Denver, CO, 1988.
32. P. Queneau and A. Troutman, "Waste Minimization Charges Up Recycling of Spent Lead-acid
Batteries," HazMat World, August, 1993, p. 34.
33 A.D. Zunkel and J.C. Taylor, "Integrated Primary/Secondary Lead Smelting," Journal of Metals,
January, 1988, p. 32.
34. Center for Hazardous Materials Research, "Quality Assurance Project Plan and Work Plan for
Using Secondary Lead Smelters for the Recovery of Lead from Rubber Battery Cases and Other
Waste Material," October, 1991, pp. 6-11.
35. Personal correspondence between the author and Charles Faust of Earth Treatment Technologies
(Dutton Mill Industrial Park, 396 Turner Way, Aston, PA 19014, February 1994).
26
-------
APPENDIX A
RECLAMATION OF MATERIALS FROM BATTERY CASE PILES
AT THE TONOLLI SUPERFUND SITE IN
NESQUEHONING, PA
Description of Evaluation and Results and Discussion
July 23, 1992
27
-------
1.0 DESCRIPTION OF EVALUATION
The purpose of the evaluation was to determine the feasibility and economics of the recovery of lead from
battery case materials. The Tonolli Corporation Superfund site in Nesquehonina Pennsylvania was used
as the source of these materials. The materials were processed at Exide's Reading, Pennsylvania,
secondary lead smelting facility.
Exide's Reading smelter is primarily engaged in recycling lead from lead-acid batteries, such as those
utilized in automobiles and the recycling of polypropylene rom battery cases for subsequent reuse. Exide
uses both reverberatory and blast furnaces to recycle lead. These furnaces and other plant operations are
typical of the secondary lead smelting industry.
This appendix incjudes a description of the evaluation, and a discussion of the results. The conclusions
from this evaluation have been ncorporated Into those found in the main body of this report.
1.1 EXPERIMENTAL DESIGN
During the evaluation, the material from the Tonolli site was fed to both the reverberatory and blast
furnaces. Initially, the test material was fed directly to each furnace using a front-end loader, which loaded
one toad of test material, then several |oads of regular feed (depending on the mix ratio), then another bad
of test material, and so on. However, it became apparent that the furnaces would better handle the test
material if it were initially mixed with the regular feed, and the resultant mixture charged to the furnaces.
The plant is equipped with two blast and two reverberatory furnaces. Generally one reverberatory furnace
and one blast furnace were utilized for processing the Tonolli material, which the second reverberatory
furnace and blast furnace remained in normal, routine operation. Conducting the test in this manner
allowed for comparison of the operation of the lest" furnaces with the "reference" or "control" furnaces.
During the evaluation, the operating parameters (fuel usage, oxygen usage, slag production, lead
production, etc.) for each set of furnaces were obtained. In addition, samples of slag and lead were
obtained from each furnace to be used to determine effects on production quality.
1.2 CHRONOLOGY OF EVALUATION
The following represents a chronology of the activities conducted in support of the evaluation, between
September Sand 13, 1991.
1.2.1 Material Acquisition and Sampling
On September 5, 1991, five dump trailers of material were loaded with battery case materials from the
Tonolli Superfund site in Nesquehoning, Pennsylvania. Because the battery case materials were stored
in piles at the Tonplli site and no separation of materials was necessary, no problems were encountered
during loading activities. No dusting problems were encountered during loading activities. A total of
167,750 pounds of material was transported in five dump trailers, ranging from 42 to 48 cubic yards of
material per trailer.
Composite samples of each battery case pile were collected during the loading process at the Tonolli site.
The samples were subsequently analyzed to determine:
bulk density;
percent of the hard rubber material in the total sample;
percent of polypropylene plastic in the total sample;
percent of other miscellaneous material (i.e. soils, concrete, etc.);
28
-------
percent of metallic lead in the total sample;
lead concentration using the Toxicity Characteristic Leaching Procedure (TCLP) on rubber
cases;
total sulfur; and
percent of lead contained in the rubber, soil, concrete, etc.
The analytical results of this material are presented in Table"! A-1. A discussion of these results appears
later in this report.
1.2.2 Major Activities
This section describes the major activities that occurred during this evaluation. These activities are
presented chronologically starting with the first day of processing through the smelter.
Day One
Activities completed during the first day of the test included density determination of test material, which
was determined by weighing trucks as they entered and exited the plant, and dividing by the truck bed
volume. One blast furnace was charged at a 7 percent test material feed ratio and one reverberatory
furnace was charged at an 11 percent feed ratio, in accordance with the QAPP. The test material feed
ratio is calculated by dividing the weight of "test" material charged to the furnace by the total weight of feed
(including "test' material and typical feed).
Day Two
Early in the second day, both blast furnaces were running sluggishly and there was a crust accumulating
within the reverberatory furnace which was being charged with test material. Based on the amount of coke
added during the previous night, the plant management concluded that the blast furnaces were overcoked.
This was unrelated to the evaluation, but Exide management still decided that it would be best to refrain
from charging additional test feed to the blast furnaces until the build up within the furnaces was reduced.
The accumulation of the crust in the reverberatory furnace was determined to be related to the experiment.
Therefore, test feed input to that furnace was suspended.
29
-------
TABLE A-1
ANALYTICAL RESULTS FROM TEST MATERIALS
Sample
1
2
3
4
5
6
7
8
Type of Sample
Composite from trucks
taken at site
Composite from trucks
taken at site
Composite from trucks
taken at site
Composite from trucks
taken at site
Composite taken at
smelter
Composite taken at
smelter
From piles on site
From piles on site
(duplicate of 7)
Bulk
Density
(lo/ty
24.4
23.2
I6.9*
28.4
•^
™
— '
Hard
Rubber
(wt%)
76.3
70.7
67.9
79.9
~
—
—
Plastic
(wt%)
22.5
27.8
27.5
18.0
—
—
—
•~
Concrete
Soil, etc.
(wt%)
1.0
2.6
2.6
2.0
— -
-—
—
3.0
Sulfur
(wt %)
3.10
_t
.•••
—
1.93
2.71
—
—
Metallic
Lead
(wt%)
0.25
0.05
0.05
0.10
—
~
—
0.10
Total
Lead
(wt%)
0.34
0.82
0.47
1.57
15.7
7.3
1.15
0.28
TCLP
Lead
(mg/L)
130.9
111.8
131.3
159.4
™~
•—
—
—
* Value was found to be anomalous by q-test at 90% confidence limit
Notes: Samples 5 and 6 contained notipeable quantities of reddish sludge, which apparently was a lead oxide compound. Therefore the total lead
concentration In these samples is significantly higher than those for other samples.
Additional bulk densities were calculated by obtaining weights and volumes from the trucks. These were calculated to be: 29.,
27.3, 23.7, 31.4 and 26.6 Ib/ft3 respectively on trucks 1 through 5. The overall average of all the values obtained was 26.7 ±1.7
Ib/ft3.
-------
The test material was not fed into the blast or the reverberatory furnace again until later in the afternoon.
At that time, the test feed was introduced at an 11 percent weight ratio into blast furnaces #1 and #2 and
reverberatory furnace #1. The test material was fed to both blast furnaces to ensure that both furnaces
would contain the same material at the start pf the third day of the experiment. During the third day, Exide
planned to use blast furnace #1 for processing of test materials.
Day Three
The test material feed ratio in Blast Furnace #1 was increased to 13 percent and Blast Furnace #2 was
fed regular feed to serve as the control. The test material was pre-mixed with the regular feed prior to
charging the furnaces to eliminate layering of test materials within the furnaces. Exide decided to reduce
coke input to the blast furnaces because the test material was serving as a supplemental source of carbon.
The reverberatory furnace remained at a feed ratio of 11 percent of the test material. Again, material
accumulated in the reverberatory furnace throughout the day.
Day Four
The reverberatory furnace showed continued signs of material buildup, crusting and sluggish performance.
Exide decided not to feed test material to the reverberatory furnaces until the next day. The blast furnace
test feed ratio was held constant at 13 percent test material.
Day Five
Beginning at 8:00 AM on the final day of the test, the reverberatory furnace was fed at a ratio of 7 percent
test material. No crust accumulation in the reverberatory furnace was noted. However, the reverberatory
furnace still performed sluggishly. Exide plant management hypothesized that the sluggish performance
was attributed to the slow rate of burning associated with the relatively large pieces of matenal in the test
feed. They concluded that it was not viable to process Tonolli site materials in the reverberatory furnaces
without size reduction of the battery case materials.
Both blast furnaces were fed test material at 20 percent by weight until the operation of one of the furnaces
was discontinued in order to complete scheduled repairs and maintenance. The testing was continued
using the second blast furnace and the overall evaluation was concluded by 10:00 PM on September 13,
1991 when all the Tonolli material had been processed.
31
-------
2.0 RESULTS AND DISCUSSION
In order to assess the viability and economics of the process, CHMR/Exide recorded the settings of various
plant parameters, the production figures for both sets of furnaces, and stack parameters. In addition
CHMR obtained samples of the input material slag and lead. The sampling was conducted in accordance
with the QAPP. The results from the sampling and analysis are presented and discussed below.
All statistical analyses were conducted according to the procedures outlined in Section 5 of the QAPP. All
confidence ranges, q-tests, analyses of variance (ANOVA analyses), etc. were performed based on 90
percent confidence limits.
In-depth analyses were performed only on the blast furnace data. No statistical analyses were conducted
on reverberatory furnace parameters, because the experienced furnace operators had already determined
that the reverberatory furnaces could not accept the battery case material feed as received from the Tonolli
site.
2.1 TEST MATERIALS
The test materials consisted of pieces of rubber and polypropylene, typically shredded to approximately 1
to 3 inches in size, combined with reddish-brown sludge, dust, bits of concrete and iron, small wood scraps
and pieces of cloth. Because of the varying particle sizes and shapes, the material was not densely
packed.
The test materials were sampled for various parameters, including density, lead content, sulfur content,
TCLP values, percentage rubber, and other parameters. The results from these analyses are presented
in Table A-1, which shows lead concentrations ranging between 0.28 percent and 15.7 percent, with an
average of 3.5 percent.
The wide range in analytical results from analyses of the Tonolli material was attributed to non-homogeneity
in the material, rather than errors in the analyses. CHMR reviewed the analytical procedure and QA/QC
notes for the analyses, and determined that no obvious errors in analyses existed. Analyses 1 through 4
were of samples comppsites of the material on the trucks loaded from the site (composited from four
different grabs of material taken from the trucks). Analyses 5 and 6 were each composites of material
obtained from five different locations in the pile of material present at the smelter. Analyses 7 and 8 were
duplicate analyses of material from one location in the pile. The composites for analyses 5 and 6 included
one grab which contained large amounts of reddish brown sludge. This sludge apparently contained a very
high fraction lead. Thus, by averaging all the analytical results, CHMR obtained an average which was
representative of approximately 30 grab samples taken of the material. Based on this, CHMR concluded
that much of the Tonolli material contained relatively low concentrations of lead, with scattered higher
concentrations, and an average concentration of 3.5%.
2.2 FURNACE INPUT PARAMETERS
The input parameters recorded included air flows, oxygen consumption, back pressure, and a record of
materials introduced into the furnaces. These parameters were recorded by Exide and CHMR personnel
during each shift. Table A-2 presents a summary of effects of the test material on the various parameters.
