v>EPA
                                United States
                                Environmental Protection
                                Agency
                                Industrial Environmental Research
                                Laboratory
                                Research Triangle Park NC 2771 1
                                Research and Development
                                EPA-600/S2-81-038  July 1981
Project  Summary
                                Level  1   Environmental
                                Assessment of  Electric
                                Submerged-Arc  Furnaces
                                Producing  Ferroalloys

                                C. W. Westbrook and D P Daugherty
                                  An EPA/IERL-RTP Level 1 multi-
                                media environmental assessment of
                                the ferroalloy industry was conducted.
                                The report contains general industry
                                statistics and the results of sampling
                                and analysis at three plants (six furnaces
                                total).
                                  The industry is facing severe pressure
                                from imported products and its con-
                                tinued viability is uncertain. In addition,
                                this report indicates that the potential
                                for serious environmental problems
                                exists within some segments of the
                                industry but does not prove that the
                                pollution problems are occurring.
                                Specifically, the pollution potential of
                                covered (mix-sealed and sealed) fur-
                                naces is substantially higher than for
                                open type furnaces, primarily due to
                                the high concentration of organics in
                                gases generated by covered furnaces.
                                The covered furnaces are estimated to
                                generate polycyclic organic material
                                (POM) at the rate of about 1,230 to
                                11,080 kg/yr (2,710 to 24,430 Ib/yr)
                                per megawatt of furnace capacity or
                                208.800 to 1,878,800 kg/yr (460,300
                                to 4,120,000 Ib/yr) for all U.S. furnaces
                                of this type. Open furnace POM gen-
                                eration rate is estimated to be 100 to
                                900 kg/yr (220 to 1,980 Ib/yr) per
                                megawatt of furnace capacity or
                                134,500 to 1,210,500 kg/yr (296,500
                                to 2,668,700 Ib/yr) for all U.S. fur-
                                naces of this type. Covered furnaces
                                comprise only 14 percent of the indus-
                                try's production capacity and no
                                growth in their use is expected. These
                                estimated nationwide POM genera-
                                tion rates (estimated rates before the
                                emission control devices) are in the
                                same order of magnitude as estimated
                                POM generation rates (before control
                                devices) of slot type  coke ovens,
                                which EPA considers to be a major
                                emitter. However, the control devices,
                                which are in use on all U.S. ferroalloy
                                furnaces, remove most of this material
                                from the gas stream. Samples from
                                one mix-sealed furnace were analyzed
                                by GC/MS which gave positive identi-
                                fication of known organic carcinogens
                                in both the clean gas discharged by the
                                scrubber (but before passing through
                                the flare which is expected to destroy
                                some organics) and in the water dis-
                                charged by the scrubber (which is
                                treated before discharge from the
                                plant). Low resolution mass spectro-
                                graphic  (LRMS) analysis indicates the
                                presence of carcinogens in the cleaned
                                scrubber discharged gas (before flar-
                                ing) or four of the five scrubber equipped
                                furnaces tested, and the  water dis-
                                charged from all  scrubbers tested
                                (before wastewater treatment), and in
                                the gases generated by one open
                                furnace  served by a baghouse (emis-
                                sions from the baghouse were not
                                determined). LRMS indicated the
                                presence of carcinogens in the waste-
                                water discharged by only one (no
                                longer operating) of the three plants
                                tested.

-------
  The report indicates areas in which
further study and/or emissions.quan-
tification is needed.
  This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory. Research Tri-
angle Park, NC,  to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see  Project Report
ordering information at back).