2.2.1 Blast Air
Air enriched with oxygen is used in the blast furnace to oxidize coke and other combustible material to
maintain a sufficiently high temperature to melt and fuse the material in the furnace. The air is enriched
with oxygen to produce higher flame temperatures, more efficient burning, and as a side benefit, smaller
bjowers are required to introduce extra oxygen than are required to move the required larger volumes of
air. The required rates of air and oxygen inputs to the furnaces are determined by the operators and set
manually. The air and oxygen settings are recorded at the beginning of each shift, and whenever they are
changed.
32
-------
TABLE A-2
SUMMARY OF EFFECTS ON BLAST FURNACE PERFORMANCE
Parameter
Air
Oxygen
Back Pressure
SO2 Emissions
Lead Production
Slag Production
Lead Concentration in Slag
Lead Concentration in Slag
Sulfur in slag
Antimony in Slag
Silica in Slag
FeO in Slag
Effect of Adding Test Feed
to Furnace
Increase on furnace #2
No observed difference
Increased
No observed difference
No observed difference
No observed difference
Increased from 2.2 to 3.7%
Reducing environment
No observed difference
No observed difference
No observed difference
No observed difference
Probable Cause
Probably due to build-up;
otherwise insignifcant
Partially due to over-coking
Over-coking
Too much carbon
Response
Small problem no
response is necessary
Decrease particle size
Reduce coke input
Reduce coke input;
increase iron input
The relative air consumptipn in Blast Furnace #2 while regular feed only was processed was 6.75±0.25,
and while the Tonolli material was processed was 7.42±0.12. The difference in air consumption in Furnace
#2 between times when test feed was processed and when it was not appears to be largely due to a
significantly decreased air consumption rate on Wednesday, September 11, which occurred when no test
feed was processed in the furnace. It also correlates to a time when the overall production rate of the
furnace decreased, possibly due to attempts by the operators to reduce coke buildup in the furnace. This
buildup was only partly attributable to the test feed.
2.2.2 Back Pressure
The furnace back pressure is a measure of the pressure against which the air blowers must push in order
to force air into the furnace. The back pressure increases as the air required by the furnace increases.
At constant air, the back pressure is a measure of the resistance of the material inside the furnace to the
flow of air. The back pressure increases when small particles are fed to the furnace or if material begins
to back up inside the furnace.
The use of test feed caused an increase in back pressure in the blast furnace. The average back pressure
while regular feed was processed was 4.95+0.6 inches of water, and 7.42+0.4inches of water when test
feed was charged to the furnace.
Although some of the apparent correlation between back pressure and test feed input may be attributable
to buildups within the furnace caused by over-coking, the evidence is strong that the Tonolli material
significantly contributed to the back pressure increase. The most likely reason for the increase in back
pressure while processing Tonolli material is that the Tonolli material was large enough that it burned
relatively slowly, but sufficiently small to fill some of the pore spaces inside the furnace feed shaft.
The increase in back pressure seen during the evaluation probably had little affect on overall furnace
performance or overall process economics. However, as the back pressure increased, the blowers had
to work harder to blow air into the furnace, and at higher back pressures (above 30 or 40 inches of water),
the blowers will stall. Therefore, based on the data, which suggest a 5 inch back pressure increase per
10 percent addition of battery case material, the maximum sustainable battery case material feed ratio is
between 30 percent and 50 percent by weight.
33
-------
2.3 FURNACE OUTPUT PARAMETERS
CHMR monitored the output of parameters such as sulfur (in the form of S02 and calcium sulfate sludge)
lead, and slag.
2.3.1 Sulfur
The Tonolli material was found to contain 2.6±1.0 percent sulfur. Overall, this concentration did not
represent a significant difference with regular feed material, which may contain 2 percent to 5 percent
sulfur. Therefore, no significant differences in overall plant sulfur emissions were anticipated. However,
CHMR/Exide monitored the fate of sulfur in the process to ensure that processing the test material did not
result in an increase in production of calcium sulfate sludge, S02 emissions, or sulfur concentration in the
slag.
Approximately 50 percent of the sulfur entering a blast furnace leaves as oxides in the flue gas stream.
Nearly 100 percent of this sulfur at Exide's Reading smelter is removed as calcium sulfate sludge from the
S02 scrubber system. However, a small portion of the sulfur in the flue gas is released as stack emissions.
Therefore, CHMR/Exide monitored both the calcium sulfate sludge generation and SC^ concentrations in
the stack emissions.
SOo Emissions
CHMR monitored both stacks at the smelter. Each stack system is connected to the corresponding blast
and reverberatory furnace. The average sulfur stack emission in system #1 was 64±6 ppm, while the
average stack emission in system #2 was 69±9 ppm. Both of these values were below Exide's permitted
average of approximately 110 ppm. No statistically significant differences exist between the emissions from
the two systems. The overall average plant SO, concentration was 66±6 ppm.
ANOVA analyses indicate no differences were demonstrated between S02 emissions when the test material
was being processed, and when it was not being processed, for either set of furnaces. In addition, there
is no significant correlation between SO, concentrations as a function of test material input ratio. Therefore,
CHMR/Exide conclude that the Tonlti feed material did not have a significant affect on the SO2
concentration in the stacks.
Calcium Sulfate Sludge Production
During each shift, specific gravity measurements of slurry generated from the air emission control
equipment (scrubbers) were taken by the operator on duty. The specific gravity is related to the calcium
sulfate concentration in the slurry, and therefore gives a measure of total sulfur uptake in the sludge.
Typical slurry specific gravity is between 1 .1 and 1 .3. Slurry specific gravities for both air emission control
systems averaged 1.18±0.02, and showed no correlation with test feed input.
At the end of each shift, the amount of slurry was calculated by counting the number of 10 cubic yard
tankers transported from the scrubbers to the on-site sludge dewatering facility. The number of tankers
varied only between 2 and 4 per shift, with a total of 9 trucks per day except Thursday, September 12,
when 8 trucks were transported. Normally, between 8 and 10 tankers per day are transported. Therefore,
CHMR/Exide conclude that there was no significant increase in calcium sulfate sludge production due to
processing the Tonolli material in the blast furnaces.
2.3.2 Slag and Lead Production and Quality
The amount of slag and lead produced by the blast furnaces were critical experimental parameters. The
amount and quality of lead produced is important for obvious reasons, as lead is the salable product from
smelter operations. The slag generation rate was also important, as slag disposal represents a significant
operating expense. The lead concentration in the slag is also important because it represents a potential
loss of salable lead product.
Lead and slag production for each furnace during each shift were measured. ANOVA statistical analyses
were performed to determine if significant differences existed between the amounts of slag or lead
produced while the test materials were run versus the amounts produced while only regular feed was
processed in the furnaces. The tests indicated no statistically significant change in lead or slag production
34
-------
rates while the test feeds were run in either furnace. The only significant drop in production during the
week occurred in Furnace #2 on September 10, when lead production dropped due to over-coking.
Lead samples from the blast furnaces were obtained during each shift. The samples were analyzed by
Exide personnel using Spectrometer Lab Test 201 equipment to determine concentrations of trace metals
in the lead. The results from the lead analyses indicated no significant increase in contaminants in the lead
product.
Composite samples of slag were taken as the slag was being tapped from the furnaces. The molten slag
samples were collected by the Exide operators by inserting a shovel into the furnaces in accordance with
the QAPP. Table A-3 shows the results of the analyses of slag samples taken from the furnaces when they
were fed regular feed , and regular feed and test feed mixed (with the test feed ratios noted). The table
shows no consistent differences in sulfur, silica, antimony, and iron oxide concentrations when the test feed
was processed in the furnaces. The results do show a significant difference in lead concentrations in the
slag. When only regular feed was processed, the lead concentration was 2.21+0.6 percent. When Tonoiii
material was added^ the lead concentration increased to 3.7+0.7 percent.
Much of this difference may be attributable to the problems encountered when both furnaces were over-
coked earlier in the week. The over-coking problem was intensified by the addition of test feed. High coke
increases the concentration of lead in the slag. On Wednesday, September 11, the plant responded by
reducing the amount of coke and increasing the iron fed to the furnace to maintain a reducing environment
with less coke. The data show a decrease in the lead content in the slag beginning during the third shift
on September 11 (samples 13 and 14) and continuing throughout the remainder of the week. By
September 12 and 13 (samples 15 and 16), the lead concentration in the slag was reduced, and a relatively
stable slag composition was maintained at 2.5 percent, which is approximately equal to the lead
concentration in slag when no Tonolli input was fed to the furnace.
35
-------
TABLE A-3
BLAST SLAG ANALYTICAL RESULTS
REGULAR FEED ONLY
Sample
1
2
3
4
5
6
7
AVERAGE
Lead (%)
2.8
3.1
1.1
2.7
1.9
2.5
1.0
2.2±0.6
Sulfur (%)
3.0
3.9
5.2
3.0
4.9
5.0
2.9
4.0+0.8
Si02 (%)
24.8
29.5
15.1
24.4
21.7
4.7
23.9
23+3.4
FeO (%)
30
32
63
36
46
30
43
40+8.6
Antimony (%)
1.5
0.1
—
0.5
0.5
1.0
-
0.72+.0.5
REGULAR FEED AND TEST FEED MIXED
(% test feed in parentheses next to sample number)
Sample
8 (7%)
9 (7%)
10 (11%)
11 (11%)
12 (13%)
13 (13%)
14 (13%)
15 (13%)
16 (20%)
Average
Lead (%)
3.3
5.3
4.8
2.6
3.3
4.7
4.5
2.5
2.5
3.7 + 0.7
Sulfur(%)
1.4
3.6
5.9
2.1
1.2
4.5
5.2
1.8
3.2
3.2+1.0
SiO2 (%)
33.9
30.0
21.4
32.6
32.6
21.2
2.36
28.0
18.8
27+3.7
FeO (%)
29
24
47
33
30
44
42
48
55
39±6
Antimony (%)
0.3
0.2
2.2
0.6
0.2
0.4
0.4
0.4
0.2
0.5+0.4
All confidence intervals based on 90% confidence level.
36
-------
APPENDIX B
RECLAMATION OF LEAD FROM THE NL
INDUSTRIES SUPERFUND SITE IN
PEDRICKTOWN, PA
Description of Evaluation and Results and Discussion
November, 1993
37
-------
1.0 INTRODUCTION
The Center for Hazardous Materials Research (CHMR) and Exide Corporation (Exide) performed a joint
research study to determine the feasibility and economics of using secondary lead smelters for the
recovery of lead from leadcontaining materials. Under the project CHMR/Exide conducted an
evaluation to determine the feasibility and economics of reclaiming lead from the NL Industries, Inc.
Superfund site in Pedricktown, NJ. This evaluation was conducted in two stages: a preliminary
investigation and a larger-scale evaluation.
As part of the preliminary investigation, a total of 19 trucks with 20 to 40 cubic yard capacities were
loaded with lead-containing material on January 29 and 30, 1992. The trucks represented different
types of material found on the Pedricktown site with lead contents ranging between 36 and 65 percent.
A total of 372 tons of material were processed during this initial investgatton. Information gathered
from the preliminary test assisted CHMR/Exide in determining the feed rates and types of material to be
processed through the secondary lead smelter.