Introduction
  Ferroalloys are alloying elements
which when added to molten iron or
steel  impart  specific  characteristics,
such as hardness, ductility, and corrosion
resistance, to the finished product. The
United States is one  of the world's
largest producers and  consumers of
ferroalloys. Annual U.S. production is
about 1.45 million tonnes and con-
sumption is about 2.1 million tonnes.
  Ferroalloys are manufactured primarily
in submerged arc  electric furnaces.
Other production and refining methods
are vacuum and induction furnaces,
exothermic (alummo-silico-thermic)
processes and electrolytic manufacture
of high purity metals.
  The submerged arc furnace, Figure 1,
consists of a refractory lined crucible
                                  with a tap hole near the hearth level to
                                  withdraw the molten product Power is
                                  supplied to the furnace through carbon
                                  electrodes which extend downward
                                  through the charge material to a point
                                  slightly above the hearth. Charge mate-
                                  rials, which include ores, scrap iron,
                                  gravel, coal, coke, and sometimes
                                  woodchips, are fed  to the furnace as
                                  required to keep the crucible filled. The
                                  electric current passing into the furnace
                                  raises the temperature of the charge
                                  into the range that  the  reduction re-
                                  actions  (basically removal of oxygen
                                  from the  metals) can occur.  Large
                                  volumes of carbon monoxide gas are
                                  produced  in  the  reduction reactions.
                                  Furnace power consumption rates range
                                  from  about 7 megawatts to over 50
                                  megawatts depending on furnace size
                                  and product being made.
                                    Furnaces are categorized by the type
                                  of furnace top cover used  There are two
                                  basic categories (open and covered) and
                                  two subtypes for each basic category
                                  The open category is composed of totally
                                  open furnaces in which there is an open
                                  gap of one meter or  more between the
                                  crucible top and the fume collecting
                                  hood, and  close hooded  in which this
                                  gap is significa ntly reduced by removable
                                  doors or panels that reduce the amount
$yU«&»!VJc&1'*§
               Molten Ferroalloy
                 Carbon Hearth
     i i  i  i  i  i  i i  i  i  i  i T
                                                   Refractory
                                                   Lining
                                                 Shell
                                                  Crucible
                                                    Tap Hole
 l^w,^. ;*;«^«*,r.^,^| i
                                                     /.ad/e

Figure 1.    Submerged-arc furnace for ferroalloy production

                                 2
of air drawn into the hood system. Them
covered category includes the  mix-
sealed furnaces in which a tight-fitting
cover is installed on the crucible and is
partially sealed by raw materials mounded
over the openings in the cover through
which the electrodes pass, and sealed
furnaces which are similar to the mix-
sealed furnace except mechanical seals
are used around  the electrodes. Two
emission control systems are used with
covered furnaces, one system to with-
draw gases from beneath the cover
(primary control system) and a hood
system above the cover to collect fumes
escaping the cover (secondary control
system).
  New Source Performance Standards
(NSPS) for emission to the atmosphere
from ferroalloy manufacture were based
on best available control technology for
the open type furnaces. EPA data col-
lected in support of  these standards
showed that a particulate emission
standard based on sealed furnace tech-
nology would have resulted in even
lower particulate emissions. This stan-
dard was not adopted because of an
objection that such a standard could
seriously affect the industry's ability to
respond to rapidly changing markets
conditions by restricting their ability to*
manufacture  different products in the
same furnace.
  EPA did, however, decide to further
investigate the subject of product flexi-
bility recognizing  that solution of this
problem could ultimately lead to stan-
dards of performance based on sealed
furnace technology  This task was
assigned to EPA's Industrial Environ-
mental Research  Laboratory (IERL) in
Research Triangle Park, N. C. As a first
step, IERL. analyzed some of the samples
previously obtained and found indica-
tions that sealed  furnaces generated
substantially more organics,  including
polynuclear aromatics (PNA), than did
open  furnaces. To verify this finding,
gases generated by one sealed furnace,
which was alternatively  producing
silicomanganese and  ferromanganese,
were sampled and analyzed. That study,
which experienced  some sampling
difficulties, did indicate that  a signifi-
cant concentration of PNAs exist in the
gases generated by the furnace and that
high energy venturi scrubbers might be
effective in their capture.
  Since these test results suggested
that a standard of performance based on
sealed furnaces might result in decreased
environmental protection,  a decision.
was made to more fully characterize