The purpose of the preliminary study was to process the different types of material through the furnaces
and determine the initial response of the furnaces to the material. This provided CHMR/Exide with the
information necessary to complete the next stage of the evaluation, which consisted of processing the
material over a longer time through the secondary lead smelter.
After receiving permission from the EPA Project Officer and the EPA Remedial Project Manager (RPM),
the second stage of the investigation was initiated. Materials were excavated, loaded, and transported
to Exide's secondary smelter in Reading, PA. The material was weighed and stored in an area isolated
from other feed material prior to beginning the evaluation. During the second evaluation, which
occurred over three months, approximately 1200 tons (2.4 million pounds) of Pedricktown material were
processed.
38
-------
2.0 DESCRIPTION OF EVALUATION
The purpose of the evaluation was to determine the feasibility and economics of processing the
material from the NL Industries, Inc. Superfund site using secondary lead smelting technology. The
initial step was a preliminary test during which 19 truckloads were processed. This was followed by a
larger-scale investigation that spanned three months.
2.1 RECLAMATION PROCESS
The reclamation of lead from Superfund and other lead-containing materials is based on existing lead
smelting procedures and basic pyrometallurgy. The materials are first excavated from Superfund sites
or collected from other sources. Then they are preprocessed to reduce particle size, and to remove
rocks, soil, and other debris. The materials are then transported to the smelter.
At the smelter, the materials are fed either to the reverberatory or blast furnaces, depending on particle
size or lead content. The two reverberatory furnaces normally treat lead from waste lead-acid batteries
as well as other lead-containing material. The furnaces are periodically tapped to remove slag, which
contains 60 to 70 percent lead, and a soft, pure lead product. The two blast furnaces treat the slag
generated from the reverberatory furnaces, as well as larger sized leadcontaining waste. These
furnaces are tapped continuously for lead, and intermittently to remove the slag, which is transported
off-site for disposal. The reverberatory and blast furnace combination at Exide can reclaim lead from
batteries and waste with greater than 99 percent efficiency.
2.2 SMELTING PROCESS DESCRIPTION
Exide's Reading, PA, secondary smelter is primarily engaged in the reclamation of lead from spent
lead-acid batteries and other lead-containing materials. Exide reclaims lead using bpth reverberatory
and blast furnaces. The plant operations are typical of the secondary lead smelting industry.
2.3 EXPERIMENTAL DESIGN
The material obtained from the Pedricktown site varied in size, lead content, and overall texture and
structure. These factors determined the method of processing through the secondary lead smelter.
The majority was processed through the reverberatory furnaces, while the blast furnaces were used for
larger pieces of material, such as lead hard heads (large chunks of metallic lead) and ingots found at
the site.
Approximately 372 tons of material were processed during the preliminary investigation. Since the
objective of the preliminary investgation was to study the effects of the Pedricktown material on furnace
performance, this stage of the evaluation featured the most diverse variety of material.
During the large-scale evaluation, CHMR/Exide processed approximately 1200 tons of material. The
majority of material was processed through the reverberatory furnaces over a two mpnth period. During
the third month, the material had to be processed through the blast furnaces due to its size and
composition.
2.4 EXPERIMENTAL FEED
Tables B-1 through B-4 provide the results of detailed analyses of the NL Industries site material during
the preliminary and large-scale investigations. Table B-1 shows the results of analyses of samples
taken during a preliminary field reconnaissance to the site during December 1991. The parameters
analyzed included lead, antimony, calcium, silicates, sulfur, arsenic, tin and iron were analyzed mainly
to determine the metallurgical aspects of the feed and its suitability in the smelting process. Further
analyses were conducted of the material which was fed to the furnaces during January 1992. The
analytical results are presented in Table B-2. Composite samples of the material were taken as it
arrived in trucks at the smelter. Analyses of lead and silica were conducted to provide baseline
metallurgical data for the materials. Fluorine and chlorine were analyzed to quantitatively demonstrate
39
-------
that the material contained negligible quantities of these elements, and no halogenated compounds.
Aluminum was analyzed because there was a concern among plant management that the material may
contain some aluminum, which is prohibited in the furnaces, and sodium was analyzed for because it
represented a potential impurity in the lead products. In general, the results showed no major areas of
concern for the material to be fed to the smelter, and a feedstock which contained over 50% lead.
Table B-3 shows the results of analyses conducted on composite samples taken of the material sent to
the smetter during the longer term evaluations, August through October 1992. The analyses included
moisture analyses for the August and September material, because moisture levels appeared to be a
potential concern to furnace operators. The results showed that the levels were not pf concern. The
material shipped during October which was fed to the blast furnace, appeared to be similar to
reverberatory slag and was composed primarily of lead and iron One load of wooden pallet material
was analyzed for Btu content, since it was anticipated to burn in the furnaces.
2.5 CHRONOLOGY OF THE EVALUATION
The following provides the schedule and different methods used for processing the material during the
preliminary and large-scale investigation.
2.5.1 Preliminary Test - January 1992
Nineteen trucks brought a wide variety of the debris from the site to the smelter at Exide Initially, test
material was fed directly into the reverberatory furnaces as 100 percent of the total feed. This caused
various malfunctions in the furnace, including build-ups in the furnace and breakdowns of the belt feed
system. The material was successfully fed at a 50 percent weight ratio into the reverberatory furnaces.
No material was fed to the blast furnaces.
2.5.2 Large-Scale Test - August through October 1992
The larg e-scal e investigation took place through the three months that were used to process the
materi al. During the first two months, the material consisted primarily of dross, slag, and debris, which
was processed mainly through the reverberatory furnaces. The material during the third month
consisted of large pieces of slag, which was processed through the blast furnaces at the smelter. The
slag also featured a lower concentration of lead than the previous material. The following paragraphs
provide a short summary of the materials, and the methods used for processing the material.
2.5.2.1 August
The initial 19 trucks for the large-scale investigation transported various types of dross material, which
were mixed with regular feed at a 40 percent (by weight) ratio and fed into the reverberatory furnace.
The larger sized material was sorted out and processed through the blast furnaces. These operations
continued until the initial truckloads of material had been depleted. The average lead concentration of
the dross material was 48 percent.
40
-------
i
3
I
1
*ST
< £
H O
ei|i
ELD SAMPLE
PRIES (PEDRM
DECEMBER 1
C w
Q
Z
_J
Z
y__
-p c
.c *
S 1^
c n.
"c
0
H
11
*- H,
!», ©
3 P
™ 1U
«s *?
12
B
E s?
- o I
(BJ C5 0
1 ||
e < •— :
o
S *"**
S T, §
— re o
| Jl
Sample Type/Notes
>5
o
T™
O
T™
O
<4«*
IT"
sf
S
o
2
V*
O
55
o
(0
r*
i
Dross from Pedricktov
o
»r~
>5
CO
o
CM
o
S«
o
«o
Duplicate of Sample A
se
CM
6s
O
*••
T™
O
S
o
j5
w
o
10
0
S3
in
Ji
o.
jS
to
1
I
g
«f
w
e
t3
00
00
g
^-*
5?
5
*~
es
o
o
4~*
•r-
S^
GO
(O
S
Q
co
o
5?
T-
C3
T-
gS
ft
a
«T"
I
CM
IB
W
S
13
S8
(O
r-
o
cs
c*> o
CM CM
0 O
# f
T- O
o o
s« a?
CN CO
CO CD
W «M
CO
o
CM 5r
T- e»
mJ9 ~*JS p^Jffil «Jp i^B ij^ffl A^M »,«i
f"i ^3T *•"• •** "*™* OO iW f'if
10 liO 0) 0) (0 t"M 10 fs«f
(0 O 10 (0 O M^
P S
«5 W
til %
*~ OL
S s?
0 $
a ,« .«
£ ^ I 1 - c
1 1 1 1 1 I
3 IS 55 S? "» -0
tj » a ? ,g Is §
S 1 1 1 1 i I I
«,y"X3t3°* s^Q
Sas?^ g'll-o
•-^rjtb— IB 0 o
D Q >-~ oo t/) 03 _1 CC
ss
fO
O
sP
5s-
O
^™
O
B«
^
^
c>
-9, «»5 »5 ^ sj
n T- tn J_ J_
SJ; 3 § 5 g
Lead debris
Soft lead dross
Orange/yellow Dross
Black Dross
AVERAGES;
0
Q.
E
m
ff)
"5
X
41
-------
TABLE B-2
FEED STREAM ANALYSES NL INDUSTRIES EVALUATION
PRELIMINARY EXPERIMENT JANUARY 1992
Quantity Concentration
Sample Type/Notes
LEAD DEBRIS
LEAD DROSS
LEAD SLAG
SLAG AND DEBRIS
SLAG AND DEBRIS
SOFT LEAD DROSS
LEAD SULFATE
SLAG AND DEBRIS
LEAD HARD HEAD MATERIAL
BATTERY CASING AND DEBRIS
MATERIAL FROM INSIDE BLDG
SLAG AND DEBRIS
COMPOSITE OF TWO PILES AT SITE
SLAG AND DEBRIS (incl. B2 P3&4)
SLAG AND DEBRIS
TOTAL
Processed Lead
(Ibs) (percent)
40,530
87,310
18,280
42,620
81,740
44,210
30,050
34,950
40,460
32,680
77,040
43,070
45,100
84,360
42,070
744,470 Ibs
57%
61%
54%
61%'
61%
54%
59%
65%
36%
53%
55%
65%
63%
59%
61%
Flouride
(percent)
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
NA
<0.0001
NA
<0.0001
<0.0001
Chloride
(percent)
0.51 %
0.06%
0.12%
0.11%
0.04%
0.08%
0.06%
0.14%
0.04%
0.08%
NA
0.14%
NA
0.24%
0.11%
Silicates
(percent)
6.9%
3.1%
3.5%
3.4%
1 .5%
1 .6%
7.9%
2.4%
5.5%
4.7%
4.2%
2.4%
3.3%
3.63%
3.4%
Aluminum Sodium Other analyses
(percent) (percentages)
0.55%
0.16%
0.18%
0.41%
0.22%
0.26%
0.12%
0.22%
0.40%
0.85%
NA
0.22%
NA
0.34%
0.41 %
1.35%
0.38%
0.28%
0.59%
0.62%
0.52%
0.12%
0.16%
0.65%
0.13%
NA S: 3.3 CA:0.9 Sb: 0.6%
0.16%
NA S: 2.7 Ca: 0.3 Sb: 0.8%
0.68%
0.19%
Weighted Average:
58% 0.00% 0.13% 3.59% 032%
0.4S%
-------
TABLE B-3
FEED STREAM ANALYSES NL INDUSTRIES EVALUATION
AUGUST - OCTOBER 1992
Quantity
Processed
Sample Type/Notes (Ibs)
Metal Concentration
Lead Moisture Antimony
(percent) (percent) (percent)
Silicates Sulfur Iron (FeO) Other
(percent) (percent) (percent) analyses (%)
AUGUST MATERIAL
Lead Dross & debris
Lead Dross & debris
Lead Dross & debris
Slag & debris
Slag & debris
133,779
112,656
105,615
182,470
262,700
43%
55%
54%
44%
47%
9.8%
10.4%
9.1%
12.6%
8.7%
0.8%
NA
NA
0.6%
3.4%
5.7%
3.4%
7.9%
4.2%
3.7%
2.7%
NA
NA
3.0%
0.9%
9.6%
NA
NA
1 .4%
4.6%
Ca 0.53%
SEPTEMBER MATERIAL
Lead debris 486,630 57%
Black dross 138,140 51%
Iron cans 13,230 <1%
Battery Cases Sdebri 60,250 44%
Pallets 39,810 7%
Baghouse bags 49,000 69%
OCTOBER MATERIAL
Wooden pallets from I 10,300 6.5%
Slag 600,900 32%
Slag & debris 261,820 28%
Total Processed: 2,457,300 Ibs
Weighted avg. composition: 43.1%
13.4%
11.7%
NA
9.6%
NA
NA
NA
NA
NA
11.3%
NA
2.3%
NA
NA
NA
NA
NA
NA
NA
2.0%
NA
NA
NA
NA
NA
NA
NA
NA
NA
4.7%
2.1%
0.8%
NA
1.0%
NA
NA
NA
NA
NA
4.0%
NA
> 80% Iron feed
5.1%
NA
NA
NA Btu/lb: 4945
2.6%
21.7% Ca: 0.14%
8.5%
2.5.2.2 September
In September, a total of 19 truckloads containing various types of lead debris were processed through the
smelter's reverberatory furnace. The material varied in overall weight and density, with an average lead
concentration of 53 percent. In addition, the material had relatively high amounts of calcium and therefore
needed to be processed through the reverberatory furnaces at tower feed rates (20 to 30 percent of typical
feed, by weight) than the material processed during the previous month. An excess of calcium in the
reverberatory furnaces causes build-ups inside the furnaces.