-------
  ollutants generated by and emitted
'from ferroalloy furnaces. The present
study is the first phase of this effort. A
complete multimedia environmental
assessment of the industry was desired;
however, funding limitations prevented
such a comprehensive study. The study
design which resulted from considera-
tion of funding limitations, and the need
to explore  the pollutant generation
potential of  several ferroalloy furnaces,
particularly the  mix-sealed type, do not
include furnace types and mode of
pollution control (i.e., baghouse or
scrubber) in the same proportions as
they exist in the industry. The design is
believed,  however, to accomplish  the
next logical step in the assessment and
to represent the best approach for  the
available funds.
  The primary objective of this study is
to determine if there is a  significant
difference in the types and amounts of
organic pollutants generated by open
and mix-sealed furnaces. To accomplish
this objective detailed testing, by EPA/
IERL-RTP Level  1 procedures, was done
at three plants. Level  1  is designed to
determine a wide variety of inorganic
and  organic species each to within at
least 1 /3 to 3 times the actual concen-
tration in the stream sampled. Some of
the information obtained, however, is
better than the overall  accuracy.  Partic-
ulate concentrations in the gas streams,
which are sampled at a single point,
should be within 1/2 to 2 times actual
values. Accuracy for gaseous compo-
nents  is not affected by the velocity
profile. Thus, total organics, which  are
determined  by extracting the samples
and weighing the residue in the extract,
should be within 1/2 to 2 times actual
stream concentrations. The final steps,
fractionation of the extracts by liquid
chromatography and  quantitation by
infrared and mass spectrographic anal-
ysis, reduce the accuracy for determining
an individual compound category to
within 1/3 to 3 times the actual concen-
tration in the stream sampled.
  Both open and mix-sealed furnaces
were tested and   products  included
ferromanganese, 50 percent ferrosilicon,
and  75 percent ferrosilicon. The study
design does  not allow a complete eluci-
dation of the separate effects of furnace
type and product  manufactured. Also,
since the gas from mix-sealed furnaces
is flared, the actual organic emission to
the atmosphere generally  cannot be
determined.
  FT wo furnaces at each of three plants
  ere tested. Scrubbers were used on
five of the furnaces and samples were
taken  of scrubber waters and of the
scrubbed gas before  it was flared. The
one furnace tested which was served by
a baghouse was sampled  before the
pollution control devices. Samples were
also taken of the plant discharge waste-
waters.

Results  and  Conclusions
  Summarized in Table 1 are the partic-
ulate generation rates by the furnaces
(before emission control). The data are
only for particutate going to the primary
emission control systems. Thus, tapping
and product handling are not included.
  With the exception of furnace A-1,
there does not seem to be a significant
difference in particulate generation
rates from variations  in product type or
type of furnace used when compared on
a kg/MW-hr basis. Furnace A-1 seemed
to be generating more secondary fume
(based  on visual estimates) than typical
mix-sealed furnaces which may account
for the low value obtained. When com-
pared  on  a  kg/Mg  of alloy produced
basis, it appears that particulate genera-
tion rates increase in the order of FeMn,
50 percent FeSi, and 75 percent FeSi.
The data are not conclusive for different
types  of furnaces since particulate
generation rates of furnaces B-1 and B-
2 are  comparable but  less than for
furnace C-2, all 50 percent FeSi product.
The difference  may  be  due to lower
efficiency (kW-hr/kg product) in furnace
C-2.
  Summarized in Table 2 are the organic
generation rate data (equivalent to Table
 1 for particulates). In this case, signifi-
 cant differences are noted when the
 generation rates are compared on either
 a kg/MW-hr or kg/Mg basis. The open
 furnaces obviously have  lower overall
 organic generation rates than the mix-
 sealed furnaces in which limited com-
 bustion was occurring.  It is interesting
 to note the variation in organic genera-
 tion rates by the different mix-sealed
 furnaces. Although the same product
 was being made in furnaces B-2 and C-
 2, the organic generation rates differ by
 almost a factor of 3 (a  wider variation
 than expected for determination of total
 organics by Level 1 procedures). This is
 probably due to more combustion under
 the cover of furnace  C-2 (indicated by
 the Orsat analyses of the furnace gases).
  Most interesting are the results for
 furnace A-1 which had almost complete
 undercover combustion. The trend
 observed for the  mix-sealed and open
 furnaces strongly indicates that  more
 complete destruction of organics would
 occur in sealed or mix-sealed furnaces
 in which complete undercover combus-
 tion was occurring.
  The efficiencies of the scrubbers for
 removal of particulate and  organic
 matter from the gases generated by the
furnaces are given in Table 3. Although
 all scrubbers  have particulate capture
 efficiencies of over 90 percent, a signifi-
cant difference in capture efficiency for
organics  is observed. As expected, the
capture efficiency increased  with an
 increase  in either pollutant inlet con-
centration or scrubber pressure drop.
Furnace Type
A-1
A -2
B-1
B-2
C-J
C-2
Mix-sealed
Open
Open
Mix-sealed
Mix-sealed
Mix-sealed
Operating
Product Power, MW kg/hr
FeMn
FeMn
50% FeSi
50% FeSi
75% FeSi
50% FeSi
11.4
15.8
48.4
48.0
15.5
16.8
47.3
174.9
470.6
447.7
196.7
187.9
kg/MW-hr
4.1
11.1
9.7
9.3
12.7
11.2
kg/Mg
alloy
10.1
26.0
49.2
46.0
103.0
68.9
Table 2.    Summary of Furnace Organic Generation Data-
Furnace Type
A-1
A-2
B-1
B-2
C-1
C-2
Mix-sealed
Open
Open
Mix-sealed
Mix-sealed
Mix-sealed
Operating
Product Power, MW
FeMn
FeMn
50% FeSi
50% FeSi
75% FeSi
50% FeSi
11.4
15.8
48.4
48.0
15.5
16.8
kg/Mg
kg/hr kg/MW-hr alloy
0.72
5.5
12.0
76.7
19.6
9.9
0.06
0.35
0.25
1.60
1.27
0.59
0.15
0.82
1.25
7.89
10.27
3.65