2.5.2.3 October
CHMR/Exide performed the final stage of the large-scale investigation in October. The material consisted
primarily of slag and debris, which was larger in size than previous material, and therefore needed to be
processed through the blast furnaces. This slag and debris material contained lead concentrations of
approximately 30 percent. The final 20 truckloads of the Pedricktown material were processed by early
November 1992.
43
-------
3.0 RESULTS AND DISCUSSION
In order to assess the viability and economics of the process, CHMR/Exide obtained production sheets,
samples, sample results and other general information of furnace performance to evaluate the processing
of the material. All statistical analyses were conducted according to the procedures outlined in Section 5
of the Quality Assurance Project Plan for the project. Analysis of variance (ANOVA) was performed on the
data sets. The ANOVA compares two experimental means to determine if a significant difference exists
between them. The null hypothesis is that no difference exists between the data set. The governing
equations are as follows:
Z=ts
and
where s is the standard deviation across all data, n is the number of samples analyzed, f is from statistical
tables (based on 90 percent probability level), and x is the calculated mean. When Z > D, the null
hypothesis is not rejected, and no significant difference between the means has been demonstrated. On
the other hand, when D > Z, then a significant difference does exist. The q-test, which rejects outlying
data, was also performed on each data set as required prior to the ANOVA.
3.1 TEST MATERIAL
The Pedricktown material was analyzed for percent moisture, silica, antimony, iron oxide, lead, sulfur, and
calcium. Table B-4 presents a summary of the analytical results of the material processed over the entire
investigation. The January material was not analyzed for all of these parameters, and the October slag
and debris material was only analyzed for total weight percent of lead.
3.2 SUMMARY OF PRELIMINARY INVESTIGATION RESULTS
CHMR/Exide conducted the preliminary investigation in which material was processed through the smelter
between January 29 and February 2,1992. During this preliminary investigation, CHMR/Exide processed
approximately 372 tons of material containing between 36 and 65 percent lead. The results from the
preliminary investigation showed that the material could be processed through the reverberatory furnaces
at feed rates of up to a maximum of 50 percent by weight test material. Higher feed rates caused build-ups
and break-downs in the furnace feed system due to the weight of the test material, which was heavier than
the typical Exide feed.
3.2.1 Lead and Slag Production and Quality
The amount and nature of the lead and slag produced by the reverberatory furnaces when processing the
test material were critical experimental parameters, since the lead is the salable product from the smelter,
and the reverberatory slag is a feed to the blast furnaces.
Blast furnace lead and slag production were not monitored during the preliminary experiment, because
the test feed material was fed exclusively to the reverberatory furnaces.
44
-------
TABLE
NL INDUSTRIES FEED MATERIAL - SUMMARY ANALYSIS
FULL-SCALE TEST
PARAMETER
% MOISTURE
% SiO2
% ANTIMONY
% FeO
% LEAD (AVG)
% LEAD (RANGE)
# OF FEED PILES
% SULFUR
JANUARY AUGUST SEPT
MATERIAL MATERIAL MATERIAL
N/A 10.1%
3.6%
N/A
N/A
56%
36 to 65
15
3%
5.0%
1.0%
3.1%
46%
42 to 52
3
1.3%
8.7%
N/A
2.3%
4.6%
52%
7 to 69
6
1.3%
OCT
MATERIAL
N/A
H/*
N/A
N/A
31%
6.5 to 31
1
N/A
3.2.2 Reverberatory Furnace Lead Production and Quality
The quality of the lead produced by the reverberatory furnaces during the January test was determined
through the use of Spectrometer Lab Test 100 analysis on-site, which measures the concentration of trace
metals in the lead product. CHMR/Exide representatives collected 20 lead samples from the test
reverberatory furnace, and 10 lead samples from the control reverberatory furnace. The average results
of these samples can be seen in Table B-5, which shows a comparison of the lead produced during
processing of the test material and the lead from the control reverberatory furnace processing typical feed.
the values given for the metals are relative units. The table shows that the quality of the lead produced
by the test reverberatory furnace is comparable with typical reverberatory lead. All of the experimental
means are statistically equivalent at a 90 percent probabi ity level except for that of antimony: 0.249+0.0028
units of antimony when test material was fed versus 0.041+0.008 when only regular feed was fed. This
was anticipated, because the test material was known to contain relativefy high concentrations of antimony.
Total lead production in the test and control reverberatory furnaces remained roughly equivalent during the
evaluation, with perhaps a slightly significant increase in production for the furnace fed test feed. The effect
of the material on total furnace production was more precisely quantified during the longer term evaluation.
3.2.3 Reverberatory Slag Production and Quality
CHMR obtained samples of the reverberatory furnace slag for both the test and control furnaces. The slag
was analyzed for total lead, antimony, silicates, sulfur, and iron. The results from the analyses are
presented in Table B-6. The results showed a significant decrease in lead concentration in the slag - from
an average of 73% in the control furnace to 48% in the test furnace. At the same time, the concentration
of silicates and iron increased in the slag from the furnace fed test feed.
Total slag production for the furnaces was only cursorily reviewed, since it was better quantified during the
longer term evaluation. It seemed to decrease slightly during the evaluation, although the decrease was
only barely statistically significant.
The combination of a decrease in lead concentration with a potential decrease in slag production indicates
a significant decrease in the amount of lead flowing from the reverberatory furnace to the blast furnace.
At the same time, lead production in the reverberatory furnace increased, Indicating that at least some of
the lead input was effectively shifted from the slag to lead product in the reverberatory furnace.
-------
TABLE B-5 REVERBERATORY LEAD FROM TEST MATERIAL VERSUS TYPICAL FEED LEAD
FOR JANUARY, 1992 PRELIMINARY INVESTIGATION
REVERBERATORY LEAD FROM TEST MATERIAL (TEST FURNACE)
AVQ LEAD FROM TEST MATERIAL (x):
STANDARD DEVIATION (S):
NUMBER OF SAMPLES (n):
TRUE MEAN (M):
Pb
13,812
483
20
13812
±
186
Sb
0.249
0.073
20
0.249
±
0.028
Sn
0.000
o.ooo
20
0
As
0.000
0.000
20
0
Cu
0.037
0.004
20
0.037
±
0.002
Ni
0.012
0.008
20
0.012
±
0.004
Bi
0.012
0.003
20
0.012
±
0.001
Ag
0.003
0.001
20
0.003
±
0.000
Fe
0.000
0.000
20
0
Te
0.004
0.001
20
0.004
+
0.000
Zn
0.000
0.000
20
0
TRUEMEANp-x±
St
V7T
VALUE FOR t BASED ON 90% PROBABILITY LEVEL
REVERBERATORY LEAD FROM TYPICAL FEED MATERIAL (CONTROL FURNACE)
AVQ LEAD FROM TYPICAL FEED (x):
STANDARD DEVIATION (S):
NUMBER OF SAMPLES (n):
TRUE MEAN (M):
Pb
13317
779
10
13317
±
446
Sb
0.041
0.014
10
0.041
±
0.008
Sn
0.000
0.000
10
0
As
0.000
0.000
10
0
Cu
0.040
0.006
10
0.040
±
0.004
NI
0.020
0.006
10
0.020
±
0.004
Bi
0.012
0.003
10
0.012
±
0.002
Ag
0.004
0.001
10
0.004
±
0.001
Fe
0.000
0.000
10
0
Te
0.004
0.000
10
O.OO4
±
0.000
Zn
0.000
0.000
10
0
TRUE MEAN p - x ±
St
77T
VALUE FOR t BASED ON 90% PROBABILITY LEVEL
-------
TABLE B-6
REVERBERATORY FURNACE SLAG COMPOSITION
DURING TEST WITH NL INDUSTRIES MATERIAL - - January 1992
Concentration
Lead Antimony Silicates Sulfur Fe
Date Sampled (percent) (percent) (percent) (percent) as % FeO
TEST FURNACE
29-Jan-92
29-Jan-92
29-Jan-92
29-Jan-92
29-Jan-92
30-Jan-92
30-Jan-92
30-Jan-92
30-Jan-92
31-Jan-92
31-Jan-92
31-Jan-92
31-Jan-92
01-Feb-92
01-Feb-92
01-Feb-92
01-Feb-92
02-Feb-92
02-Feb-92
02-Feb-92
03-Feb-92
AVERAGE:
STD. DEV.
CONTROL FURNACE
29-Jan-92 71.9% 3.9% 6.4% 2.60%
29-Jan-92 76.9% 3.7% 4.7% 2.20%
29-Jan-92 44.3% 3.6% 2.6% 2.70%
29-Jan-92 79.4% 3.8% 3.7% 2.70%
30-Jan-92 86.5% 3.9% 3.4% 2.90%
71.6% 2.7% 1.8% 5.2% 0.5%
31-Jan-92 78.9% 3.4% 4.1% 2.30%
31-Jan-92 79.7% 3.6% 4.2% 2.80%
68.0% 3.3% 3.8% 2.9% 0.8%
AVERAGE: 73.0% 3.5% 3.9% 4.0% 2.1%
STD. DEV. 11% 0% 1% 1% 1%
47
78.7%
58.3%
43.9%
49.7%
42.2%
49.7%
33.5%
54.0%
40.1%
43.7%
41.6%
47.1%
71.5%
47.8%
82.8%
57.4%
28.3%
30.2%
29.9%
34.3%
29.1%
20.2%
12.4%
25.9%
48.2%
17%
3.4%
4.2%
5.6%
6.8%
6.6%
5.4%
6.6%
4.4%
9.9%
11.4%
9.9%
7.3%
2.9%
6.0%
2.2%
2.0%
2.3%
3.2%
4.7%
3.8%
3.6%
3.2%
1.3%
4.0%
5.4%
3%
3.3%
5.8%
20.3%
15.3%
14.9%
11.3%
15.1%
9.5%
13.4%
9.3%
13.1%
12.1%
5.1%
14.5%
1.5%
3.8%
23.0%
21.1%
18.8%
17.6%
19.2%
22.8%
30.3%
21.5%
12.4%
7%
0.7%
4.4% 3.6%
10.6%
8.7%
2.2% 3.6%
10.4%
19.2%
12.9%
16.1%
17.3%
16.7%
12.0%
5.3%
11.5%
3.8%
0.3% 9.8%
19.4%
20.4%
21.3%
22.1%
25.4%
26.5%
26.7%
24.6%
2.3% 12.3%
2% 8%
-------
CO
z
13
HI
5
HI
to
z
o
CO
CO
LU
CM
O
CO
200
150
100 | i
50
70
HOURS FROM START OF TEST
_»_ JEST FURNACE — CONTROL FURNACE
ANOVA ANALYSIS OF EMISSION DATA
VARIABLE (EXPANATION)
x1,x2 (AVERAGE DATA VALUES)
s1,s2 (STANDARD DEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1, t2 (FOR 90% CONFIDENCE LIMIT)
s (STANDARD DEV. ACROSS ALL DATA)
Z = ts ((n1 + n2)/(n1 n2)) A .5
D = abs(x1 - x2)
Signif. difference iff D > Z:
Conclusion:
TEST
FURNACE
0.49
0.49
45
1.64
0.60
0.21
0.51
CONTROL
FURNACE
1.00
0.59
45
1.64
Significant difference
x2 - x1 = 0.51 ± 0.19 |
FIGURE B-1 Sulfur Dioxide Emissions from the Test and Control
Furnaces - January 1992
these analyses were not made available to CHMR, although reportedly the vast majority of slag samples
were non-hazardous. The amount of blast slag produced is a far more important parameter, and if is
discussed below.
3.3.2 Lead and Slag Production
Figures B-2 through B-7 present the lead and slag production for the three months of the large-scale
investigation. The data used to generate the figures was taken from Exide's daily productions sheets. All
of the values In the figures are normalized so that the average production of the control furnace is 1. This
was done to protect confidential Exide production information.
48
-------
ANALYSIS OF VARIANCE (ANOVA)
TEST
VARIABLE (EXPLANATION) FURNACE
x1 , x2 (AVERAGE DATA VALUES)
s1,s2 (STANDARDDEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1 , t2 (FOR 90% CONFIDENCE LIMIT)
(STANDARD DEV. ACROSS ALL DATA)
Z = ts ((n1 + n2)/(n1 n2)) A .5
D = abs(x1 - x2)
1.22
0.24
25
1.66
0.24
0.12
0.22
CONTROL
FURNACE
1 .00
0.18
20
1.71
-i
Signif. dfference iff D > Z:
Conclusion:
Significant difference
x2 - x1 = -0.22 ± 0.10
AUGUST REVERBERATORY SLAG PRODUCTION
3 2
£
n zo n tf-
j i •7 i •10 ii n ran n ™ n
DAYS PROCESSING*
-TEST MATERIAL REVERB SLAG CONTROL FURNACE REVERB SLAG
FIGURE B-2 Comparison of Reverberatory Slag Production- Test and Control Furnaces -
August 1992
Different figures are given for lead and slag production for each month of the investigation. Each figure
contains a statistical analysis of each month's production based on which furnace the test material was Fed.
For example, if the reverberatory furnace was being fed with test material, then the analysis of variance
was performed on the reverberatory slag and lead produced.
3.3.2.1 August
Figures B-2 and B-3 show that test material processed through the reverberatory furnaces in August
generated a significant statistically increase in lead and slag productions. This was an unanticipated result,
since the test material purportedly contained a lower concentration of lead than the regular feed. No
adequate explanation was found, except perhaps that the test feed material could be processed faster than
the regular feed, thereby increasing overall furnace throughput.
49
-------
ANALYSIS OF VARIANCE (ANOVA)
VARIABLE (EXPANATION)
TEST CONTROL
FURNACE FURNACE
x1, x2 (AVERAGE DATA VALUES) 1 .09 1 .00
s1 , s2 (STANDARD DEVIATIONS) 0.15 0.10
n1,n2 (NUMBER OF DATA) 25 21
t1 , t2 (FOR 90% CONFIDENCE LIMIT) 1 .66 1 .70
s (STANDARD DEV. ACROSS ALL DATA) 0.14
z = ts ((n1+n2)/(n1 n2))A .5
D= abs(x1 -x2)
Signif. difference iff D > Z:
Conclusion:
0.07
0.09
Significant difference
x2 - xl = -0.09 ± 0.06
AUGUST REVERBERATORY LEAD PRODUCTION
§,
s
5
•*•
ft 3*5
22 23 24
4 C • 7 • • 10 11 12 13 14
DAYS PROCESSING*
-TEST MATERIAL REVERB LEAD CONTROL FURNACE REVERB LEAD
FIGURE B-3 Comparison of Reverberatory Lead Production- Test and Control Furnaces -
August 1992
3.3.2.2 September
The statistical analyses for the reverberatory lead and slag production in September (figures B-4 and B-5)
ielded the same quantitative results as in Agiust, with statistically significant increases in the reverberatory
rn
throughput rates.
50
-------
ANALYSIS OF VARIANCE (ANOVA)
VARIABLE (EXPANATION)
xl, x2 (AVERAGE DATA VALUES)
s1, s2 (STANDARD DEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1,t2 (FOR 90% CONFIDENCE LIMIT)
s (STANDARD DEV. ACROSS ALL DATA)
Z = ts ((nl + n2)/(n1 n2)) A .5
D = abs(x1 - x2)
Signif. difference iff D > Z:
Conclusion:
TEST CONTROL
FURNACE FURNACE
1.24
0.34
32
1.64
0.31
0.14
0.24
Significant difference
x2 - xl = -0.24 ± 0.12
1 .00
0.23
26
1.65
SEPTEMBER REVERBERATORY SLAG PRODUCTION
a
Q
o
* 1
DAYS PROCESSING*
-TEST MATERIAL REVERB SLAG CONTROL FURNACE REVERB SLAG
FIGURE B-4 Comparison of Reverberatory Slag Procfuctton- Test and Control Furnaces -
September 1992
3.3.2.3 October
Figures B-6 and B-7, which show October blast furnace lead and slag production, indicate that the average
amount of lead produced in the furnace fed test feed was statistically higher than that produced than that
in the control. This difference was most probably due to a higher throughput caused by the use of the test
material. There was no significant difference demonstrated in the amount of slag generated in October
between the furnace fed test materials and the control furnace.
51
-------
ANALYSIS OF VARIANCE (ANOVA)
1 VARIABLE (EXPANATION)
xl, x2 (AVERAGE DATA VALUES)
31, s2 (STANDARD DEVIATIONS)
n1 , n2 (NUMBER OF DATA)
t1,t2 (FOR 90% CONFIDENCE LIMIT)
TEST
FURNACE
1.14
0.16
29
1.64
CONTROL
FURNACE
1.001
0.11
25
1 .66 !
s (STANDARD DEV. ACROSS ALL DATA) 0.15
Z= ts ((nl + n2)/(n1 n2)) ~.5 0.07
-jc2) _ _ i JL14 _ _____
Sign¥ differehcelff D"> Z: Significant difference
Conclusion: x2 - xl = -0.14 ± 0.06
SEPTEMBER REVERBERATORY LEAD PRODUCTION
ra
I
Q
EL
9"
DAYS PROCESSING*
-TEST MATERIAL REVERB LEAD CONTROL FURNACE REVERB LEAD
FIGURE B-5 Comparison of Reverberatory Lead Production- Test and Control Furnaces -
September 1992
3.3.3 Overall Blast Slag Production
Since blast slag is costly to dispose of, an increase in blast slag production would cause a commensurate
increase in costs. The amount of blast slag produced during each shift was determined based on Exide's
daily production reports and compared for the furnace systems fed test and regular feed. Table B-7
presents an ANOVA analysis for the blast slag production during the three month investigation. The
analysis showed that there was no significant difference demonstrated between the average test furnace
slag production and the average control furnace slag production. Therefore, it was concluded that the use
of the test feed would not cause a significant increase in disposal costs.
52
-------
ANALYSIS OF VARIANCE (ANOVA)
TEST
VARIABLE (EXPANATION) FURNACE
x1,x2 (AVERAGE DATA VALUES)
s1 , s2 (STANDARD DEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1,t2 (FOR 90% CONFIDENCE LIMIT)
s (STANDARD DEV. ACROSS ALL DATA)
Z = ts ((n1 +n2)/(n1 n2)) A .5
D = abs(x1 - x2)
Signif. difference iff D > Z: No
Conclusion: x =
0.95
0.15
16
1.75
0.13
0.08
0.05
difference
0.97 ± 0.04
CONTROL
FURNACE
1.001
0.10
16
1.75
demonstrated
OCTOBER BLAST FURNACE SLAG PRODUCTION
m
jj)
-*-
-4-
DAYS PROCESSING'
TEST MATERIAL BLAST SLAG CONTROL FURNACE BLAST SLAG
FIGURE B-6 Comparison of Reverberatory Slag Production- Test and Control Furnaces -
October 1992
3.3.4 Sulfur Dioxide Emissions
CHMR did not study the sulfur dioxide emissions during the large-scale investigation because the
preliminary investigation revealed that the emissions did not change significantly when test material was
processed through the furnaces and the emissions are well controlled by Exide's scrubber system
Therefore, CHMR assumed that the sulfur emissions would remain relatively constant, and did not obtain
or analyze data on them. Plant personnel indicated that no extraordinary exceedances of permit conditions
occurred during the three months that the tests were performed.
53
-------
ANALYSIS OF VARIANCE (ANOVA)
VARIABLE (EXPANATION)
x1,x2 (AVERAGE DATA VALUES)
s1, s2 (STANDARD DEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1, t2 (FOR 90% CONFIDENCE LIMIT)
s (STANDARD DEV. ACROSS ALL DATA)
Z= ts ((n1 + n2)/(n1n2))A.5
D = abs(x1 - x2)
Signif. difference iff D > Z:
Conclusion:
TEST CONTROL
FURNACE FURNACE
1.25
0.22
20
1.71
0.23
0.11
0.25
Significant difference
x2 - x1 = -0.25± 0.10
1.00
0.16
23
1.66
i
I
OCTOBER BLAST FURNACE LEAD PRODUCTION
I
o.
3"
i — i — t
t — 4-
DAYS PROCESSING*
TEST MATERIAL BLAST LEAD CONTROL FURNACE BLAST LEAD
FIGURE B-7 Comparison of Reverberatory Lead Production- Test and Control Furnaces-
October 1992
3.3.5 Calcium
Several loads of material at the site contained up to 50 percent calcium. Small quantities of such materials
could be blended with Exide's regular feedstock material. However, it is not possible for Exide to process
large quantities of such material through its blast or reverberatory furnaces, as such materials tend to
impair furnace production.