-------
  The concentrations of particulates
and organics  in the plant discharge
wastewaters are given in Table 4. These
effluents do not contain cooling or
sanitary water.
  All samples collected  during the test
were extracted with methylene chloride
and analyzed by infrared (IR) and  low
resolution mass spectograph (LRMS).
The analyses are not adequate for
individual compound  identification but
do  indicate compound categories and
potential compounds present. Both the
cleaned gas and the water discharged
by  the scrubber used  for control of
fumes generated by furnace C-2 were
analyzed by gas chromatograph-mass
spectrograph (GC-MS) for exact com-
pound identification.
  The IR and LRMS analyses of furnaces
A-1, A-2, and  B-1, all of which were
achieving nearly complete combustion
of the furnace gas,  indicate a low con-
centration of most  organic categories.
Potentially low concentrations of the
carcinogens,  indeno(1,2,3-cd)pyrene
and dibenzochrysene  isomer, in emis-
sions to the air from furnace A-2  are
indicated by LRMS responses at masses
276 and 302,  respectively. Similarly,
low concentrations  of the carcinogens,
benzanthracene and benzo(a)pyrene, in
gases generated by furnace B-1 (before
emission control equipment) are indi-
cated by LRMS responses at masses
228 and 252, respectively. No evidence
of potential carcinogens was found in
emissions to the air (primary emission
control  system) from furnace A-1. The
scrubber discharge water from furnace
A-1  contained organic compounds with
masses  (LRMS analysis) of 228, 252,
256, and 302 which could be the car-

Table 3.     Scrubber Efficiencies, Percent"
          cmogens, benzanthracene, benzo(a)
          pryene, dimethylbenzoanthracene, and
          dibenzochrysene  isomer,  respectively.
          The scrubber discharge water from
          furnace A-2 contained, in addition to the
          cited organic for furnace A-1, masses at
          266 and 276 (dibenzofluorene and
          indeno(1,2,3-cd)pyrene,  respectively).
            The scrubbed gases from the covered
          furnaces B-2, C-1, and  C-2 (measured
          before the flares) all coniam similar
          types  of organic  compounds although
          the concentration from the B-2 furnace
          is lower than that of the other two,
          presumably due to the higher scrubber
          efficiency for furnace B-2. For  these
          furnaces, the  LRMS analysis indicates
          significant concentrations of fused
          aromatic organics at masses 252, 266,
          276, and 302 which could be carcino-
          gens, benzo(a) pyrene, dibenzofluorene,
          mdeno(1,2,3-cd)pyrene,  and dibenzo-
          chrysene isomer, respectively. All scrub-
          ber discharge waters from these fur-
          naces contain relatively high concen-
          trations  of organics with  masses 228,
          252, 256, 266, 276, and 302 which
          could  be the  carcinogens cited previ-
          ously. Evidence for potential carcino-
          gens (at masses 228 and 252) was
          found  only in the treated process dis-
          charge water from plants C. No evidence
          of organic carcinogens  was found for
          the treated water discharged from
          plants A and B.
            The GC-MS analysis of the scrubbed
          gases from furnace C-2 (before flaring
          which should destroy some organics)
          gave positive  identification of 13 poly-
          cyclic   aromatic  hydrocarbons  (PAH)
          including the known carcinogens, benz
-------
  tble 5.    Estimated Concentrations of Identified PAHs
                                                                                                Estimated Concentrations
                                                                                                     in Unftared Gas
Compound
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzfajanthracene
Chrysene
Benzo(e)pyrene
Benzofkjf/uoranthene
Perylene
Benzo(a)pyrene
Inden o( 1 , 2, 3 - cd)p yrene
Benzo(ghi)perylene
Coronene
Flu or ene*
9-Methylphenanthrene*
Cyclopentafdef) phenanthrene*
Benzo(a)fluorene*
Methyl Pyrene*
Benzo(b)fluorene *
Benzofghijfluoranthene *
Benzo(jjfluroanthene *
Benzo(e)acephenanthrylene *
Anthanthrene*
Mass
178
178
202
202
228
228
252
252
252
252
276
276
300
166
192
190
216
216
216
226
252
252
276
Normalized
Carcinogen Relative Sample
Rating " Concentration
— 6.8
— 6.8
— 10.2
— 10.6
+ 3.9
± 3.0
— 0.30
— 0.06
— 0.16
+++ 0.61
+ 0.41
— 1.20
— 0.38
— 27.9
— 0.56
— 4.0
— 0.20
— 0.025
— 0.025
— 3.3
++ 1.3
? 1.3
— 0.31
mg/Nm3
18.3
18.3
27.4
28.5
10.5
8.1
0.81
0.16
0.43
1.64
1.10
3.2
1.0
75.0
1.5
10.7
0.54
0.07
0.07
8.9
3.5
3.5
0.83
DMEG
Air Health
Limit, mg/Nm3
1.6
56
90
230
0.045
2.2
3.0
1.6
—
2 x W's
1.6
—
—
—
—
—
—
—
—
—
6.5
—
—
  a±Weakly carcinogenic, + carcinogenic, ++ and +++ strongly carcinogenic, — not carcinogenic.
  *Tentative identification.
Table 6.    Organic Extract Summary Table, Sample No. CI-X
                                   LCI
                                               LC2
                                                           LC3
                                                                       LC4
                                                                                   LC5
                                                                                              LC6
^'Quantity Not Sufficient        '"Possible Contamination.
** The data are presented as assigned intensity (from IK and/'or LRMS)/'concentration.
                                                                                                          LC7
Total Organics, mg/m3
TCO, mg/m3
GRAV, mg/m3
Category
Aliphatic Hydrocarbons
Halogenated Aliphatics
Aromatic Hydrocarbons
Halogenated Aromatics
Silicones
Heterocyclic Q Compounds
Nitroaromatics
Ethers
Aldehydes
Phosphates
Nitnles
Heterocyclic N Compounds
Heterocyclic S Compounds
Alcohols
Phenols
Ketones
Amines
Alkyl S Compounds
Sulfur ic Acids
Suit oxides
Amides
Carboxylic Acids
Esters
2640
262.7
1.3

100/66.0"
1 0O/66. 0
tOO/66.0"
1OO/66 0"
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
81 0
59.5
21.5

	
100/16.2
100/16.2
100/16.2
100/16.2
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
100/16.2"
31.0 30.4 177
5 70 19.0 8.9
25.3 114 8.8
Assigned Intensity - mg/lm3)
- Q/VS*
—
100/280 —
—
10/0.80
10/0.80
10/0.80
100/80
10/0.80
10/0.80
10/080
10/0 80
10/0 80
—
—
100/8.0"
Q/VS* Q/VS*





10/3.0 100/8.0"
62.0 5. 1
19.0 0
43 5.1

- Q/VS*
—
—
—
—
—
—
—
—
1 0/0 94
—
1OO/9.4
—
100/9.4
1 00/9.4
10/0.94
100/9.4 Q/VS*
70/0.94
10/0.94
10/0.94
100/9.4
100/9.4
10/0.94
491 2
374.7
1165

66.0
82.2
110.2
82.2
17.0
08
08
8.0
0.8
1.74
08
102
0.8
94
9.4
8.94
94
094
094
0.94
9.4
9.4
28.14

-------
Table 7.    Estimates of Furnace Generated POM*
            Captured by Control Device
Furnace      kg/yr/MW of Capacity
      Total Generated by Furnace
        kg/yr/MW of Capacity
A-2
B-1
B-2
C-1
C-2
280
ND+
3,600
2,800
1,200
690
280
4.400
3,900
2,800
"Furnace A-1 is atypical.
+/VD = Not Determined.