54
-------
TABLE B-7
ANOVA Analysis of Blast Slag Production - August -October 1992
VARIABLE (EXPANATION)
x1,x2 (AVERAGE DATA VALUES)
s1 , s2 (STANDARD DEVIATIONS)
n1, n2 (NUMBER OF DATA)
t1,t2 (FOR 90% CONFIDENCE LIMIT)
s (STANDARD DEV. ACROSS ALL DATA)
Z = ts ((n1 +n2)/(n1 n2))°-5
D = abs(x1 - x2)
Signif. difference iff D > Z:
Conclusion:
TEST CONTROL
FURNACE FURNACE
0.87
0.44
60
1.64
0.44
0.129
0.128
1.00
0.43
65
1.64
No difference demonstrated
x = 0.94 ± 0.06
ALL PRODUCTION VALUES HAVE BEEN NORMALIZED TO 1 BY DIVIDING THEM BY THE "NORMAL FEED"
TO PROTECT EXIDE CONFIDENTIAL INFORMATION
55
-------
APPENDIX C
THE USE OF SECONDARY SMELTING
TECHNOLOGY TO RECLAIM LEAD FROM IRON-SHOT
BRIDGE BLASTING MATERIAL
Description of Evaluation and Results and Discussion
November, 1993
56
-------
1.0 INTRODUCTION
The Pennsylvania Department of Transportation (PennDOT) uses iron shot abrasive material to blast old
paint from some state bridges. PennDOT collects the spent material and stores it in 55 gallon drums for
disposal. CHMR obtained 16 drums (approximately 6.5 tons) of the blasting material from PennDOT to
perform a qualitative evaluation to determine if the material could be processed through a secondary lead
smelter for lead recovery. The main goal of the investigation was to determine if the material could be
processed without causing any mechanical or production problems in the furnaces at Exide's secondary
lead smelter in Reading, PA. The material used was from a bridge in Belle Vernon, PA. CHMR observed
the modifications, if any, that were necessary to process the material in the most efficient manner.
This appendix details the evaluation conducted on iron shot abrasive material used by the Pennsylvania
Department of Transportation (PennDOT) to blast the paint from a bridge in Belle Vernon, PA.
57
-------
2.0 METHODOLOGY
The material arrived at the smelter in sixteen 55 gallon drums. The material was inspected by Exide
personnel upon its arrival at the site, and they determined that the material contained too much moisture,
which was anticipated to cause problems in the reverberatory furnaces. Plant personnel advised that the
material, including the barrel, be placed directly into Blast Furnace #2. The other blast furnace (#1) was
processing regular feed, and this was to be used as a control furnace.
Approximately 4 hours into the experiment, CHMR personnel at the site were informed that the control
furnace had been fed with a different type of "typical" feed material than the test furnace. This precluded
CHMR from making any direct comparisons between the two blast furnaces. This was later verified upon
analysis pf the lead produced by each blast furnace prior to processing any test material. Even before
the experiment, blast furnace #2 was producing lead that was significantly higher in antimony, arsenic, and
tin. CHMR decided to compare the production during the experiment with the production in the same
furnace before processing the test material.
A front-end loader was used to feed the barrels into the blast furnace at the rate of one barrel per charge,
with the furnaces being charged approximately twice every hour. All 16 of the barrels were fed into the
test furnace approximately 7 hours after the start of the experiment. Residence time in the furnaces
ranges up to 3 hours. Therefore, all test material was processed roughly 10 hours after feeding the first
barrel. Furnace operating parameters, such as back pressure and air and oxygen rates, were recorded
and blast furnace slag samples were collected by CHMR personnel approximately every hour.
The material obtained from PennDOT varied in size and overall composition. A composite sample of the
blasting material was obtained by removing some of the material from each barrel prior to processing
through the blast furnace. The material was dirty brown in color with visible blue-green specks in it. The
specks were most likely paint from the bridge. The sample also seemed to contain some dirt and small
wood pieces. Table C-1 provides the analysis for the composite sample of the material as received.
TABLE C-1
PennDOT BLASTING MATERIAL
Parameter
Iron
Calcium
Lead
TCLP Lead
Total Sulfides
Other
Composition
39.7 to 63.4%
1.8%
3.2%
2.5 mg/L
7.0 mg/kg
31.6 to 55.3%
These results show that the main constituent of the blasting material is iron. The lead concentration is
relatively low compared with reverberatory slag, which is typical blast furnace feed and which contains
between 60 and 70 percent lead. The blasting material is not TCLP hazardous for lead, as the test yielded
a value less than 5 mg/L.
58
-------
3.0 RESULTS AND DISCUSSION
CHMR obtained daily production sheets, sample results, and other general information about furnace
performance from Exide personnel and through direct observation in order to evaluate the processing of
the bridge blasting material through the blast furnaces at the smelter.
3.1
TEST MATERIAL
A composite sample of the 16 drums of the abrasive blasting material was collected prior to testing.
test material was low in lead (3 to 4 percent) compared to typical blast feed, which averages 65 to 70
percent lead. The main component in the material is iron, since it is an iron shpt abrasive blasting
material. The iron is assumed to be beneficial to the blast furnace performance since iron is usually added
to the furnace as a fluxing agent. The 55 gallon barrels that contained the blasting material also provided
iron in the furnace.
3.2
FURNACE PERFORMANCE
Discussions with plant management and furnace operators indicated that the test material did not cause
any back-ups or other mechanical problems in the blast furnace. The observed results were expected
since the amount of material being fed into the furnace was very small compared to the amount of typical
feed. Plant personnel also mentioned that the smelter should be able to accept other iron-shot based
abrasive material. However, they can not accept other types of blasting material that are silica based
(such as "Black Beauty") because a high silica content has the potential to cause problems in the furnaces.
3.3
LEAD AND SLAG QUALITY AND PRODUCTION
The amount and nature of the slag and lead produced by the furnace are critical experimental parameters.
The quality of the lead is important since the lead is the marketable product from the smelter. The blast
furnace slag is one of the waste streams from the smelter, and disposal is costly.
3.3.1 Lead Quality
The results of the lead analyses are presented in Table C-2. The analyses were performed on-site by
Exide personnel using Spectrometer Lab Test 201 equipment. This equipment determines the amounts
of trace metals in the lead sample. The "Lead Before Experiment" are samples taken from each blast
furnace prior to processing the test material. These initial samples show that the lead produced in blast
furnace #1 contains significantly different amounts of antimony, tin, arsenic, copper, and cadmium than
the lead produced in blast furnace #2. This was due to the different feed piles that were being used for
each furnace, according to Exide personnel. The difference in the feeds is what led CHMR to abandon
furnace #1 as a valid control furnace.
TABLE C-2
TEST MATERIAL BLAST LEAD ANALYSIS
LEAD BEFORE EXPERIMENT:
BLAST # 1 03:05 PM
BLAST #2, 03:05 PM
TEST MATERIAL BLAST LEAD:
BUST #2,04:00 PM
BLAST #2,05:35 PM
BLAST X2. 07:00 PM
MEAN TEST BLAST LEAD:
Pb^
9.636
8,568
8,566
10.236
11.243
... — i
1.5001
Sb Sn
1.95 0.43
5.08 1.14
5.08 1.17
5.15 1.08
4.62 1.00
4.95 1.08
± ±
0.32 1 0.09 1
L A»
0.08
0.47
0.44
0.42
0.25
0.37
07121
Cu
0.45
0.33
0.31
0.28
0.28
0.29
•4-
OX)2I
Ni
0.04
0.04
0.04
0.03
0.02
0.03
o!bi
Cd
0.15
0.57
J.
0.53
0.46 1
0.35
0.44 i
0~10I
Ca
0.11
0.11
0.01
0.10
0.10
0.07
•4-
OX«I
So
001
0.00'
0.01
0.01
0.01
0.01
oJbo
s
0.03
0.05
0.04
0.03
0.03
0.03
4-
OXM
Al
o.or
0.01
0.00
0.02
0.02
OA A
.01
± '
0.01>
59
-------
The section in Table C-2 labeled "Test Material Blast Lead" refers to the samples taken from the test
furnace during the processing of the abrasive blasting material. The "Lead Before Experiment" are.initial
samples taken prior to testing. The table shows no significant difference, based on a 90 percent
confidence limit, between the metals concentrations taken before and during the experiment.
3.3.2 Slag Quality
Table C-3 provides the analyses of the slag from the test blast furnaces. The results from blast furnace
#1 were not used for the evaluation. CHMR decided to concentrate on the slag taken from blast #2 the
test furnace, to see how the composition of the slag changed over the course of the evaluation. An initial
slag sample was taken from the test furnace prior to testing, followed by five additional samples at various
times during the processing of the test material. These samples were composites of two to three furnace
taps.
TABLE C-3
ANALYSIS OF SLAG SAMPLES
Sample
Initial
1
2
3
4
5
% Pb
3.75
5.10
3.90
2.45
3.55
6.25
%Fe
38.7
44.7
40.8
42.5
46.9
47.4
Reactive Sulfides !
1.12 ppm i
ND* '
ND* i
0.23 ppm ,
0.24 ppm !
0.25 ppm
The data indicate that the test material had some small effects on the composition of the slag. The
percent iron in the slag showed a slight increase, which may be due to the high iron content of the test
material. The reactive sulfides present in the slag were slightly lower during the processing of the test
material.
3.3.3 Slag and Lead Production
It is difficult to make a determination about the production of lead and slag since the material was only
being fed for a 7 hour period. CHMR obtained Exide's daily production sheets from 9 days before the
evaluation through the day after. The production sheets indicated that there was basically no affect on
the amount of lead and slag produced during the test when compared to production data from before the
test. CHMR concludes this is because of the relatively low feed ratio (10 percent by weight) of the test
material to the furnace.
3.4
SULFUR DIOXIDE EMISSIONS
CHMR did not obtain sulfur dioxide emissions data for this evaluation. The iron-shot abrasive blasting
material was initially low in sulfides (7 ppm) and was not expected to raise the sulfur dioxide emissions.
Previpus evaluations conducted with materials containing 2 to 3 percent sulfur showed no effect on
emissions, which are well controlled.
3.5 FURNACE PARAMETERS
Certain furnace parameters, such as oxygen and air usage, and furnace back pressure, were monitored
during the evaluation, and the results are given in Table C-4. The amount of oxygen needed in the test
blast furnace was increased slightly (approximately 100 ft3 per hour) after roughly 3 hours because of the
60
-------
test material. Air usage did not fluctuate at all, and back pressure seemed to behave normally. There
were no reports of any build-ups or other complications in the furnace that could be attributed to the test
material.
TABLE C-4
TEST BLAST FURNACE OPERATING PARAMETERS
TIME
BLAST AIR BLAST O2
(ft3/min)
12:30PM 1000
01:00 PM 1000
01:45 PM, 1000
02:40 PM 1000
03:00 PM
03:40 PM
04:30 PM
05:00 PM
06:00 PM
07:00 PM
1000
1000
1000
1000
1000
1000
BACK
(fP/hr) . PRESS (psi).
4500
4500
4500
4500
4600
4600
4700
4700
4600
4600
1.88
1.88
COMMENTS
START FEEDING AT 12:45 PM
FEEDING 1 BARREL PER CHARGE
1 .88 ONE CHARGE PER HALF HOUR
1.75
1.81
1.94
1.63
1.81
1.69
2.06
LAST BARREL AT 07:00 PM
61
-------
4.0 CONCLUSIONS
Based on the study, CHMR concluded the following:
The test material was fed to the furnace at a feed ratio of approximately 10 percent of the typical
feed that is processed through the smelter. No effect was found on lead and slag production and
quality, production costs, sulfur dioxide emissions, and furnace performance.