the ferroalloy industry  or about 9,100
tonnes (10,000 tons), on the average,
for each plant. About 30 percent of this
material may contain wastes specifically
listed as hazardous by proposed section
3001 of  the Resources Conservation
and Recovery Act (RCRA). About 85
percent of the waste is disposed of in
landfills or lagoons which are unlined.
The dusts and sludges  from open fur-
naces may contain about 0.1 percent
organic matter. Sludges from covered
furnaces may contain up to 8 percent
organic matter. Sludges, from covered
furnaces in particular, may contain high
concentrations of polynuclear aromatic
hydrocarbons including  known carcino-
gens. Information is presented which
indicates that the POM concentration in
the clarified scrubber water should be
less than its solubility in pure water
(POMs are preferentially absorbed on
suspended solids). Since suspended
solids are generally removed from the
scrubber  water before chemical waste-
water treatment and since previous
research  has shown that POMs degrade
at a slow rate, it is likely that most POMs
collected by the scrubber accumulate in
solid waste disposal sites and disposal
lagoons. Industry tests indicate that the
dusts from a hard, fairly impermeable
mass (permeability K values of 10"4 to
10~" cm/sec) when wetted and allowed
to dry. Industry data from monitor wells
show virtually no contamination of
groundwater based on analysis for five
metals (Ba, Cd, Cr, Pb, and Hg). No data
are available on organic leaching from
these sludges. To the best of our knowl-
edge, there  is no evidence available to
prove or disprove that sealing occurs.

  The conclusions of this report are
based, in part, on sampling and analysis
data obtained using EPA/IERL-RTP
Level  1 assessment procedures which
yield final results accurate to within at
least 1 /3 to 3 times the actual value of
the stream sampled. This approach is
used to identify potential environmental
problems and is not in  itself sufficient
proof that a problem exists. Appropriately,
therefore, the data are interpreted using
the worst case approximation unless
data exist to prove this approximation
invalid. The  major conclusions of this
report are as follows.
   1.  There  are basically two  types of
      furnaces, open,  86 percent of
      installed capacity, in which com-
      bustion of the furnace gas occurs
      before the emission  control
      equipment; and covered, 14 per-
      cent of  installed capacity, in
      which  the gas is combusted after
      passing through the emission
      control equipment.
   2.  The pollution potential of covered
      (mix-sealed) furnaces is substan-
      tially higher than for open fur-
      naces, primarily due to much
      higher organic  generation rates
      by the covered furnaces. However,
      mix-sealed  furnaces appear to
      vary in the  rate of organic pro-
      duction (kg/MW-hr basis) prob-
      ably due to varying rates  of
      combustion under the  furnace
      cover.  Open furnaces are esti-
      mated to generate POM at the
      rate of about 100 to 900 kg/yr
      (220 to 1,980 Ib/yr) per megawatt
      of furnace capacity or 134,500 to
      1,210,500  kg/yr (296,500  to
      2,668,700 Ib/yr) for all U.S. fur-
      naces  of this type The  covered
      furnaces are estimated to gener-
      ate  POM at the rate of about
      1,230  to 11,080 kg/yr (2,710 to
      24,430 Ib/yr) per megawatt of
      furnace  capacity or 208,800 to
      1,878,800  kg/yr (460,300  to
      4,120,000 Ib/yr) for all U.S. fur-
      naces  of this type. Control de-
      vices, which are in use on all U.S.
      furnaces, remove most of this
      material from the furnace gas.
      Thus, the estimated  nationwide
      POM generation rates (estimated
      rates before the emission control
      devices) are in the same order of
   magnitude as POM generation^
   rates (before control devices) of^
   slot type coke ovens, a major
   POM emitter, which are estimated
   to be 317,000 to 3,200,000 kg/yr
   (700,000 to 7,000,000 Ib/yr) for
   all U.S. coke ovens.
3  The industry generates about
   363,000 tonnes (400,000 tons)
   of solid waste annually, about 85
   percent of which is disposed of in
   unlined lagoons and landfills.
   Although the wastes contain
   known and/or suspected hazard-
   ous  inorganic and organic mate-
   rials, there is some evidence that
   the  wastes  are self-sealing  and
   that heavy metals do not leach
   into the groundwater.
4.  The  industry consumes about  9
   million megawatt hours of elec-
   tricity  annually, 6  percent of
   which is used for pollution con-
   trol. Open and mix-sealed fur-
   naces use up to 5 times as much
   energy for pollution control as
   does a typical totally sealed fur-
   nace.
5.  For the six furnaces tested, there
   appears to be no significant dif-
   ference  in the kg of paniculate
   generated/megawatt hour of
   furnace  power (before emission
   control)  as a function of furnace
   size, type, or product being man-
   ufactured. There does appear to
   be a difference in the kg of partic-
   ulate (per megawatt hour of fur-
   nace power) in the gas discharged
   from the scrubber, which appears
   to be related to scrubber design
   and  pressure drop, but may also
   be a function of furnace  type
   and/or product being manufac-
   tured.
6.  Scrubbers appear to be less ef-
   ficient for capturing organics
   than for particulate capture.
7.  Low resolution mass spectro-
   graphic  analysis  indicates  the
   potential presence of carcinogens
   in the cleaned  gas from  the
   scrubbers, before  it was flared,
   from four of five furnaces tested
   (the  exception being one  mix-
   sealed furnace in which complete
   undercover combustion was ap-
   parently occurring), and in the
   gas  from one open furnace which
   was tested before emission
   control.
8.  Low resolution mass spectro-
   graphic  analysis  indicates  the
   presence of potential carcinogens