The material can be easily fed into the blast furnace. The blasting material arrived in 55 gallon
drums, and were charged directly into the furnace when processing this material.
Exide plant management believe that they can accept iron shot abrasive blasting material. They
cannot accept silica-based blasting material.
Costs of processing the material in a secondary lead smelter are estimated to be below $150 per
drum.
These conclusions were based on one relatively short term test, with a relatively small amount of material
processed. CHMR recommends additional testing and further monitoring of daily production data to
determine if the blasting material has adverse effects on the slag and lead production and to better
determine the process economics.
62
-------
APPENDIX D
RECLAMATION OF LEAD FROM BATTERY CASE MATERIALS FROM THE HEBELKA SITE
Description of Evaluation and Results and Discussion
February 5, 1992
63
-------
1.0 INTRODUCTION
The Center for Hazardous Materials Research (CHMR) and Exide/General Battery Corporation (Exide)
engaged in a joint research study to determine the feasibility and economics of using Exide's
secondary lead smelters for the recovery of lead from lead-containing materials from Super-fund sites
and other facilities. As part of that project, CHMR/Exide processed approximately 8 tons of battery
case materials from the Hebelka Corporation (Hebelka) site in Upper Macungie Township,
Pennsylvania.
The primary purpose of the evaluation was a qualitative determination of whether or not battery case
materials have negative affects on the reverberatory furnaces when the size is reduced significantly. A
previous investigation showed that large pieces of battery cases slow the reverberatory furnace
production. The Hebelka Superfund site was originally an auto salvage yard containing approximately
1,000 cubic yards of rubber battery case material from an unknown battery breaker source. The
materials consisted of large broken pieces of rubber and polypropylene battery cases.
CHMR/Exide received permission in January 1992 from Patrick Augustin, the EPA Project Officer for
the research, and Fred McMillan, the Hebelka Remedial Project Manager from EPA Region III, to
proceed with the evaluation. The material was processed on February 5, 1992.
The main conclusion from the evaluation was that the battery case materials, after being milled to less
than 1/4 inch, could be successfully fed to the reverberatory furnaces at a weight ratio of at least 17%.
64
-------
2.0 DESCRIPTION OF EVALUATION
The purpose of the evaluation was to determine if reducing the size of battery case materials affects
reverberatory furnace performance. The Hebelka Super-fund site in Upper Macungie Township,
Pennsylvania, was used as the source of these materials. The materials were processed over a five
hour period at Exide's Reading, Pennsylvania, secondary lead smelting facility.
2.1 RECLAMATION PROCESS
The reclamation of lead from Superfund and other lead-containing materials is based on existing lead
smelting procedures. The material are first collected or excavated. Next, they may be preprocessed
to reduce particle size, and to remove rocks, soil, and other debris. The materials are then transported
to the smelter. An overview of the reclamation and smelting process can be found in the main report.
2.2 EXPERIMENTAL DESIGN
The main objective of the evaluation was to feed Hebelka material to the furnaces at a known feed
ratio and determine if the material significantly changed furnace operation or parameters. This was a
short term test lasting only five hours, but the intention was to see if reducing the size of the battery
case material would allow it to be processed in the reverberatory furnaces. During the Tonolli
experiment, one of the conclusions was that large battery case pieces (up to 14") cause build-ups in
the reverberatory furnaces. This is because the larger pieces burn more slpwly. Smelter management
determined that a relatively short term test would be sufficient to detect buildups within the
reverberatory furnace.
The material from the Hebelka site was reduced in size in a hammermill until it passed through a V*
inch screen. After size reduction, the material was fed to one reverberatory furnace. No test material
was fed to the blast furnaces, as the ability of blast furnaces to reclaim lead from battery case material
was evaluated in a previous study (Tonolli). The second reverberatory furnace was fed typical
material. Conducting the test in this manner allows for comparison of the operation of the "test"
furnace with the "control" furnace.
During the evaluation, the operating parameters (fuel usage, oxygen usage, slag production, lead
production, etc.) for each set of furnaces were obtained, m addition, samples 9? slag and lead were
obtained from each furnace to be used to determine effects on production quality.
2.3 EXPERIMENTAL MATERIAL
The Hebelka material processed during this evaluation consisted primarily of chunks of rubber battery
cases, ranging in size from a fraction of an inch to full-sized (approximately 12-inch by 8-inch x 6-inch)
battery cases. The cases were black in color, and appeared to be a mixture of ebonite rubber with
some polypropylene cases. Much of the material was covered with a thick reddish-purple sludge,
which contained approximately 80% of the lead. Tests of the material showed average concentrations
of 14.7% lead. The composition of the material is further characterized in Table D-1.
2.4 CHRONOLOGY OF THE EVALUATION
The following represents a chronology of the activities conducted in support of the evaluation between
January 24 and February 5, 1992.
2.4.1 Material Acquisition (1/24/92)
Eight tons of the battery case material were excavated and loaded in two trucks by an EPA contractor.
After the excavation and loading of the material, the equipment and trucks were decontaminated under
EPA supervision. The material was transported to Exide's smelting facility weighed, and then stored
in an area isolated from other feed material.
65
-------
2.4.2 Size Reduction (2/5/92)
The waste battery case material varied in size between a fraction of an inch and large pieces up to 14
inches. Based on a previous experiment with rubber battery case material from the Tonolli Super-fund
site, it was determined that rubber battery case material would need to be reduced in size prior to
processing the material.
The material leaving the hammermill ranged in size from a fine dust to a particle size around % inch.
The material was collected in cubic yard boxes and mixed on-site with regular feed material before
processing. The material in each box weighed approximately 800 pounds.
2.4.3 Processing in the Furnace (2/5/92)
The regular reverberatory furnace feed consisted of spent battery parts, lead oxides, sludge and other
lead containing material, mainly from Exide's sink/float separation system. The rubber battery cases
were mixed with regular feed material at a 17% weight ratio (e.g., 17 pounds of material to 83 pounds
of regular feed). The mixed material was fed into reverberatory furnace #1. Reverberatory furnace #2
served as the experimental control. The material was processed between 12:20 PM and 5:20 PM on
February 5, 1992.
During the experiment, reverberatory furnace #1 was tapped twice for slag. The slag from the furnace
flowed well and there was no noticeable build-up within the furnace. CHIVTR personnel continued to
monitor the furnace performance until 7:30 PM, with no problems noted in the furnace.
66
-------
3.0 RESULTS AND DISCUSSION
In order to assess the viability of the process, CHMR/Exide recorded the settings of various plant
parameters, the production figures for both reverberatory furnaces, and stack parameters. In addition,
CHMR obtained samples of the input material, and output slag and lead. The sampling was conducted
in accordance with the EPA approved Quality Assurance Project Plan (QAPP). The results from the
sampling and analysis are presented and discussed below.
3.1 TEST MATERIALS
The test material was analyzed for lead, silica, sulfur, carbon, and iron concentrations, as well as bulk
density, and heat value. Table 1 shows the results of analytical tests on the test materials. The
results for silica and sulfur are below the thresholds at which the material would impair the furnaces.
The carbon and iron percentages are favorable for smelting. The amount of lead in the material
(14.7%) is considerably below that of normal reverberatory furnace feeds, which average about 60%
lead.
Table D-1. Hebelka Battery Case Material Analysis
Sample
#
2
3
Average
Pb
(%)
13.9
18.4
11.7
14.7
S
(%)
3.3
1.3
2.3
Si02 (%)
5.2
4.2
4.7
C
(%)
52
60
56
Fe
(%)
0.45
0.36
0.41
Density
(Ib/ft3)
58
51
55
Heat
(BTU/lb)
9670
7820
8750
3.2 SLAG AND LEAD PRODUCTION AND QUALITY
The amount and nature of slag and lead produced by the furnaces were critical experimental
parameters. The amount and quality of lead produced is important because lead is the salable product
from the smelter. The reverberatory slag characteristics also important, because this slag is used as
feed material for the blast furnace.
Composite slag samples were taken as the slag was being tapped from the furnaces throughout each
shift from reverberatory furnace #2. Grab samples of sjag were taken from furnace #1. The molten
slag samples were collected by the Exide operators by inserting a shovel into the furnaces and
extracting the sample, in accordance with the QAPP. The results of the slag analysis are presented in
Table D-2.
The amount of lead production was determined by collecting Exide's daily production reports, which
detail the entire smelter activity for each day. The quality of the lead was determined through analysis
with a Spectrometer Lab Test 100. This analyses, which is performed on-site by Exide personnel,
provides the amount of several trace metals in lead samples taken from the furnace. The results from
these tests can be seen in Table D-3.
Performing a paired two-sample T-test for means for each metal in the table revealed that only the
antimony showed a difference between the two means. The test was performed with a 95%
confidence level.
67
-------
Table D-2. Composition of Slag Samples from Reverberatory Furnaces
Sample
1
2
3
4
5
6
!f? •*• ^'v ' %
«; *• „ "•
' Sample
7
8
Furnace
Test
Test
Test
Test
Pb (%)
72.0
42.2
62.4
73.1
Test 59.1
Test
^tti&vtttti*'*
f\%*f**f!J*r
Furnace
Control
Control
'&vefag.
Pb (%)
S (%)
8.2
1.3
7.1
10.5
7.3
6.9
'', && ^
;//, >">»* JJj
S (%)
52.5 6.5
60.8
6.1
Sb (%)
3.0
5.8
2.8
1.5
2.7
3.0
*4 -
-------
Table D-4. Furnace Parameters During Test.
Time
13:30
14:00
14:30
15:00
15:30
16:00
16:30
17:00
17:30
Oxygen usage (cfm)
Test
1500
1500
1500
2100
2200
2200
2200
2000
2100
Control
1500
1900
1700
1800
1800
1800
1900
1800
1800
Gas usage (rel. units)
Test
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.8
Control
8.0
6.0
8.0
6.0
6.0
6.0
6.0
6.0
6.0
3.4.1 Sulfur Dioxide Emissions
In order to feasibly process the Hebelka material, it is important that the test materials do not cause an
increase in sulfur dioxide (S02) emissions above Exide's permitted emissions limits of approximately 110 parts
per million (ppm). No such increase was anticipated because the concentration of sulfur in the normal furnace
feed is about 1 to 2%, and also because Exide has a sulfur dioxide control system. However, CHMR/Exide still
monitored sulfur dioxide emissions to determine if an increase occurred.
CHMR/Exide obtained data concerning sulfur dioxide emissions from the facility's SO? Continuous Emission
Monitor (CEM). Average and initial data values were collected by the operator and CHMR/Exide project
personnel on duty during each shift. Each stack system is connected to the corresponding blast and
reverberatory furnaces. The average stack emission in system #1, to which the test material was fed, was 56
ppm, while the average stack emission in system #2 was 78 ppm. Thus, the emissions remained considerably
below the permit levels of 110 ppm.
3.4.2 Calcium Sulfate Sludge
During each shift, specific gravity measurements of slurry generated from the air emission control equipment
(scrubbers) were taken by the operator on duty. The specific gravity is related to the calcium sulfate
concentration in the slurry, and therefore gives a measure of total sulfur uptake in the sludge. Slurry specific
gravity measurements for both stacks averaged 1.25, which indicates no significant difference between the
slurry produced in the furnace system fed the Hebelka material and that produced in the system fed only
normal feeds.
At the end of each shift, the arrount of slurry was calculated by counting the number of truckloads transported
from the scrubbers to the on-site sludge dewatering facility. No increase was found in the number of
truckloads produced each shift.
69
-------
APPENDIX E
RECLAMATION OF LEAD FROM DEMOLITION MATERIAL FROM A HUD
RENOVATION
Description of Evaluation and Results and Discussion
December 1991
70
-------
1.0 INTRODUCTION
The Center for Hazardous Materials Research (CHMR) and Exide/General Battery Corporation (Exide)
engaged in a joint research study to determine the feasibility and economics of using Exide's
secondary lead smelters for the recovery of lead from lead-containing materials found at Superfund
sites and other sources. As part of this study, CHMR/Exide processed demolition material (primarily
wood) from the renovation of a women's shelter in Montgomery County, PA.
One common, but often overlooked, source of lead is demolition material from residential housing
coated with lead based paints. Many houses, especially older ones, have interior and exterior wood
which can contain lead in excess of 1%. The current treatment for lead-containing demolition waste is
to place it into sanitary landfills with other municipal wastes. As an alternative to landfilling, CHMR and
Exide evaluated the potential for processing lead-containing demolition material in a secondary lead
smelter.
The demolition material is from the front and back porches of the Laurel House, a women's shelter in
Montgomery County, PA. The deteriorating porches were coated with lead based paint. The
Montgomery County Housing and Community Development (MCHCD) Office contracted an
environmental firm to perform the necessary abatement. Nearly five tons of lead-containing waste was
generated.
The main goal of this short term experiment was to determine the effects that a wooden feed would
have on the reverberatory furnaces. After receiving consent from the EPA, the demolition material was
transported to Exide's secondary lead smelter in Reading, PA, and accumulated until December 16,
1991, when the material was processed through the smelter.
71
-------
2.0 DESCRIPTION OF EVALUATION
The purpose of this evaluation was to determine the affects on furnace production that a wood-based
feed would have. Material was acquired from the Laurel House, a woman's shelter in Montgomery
County, PA. The demolition materials were processed over a ten hour period at Exide's Reading, PA
smelter.
2.1 RECLAMATION PROCESS
The reclamation of lead from Superfund and other lead-containing materials is based on existing lead
smelting procedures. The material are first collected or excavated. Next, they may be preprocessed
to reduce particle size, or to remove rocks, soil, and other debris, The materials are then transported
to the smelter. An overview of the reclamation and smelting process can be found in the main report.
2.2 EXPERIMENTAL DESIGN
The main objective of this evaluation was to process the demolition material in the reverberatory
furnaces at a known feed ratio and determine if the material significantly changed furnace operation or
performance. Unlike other evaluations, material was charged to both reverberatory furnaces.
Therefore, this test did not have a "test" and "control" furnace and so comparisons can only be made
to determine how both furnaces performed with respect to each other while processing the same feed.
The preprocessing phase for this evaluation involved reducing the size of the demolition material with a
pallet shredder, and then storing in an area separate from other feed material at the smelter. During
this evaluation, certain operating parameters for each reverberatory furnace were obtained, including
air and natural gas usage. Samples of slag and lead were obtained from the reverberatory furnaces
for analysis.
2.3 EXPERIMENTAL MATERIAL
The demolition material consisted primarily of pieces of wood coated with lead-based paint. The
material arrived at the smelter on pallets in three to six foot lengths. The material had an average lead
concentration of 1.2%, and an average TCLP lead level of 13.7 mg/l. The ash content of the material
was determined to be 4% by weight.
2.4 CHRONOLOGY OF THE EVALUATION
The following is a chronology of the activities conducted for this evaluation between November 21 and
December 16, 1991.
2.4.1 Material Acquisition
MCHCD contacted CHMR in early November and expressed interest in working with the lead
reclamation project. CHMR consulted with Patrick Augustin of the EPA, who agreed that the material
could be processed as part of the research project. Starting on November 21, the environmental
contractor responsible for the lead abatement transported several pallets of the demolition material to
the smelter in Reading, PA. Once at the smelter, the material was weighed, processed through a
72
-------
pallet shredder, and stored separate form other smelter feed material.
2.4.2 Day One (12/15/91)
CHMR personnel arrived at the site and were introduced to smelter management and supervisors.
CHMR personnel also received health and safety training regarding the operations and equipment at
the smelter.
The density of the demolition material was determined by weighing a known volume of the material
This procedure was duplicated seven times, and the average density was found to be 15.7 Ibs/ft3.
Three composite sample of the initial demolition material were collected in accordance with section 3.6
of the Quality Assurance Project Plan (QAPP). Grab samples of the lead and slag produced by the
reverberatory furnaces prior to processing any test material were also collected.
2.4.3 Day Two (12/16/91)
More samples of the reverberatory lead and slag were taken before feeding the demolition material
The test material and regular feed were mixed prior to starting the experiment.
At 10:30 AM, a mixture containing 10% by weight (50% by volume) demolition material was charged to
both furnaces. At approximately 1:30 PM, plant management noticed a buildup of material in both of
the reverberatory furnaces. Test feed to the furnaces was suspended and oxygen flow was increased
to try and correct this problem.
Exide plant management decided that the feed ratio should be lowered to 5% by weight to try and
prevent any further buildups from occurring in the furnaces. The mixed feed was started again at 3:00
PM and continued until 8:45 PM, with no other problems noticed in the furnace. All slag and lead
samples taken during the course of the investigation were taken in accordance with the QAPP.
73
-------
3.0 RESULTS AND DISCUSSION
CHMR obtained enough data to determine whether or not the material could be successfully run at the
tested feed ratios. The test was not long enough to determine the total effects that processing the
demolition material had on the furnaces. The following sections give more detailed information on the
operating parameters for the two reverberatory furnaces.
3.1 INPUT PARAMETERS
CHMR personnel recorded air flows, oxygen volumes, natural gas usage, and types of material
charged into the furnace. No major changes were noted in air or natural gas usage throughout the
length of the experiment. The percent oxygen enrichment had to be increased for a short time to
reduce the buildups in the reverberatory furnaces that occurred from processing the test material at a
10% by weight mixture. Overall the results were inconclusive: insufficient material was processed to
adequately determine if changes in process settings, i.e., natural gas, oxygen, were necessary to
process the demolition material.
Three samples of the test material were collected prior to the start of the evaluation, and two were
sent to separate laboratories. Table E-1 shows the composition of the feed material. The material has
a lead concentration of 0.96% in one analysis and 1.5% in another. The average leachable lead is
13.7 mg/l, based on TCLP analysis.
Table E-1. Analysis of Demolition Material
Feed
Sample
1
2
TCLP Pb
(mg/l)
11.3
16.1
Pb
(ppm)
9,594
—
Pb
(%)
0.96
1.5
Ash
(%)
4
3.2 SLAG PRODUCTION AND QUALITY
The reverberatory slag production and quality are important concerns because the slag is used as feed
to the blast furnaces. CHMR acquired Exide's daily production sheets for the week of December 13
through December 18, 1991. The test appeared not to affect the slag production at all, however, since
the test was short term with only 9,000 pounds processed, no conclusions can be drawn about how
the demolition material affected the slag production.
Composite slag samples were taken as the slag was tapped throughout each shift. These samples
were taken by inserting a shovel into the furnace and removing the slag, in accordance with section
3.6 of the QAPP. At the end of each shift, CHMR personnel collected the samples and split them with
Exide. The reverberatory slag samples were analyzed for percent lead, tin, antimony, and arsenic.
Tables E-2 and E-3 below show the analysis of the slag samples taken before, during, and after the
evaluation. Although furnace #1 showed a slight decrease in slag lead content, Exide personnel said
that it is in the acceptable range for the reverberatory slag. Furnace #2 showed a slight increase, but
leveled out during the evaluation.
74
-------
Table E-2. Reverberatory Furnace #1 Slag Samples.
When
Before
During
During
After
Pb (%)
60.4
55.3
50.0
47.2
Sb (%)
3.40
5.07
1.25
6.56
Sn (%)
0.80
1.09
0.25
1.50
As (%)
0.22
0.29
4.25
0.36
Table E-3. Reverberatory Furnace #2 Slag Samples.
When
Before
During
During
After
Pb (%)
51.4
66.5
69.3
66.2
Sb (%)
5.34
3.60
0.56
3.28
Sn (%)
1.52
0.66
0.06
0.62
As (%)
0.43
0.24
2.80
0.22
3.3 LEAD PRODUCTION AND QUALITY
The production and quality of lead is an important parameter because lead is the salable product from
the smelting process. CHMR anticipated that the demolition material would cause a decrease in the
production of lead. This would not be caused by any change in the slag, but because the demolition
material comprised much of the volume of the feed. Therefore, any decrease would be due to the
decreased throughput. However, the material was not run long enough to notice any affects on the
lead production.
Lead samples were collected throughout the evaluation and analyzed with a Spectrometer Lab Test
100. This analyses, which is performed on-site by Exide personnel, provides the amount of several
trace metals in lead samples taken from the furnace. CHMR could not get this data, however Exide
plant supervisory personnel noted that there was no change in lead quality throughout the evaluation
3.4 SULFUR
Once material is inside a furnace, the sulfur present in the demolition material can exit the process by
three means: through stack gas emissions, through the calcium sulfate sludge generated by the
emissions control system, or in the blast and reverberatory slag.
3.4.1 Sulfur Dioxide Emissions
Sulfur dioxide (S02) emissions cannot exceed Exide's permitted emissions limits of 110 ppm. S02
emissions at Exide are constantly monitored and recorded on a strip chart recorder. However, at the
time of the evaluation, Exide was having mechanical problems with their S02 Constant Emission
Monitor (CEM). Plant personnel estimated that the sulfur dioxide levels never exceeded 100 ppm
based on the limited data that they could acquire.
75
-------
3.4.2 Calcium Sulfate Sludge
Calcium sulfate sludge is generated by the emissions control equipment (scrubbers) at the smelter.
Specific gravity measurements of the slurry were taken approximately every hour by Exide personnel.
The specific gravity is related to the amount of calcium sulfate in the slurry and therefore allows for an
estimate of sulfur uptake in the sludge. The average specific gravities over the course of the
experiment were 1.22 and 1.20 for furnaces 1 and 2, respectively. There was also no increase in the
daily amount of calcium sulfate sludge generated.
76
-------
4.0 CONCLUSIONS
Based on the limited evaluation conducted, CHMR/Exide concluded the following:
It is technically feasible to reclaim lead from wood demolition materials, provided they are
appropriately sized to fit in the furnaces.
The material burns well in the furnaces, and did not present any major difficulties based on its
combustive properties (i.e., did not cause a build-up within the reverberatory furnaces).
The economics of recovering lead from such materials are difficult to assess. However, the
value of the material as a feed to the smelter is minimal because it contains such a low
concentration of lead. The use of secondary lead smelters is only likely to be economical
when the demolition material is a hazardous waste, and therefore cannot be disposed
economically in a municipal or demolition landfill.
77
------- |