-------
      in all scrubber discharge waters
      and  in the plant discharge water
      from only one plant (no longer
      operating) of the three tested.
    9. Analysis of samples of one mix-
      sealed furnace by GC-MS tech-
      niques gave positive identification
      of known carcinogens in the
      cleaned  gas discharged by the
      scrubber (but before passing
      through the flare  which  may
      destroy some of the organics) and
      in the scrubber discharge  water
      (before wastewater treatment).
      Two of these carcinogens  could
      exceed DMEG1 values by factors
      of up to 200 and 80,000 respec-
      tively, if  significant destruction
      does not occur in the flare. These
      data provide strong evidence that
      the  preliminary identifications
      listed above in 7 and 8 are prob-
      ably correct.
   10. U.S. production of ferroalloys has
      declined  during the last decade to
      about  1945 levels. Imports have
      risen from about 2.4 percent of
      domestic consumption in 1945 to
      over 40 percent in the years since
      1975.
   11. Unless action is taken soon to
      stem the tide of imports, the
      continued viability of the  U.S.
      industry  is questionable.
   12. There are noplanstoexpandU.S.
      production capacity. Rather, some
      furnaces are idle, some plants
      are being closed, and some older
      furnaces are being  replaced by
      larger, more efficient furnaces.
   13. Based on information obtained in
      these tests, we  must conclude
      that  a potential for a significant
      multimedia environmental  prob-
      lem  exists with ferroalloy manu-
      facture and that  this potential is
      significantly greater for plants
      using mix-sealed and sealed fur-
      naces than for those using open
      furnaces. It has not been estab-
      lished that a real environmental
      problem exists in any of the three
      media—air, water, or solid waste.

Recommendations
  More accurate testing should be done
to quantify the  pollutants produced by
'Kmgsbury, G L , etal "Multimedia Environmental
Goals for Environmental Assessment - MEG Charts
and Background Information Summaries," Vol Ill-
Categories 1-12, EPA-600/7-79-176a (NTIS PB80-
 15108), and Vol IV-Categones 13-26, EPA-600/
 -79-176b(NTIS PB80-115116), August 1979
 the furnaces and determine how much
 is ultimately discharged to the environ-
 ment through any and all three media. If
 these tests should prove that unaccept-
 able amounts of pollutants are emitted,
 or are disposed of in an environmentally
 unsound manner, work should be initi-
 ated to determine if the public is being,
 or  is likely to be, endangered. If these
 studies indicate public endangerment,
 studies should be undertaken to reduce
 pollutant releases from the industry
  Specifically, the following additional
 work is recommended.  More accurate
 sampling (i e , isokmetic, duct traverse,
 integrated composite water sampling)
 and analysis (GC-MS, for example) need
 to be used to quantify discharges from
 the plants to all media. For plants using
 only open furnaces and capturing and
 disposing of only dry dust (baghouse
 control system), sampling will be re-
 quired for emissions from the baghouse
 and for surface water runoff and ground-
 water intrusions from the dust disposal
 site A few locations control emissions
 from open  furnaces  with scrubbers or
 slurry the  dust captured by the  bag-
 house. The number  and size of these
 facilities are probably not large enough
 to warrant detailed testing. Sampling in
 the gas stream before the control device
 (baghouse) and of the collected  bag-
 house dust is also recommended since
 these  tests will  allow  a measure  of
 control efficiency for  the contaminants,
 a measure of contaminants entering the
 disposal sites,  and an indication  of
 possible emissions in the event  of
 control device failure (bag rupture, etc.).
  Quantifying emissions to the air from
 covered (mix-sealed and sealed) fur-
 naces is extremely difficult since the gas
 is flared on discharge to the atmosphere
 At  present, there are  no established
 techniques for  measuring emission
 rates from flares. It  is  recommended,
 therefore, that the gas be sampled in the
 duct after the scrubber and before the
 flare This should provide a  reasonable
 estimate of particulate emissions, al-
 though some change in mass is to be
 expected since flaring may change the
form of some of the particulate compo-
 nents and is expected to burn off some
 of the organics on the particulate matter.
 Determining the actual organic emis-
 sion rate is complicated by the fact that
the  flare will destroy some of the organic
 matter and the percentage destruction
(for total organics or for  individual
compounds) cannot be accurately mea-
sured. As a first approximation, it can be
assumed that the flare is 100 percent
effective and the emission rate calcu-
lated based on the percent of time that
the flares are not operating. Other
assumptions about flare efficiency
could be made. If adequate methods are
developed, an actual assessment  of
flare effectiveness should be made.
  The  wastewater discharged by the
plant should be  analyzed for priority
pollutants including polynuclear aro-
matics.  The  possibility of  leaching
inorganics and organics intotheground-
water  at disposal sites and  lagoons
should be examined.
  It is recommended that, in conjunction
with the  above tests, the water dis-
charged by the scrubbers on the furnace
be tested since this provides  informa-
tion as to the control efficiency of both
the scrubber and the wastewater treat-
ment system.
  If the above test should prove that
unacceptable amounts of pollutants are
emitted or are disposed of in an environ-
mentally unsound manner, work should
be initiated to determine  if the public  is,
or is likely to be, endangered. To ac-
complish this, modeling studies for the
pollutants of concern should be done to
determine the potential  impact on the
population surrounding a plant.
  If the weight of evidence gathered
indicates public endangerment, work
should be initiated to reduce pollutants
emitted by the industry. While we
cannot predict with certainty which
pollutants would be involved or which
media would  have the most impact, we
can suggest some areas  in which addi-
tional work might be fruitful. Included in
the suggested efforts below are some
already being instituted by the industry.

  1. Improve flare design and opera-
     bility.
  2. Improve scrubber efficiency,
     particularly for organics.
  3. Reduce gas volume  from open fur-
     naces, possibly by the use of close
     hooding.
  4. Investigate the possibility of con-
     trolled undercover combustion  in
     mix-sealed and sealed type fur-
     naces for organic matter  destruc-
     tion. *
  5.  Investigate improved water treat-
     ment methods, including clarifica-
     tion  and filtration  for improved
     suspended solid removal  and an
     investigation of the  applicability  of
     reuse and/or recycle of waste-
     water since this has the potential
     for significantly reducing mass
     emissions of suspended solids (on
       US GOVERNMENT PRINTING OFFICE 1981  757-012/7243

-------
   which polycyclic aromatic  hydro-
   carbons can be absorbed) and
   dissolved materials.
6.  Investigate alternate methods for
   treatment or disposal of solid
   wastes generated.
                                                                                                                      4
                                                C.  W.  Westbrook and D.  P. Daughterly are with Research Triangle Institute,
                                                  Research Triangle Park, NC 27709.
                                                Robert C. McCrillis is the EPA Project Officer (see below).
                                                The complete report, entitled "Level 1 Environmental Assessment of Electric
                                                  Submerged-Arc Furnaces Producing Ferroalloys," (Order No. PB81 -210 106,
                                                  Cost: $24.50, subject to change) will be available only from:
                                                       National Technical Information Service
                                                       5285 Port Royal Road
                                                       Springfield.  VA 22161
                                                       Telephone  703-487-4650
                                                The EPA Project Officer can be contacted at:
                                                       Industrial Environmental Research Laboratory
                                                       U.S. Environmental Protection Agency
                                                       Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
                             Center for Environmental Research
                             Information
                             Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
          V i i
            b
                           jTr.cn '
                    II

-------