United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Cincinnati, OH 45268
EPA-600/9-84-022
September 1984
Research and Development
Incineration and
Treatment of
Hazardous Waste

Proceedings of the
Tenth Annual
Research Symposium

-------

-------
                                              EPA-600/9-84-022
                                              September  1984
       INCINERATION AND TREATMENT OF HAZARDOUS WASTE
    Proceedings of the Tenth Annual Research Symposium
        at  Ft. Mitchell, Kentucky, April 3-5, 1984

Sponsored by the U.S.  EPA, Office of Research & Development
        Tndustrial Environmental Research Laboratory
             Energy Pollution Control Division
                           and
        Municipal Environmental  Research Laboratory
        Solid and Hazardous Waste Research Division
                      Coordinated by:

                        JACA Corp.
           Fort Washington, Pennsylvania  19034
                  Contract No. 68-03-3131
                     Project Officer

                      Harry Freeman
              Energy Pollution Control Division
         Industrial Environmental Research Laboratory
                  Cincinnati, Ohio  45268
      INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
          OFFICE OF RESEARCH AND DEVELOPMENT
         U. S. ENVIRONMENTAL PROTECTION AGENCY
                CINCINNATI, OHIO  45268

-------
                                   NOTICE
      This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication.  Mention of trade
names or commercial products does not constitute endorsement or recommen-
dation for use.
                                    -ii-

-------
                                 FOREWORD
     We at the Industrial Environmental Research Laboratory in Cincinnati
have always believed that an important part of our responsibility is to
provide to the technical community the results of our various projects in as
timely a manner as possible.  This Proceedings document, containing papers
presented at our 1984 Research Symposium, presents project summaries and
interim results of our current work for this purpose, since much of the work
discussed may not be available in final reports for some time. We hope that
you will find these papers useful.  Any questions or comments that you
might have should be directed to the Chief, Incineration Research Group, IERL.
                                         David G. Stephan
                                         Director
                                         Industrial Environmental
                                           Research Laboratory
                                       -iii-

-------
                                       PREFACE
     These Proceedings are intended to disseminate up-to-date information on extra-
mural research projects concerning land disposal, incineration, and treatment of
hazardous waste.  These projects are funded by the U.S. Environmental Protection
Agency's Office of Research and Development and have been reviewed in accordance with
the requirements of EPA's Peer and Administrative Review Control System.
                                       -IV-


-------
                                 ABSTRACT
     The Tenth Annual  Research Symposium on Land Disposal,  Remedial  Action,
Incineration and Treatment of Hazardous Waste was held in  Fort Mitchell,
Kentucky, April  3 through 5, 1984.  The purpose of the symposium was to
present to persons concerned with hazardous waste management the latest
significant findings of ongoing and recently completed research projects
funded by the Industrial Environmental  Research Laboratory's Energy  Pollution
Control Division and the Municipal Environmental Research  Laboratory's Solid
and Hazardous Waste Research Division.

     This volume is a compilation of speakers' papers and  poster presenters'
abstracts for Session B, Hazardous Waste Incineration and  Treatment.
Subjects include innovative technology, lab-scale research  projects, pilot
and full-scale incineration projects, combustion of hazardous waste  in
boilers and in industrial processes, and hazardous waste treatment and cost
studies.

     This document covers hazardous waste incineration and  treatment only.  A
separate document for Session A, Hazardous Waste Land Disposal, is available
from the Municipal Environmental Research Laboratory.
                                    -v-

-------
                                 CONTENTS
                                                                   Page
Environmental Characterization of the Combustion of Waste
Oil in Small Commercial Boilers
     Paul F. Fennelly, Mark McCabe, and Joanna Hall
     GCA Corp	     1

Demonstration of Wet Air Oxidation of Hazardous Waste
     William Cop'a, James Heimbeck, and Phillip Schaefer
     Zimpro, Inc	     9

Evaluation of Pilot-Scale APCDs for Hazardous Waste
Combustion Sources
     C.W. Westbrook
     Research Triangle Institute 	 	    19

Disposal of Hazardous Waste in Aggregate Kilns
     James Peters and Duane Day, Monsanto, and
     Robert Mournighan, U.S. Environmental Protection Agency .  .    24

Effects of Disposal of Hazardous Wastes in Cement Kilns on
Conventional Pollutant Emissions
     James Peters, Monsanto, and Robert Mournighan,
     U.S. Environmental Protection Agency	    38

Evaluation of Hazardous Waste Incineration in a Lime Kiln
     Duane Day, Monsanto, and Robert Mournighan,
     U.S. Environmental Protection Agency	    48

Field Tests of Industrial Boilers Cofiring Hazardous Wastes
     Carlo Castaldini, Howard Mason, Robert DeRosier, and
     Stefan Unnasch
     Acurex Corp	    57

Field Tests of Industrial Boilers and Industrial Processes
Disposing of Hazardous Wastes
     Radford Adams, Michael Hartman, and Denny Wagoner
     Radian Corp	    62

Summary of Field Tests for an Industrial Boiler Disposing of
Hazardous Wastes
     John Chehaske, Engineering-Sciences,  and
     Gregory Higgins, SYSTECH Corp	    70

Parametric Experimentation with a Pilot-Scale Boiler Burning
Hazardous Compounds
     C. Wolbach, Acurex Corp	    76
                                     -vi-

-------
                            CONTENTS (Cont'd)
Products of Incomplete Combustion from Hazardous Waste
Incinerators
     Andrew Trenholm and Roger Hathaway,-
     Midwest Research Institute, and Don Oberacker,
     U.S. Environmental Protection Agency	
Preliminary Assessment of Costs and Credits for Hazardous Waste
Co-Firing in Industrial Boilers
     R. McCormick and L. Weitzman
     Acurex Corp	     

Hazardous Waste Pretreatment as an Air Pollution Control
Technique
     James Spivey, C. Clark Allen, and Robert Stall ings,
     Research Triangle  Institute, and Benjamin Blaney,
     U.S. Environmental Protection Agency	
 84
 96
Programs of the Industrial Waste Combustion Group at the U.S.
EPA Center Hill Facility
     George Huffman, U.S. Environmental Protection Agency. . . .

Determination of the Thermal Decomposition Properties of 20
Selected Hazardous Organic Compounds
     Barry Bellinger, Juan Torres, Wayne Rubey, Douglas Hall,
     and John Graham, University of Dayton, and Richard Carnes,
     U.S. Environmental Protection Agency	
104
114
116
Hazardous Waste Destruction Using Plasma Arc Technology
     Thomas Barton, Pyrolysis Systems Incorporated;
     Nicholas Kolak, New York State Department of Environmental
     Conservation; and Chun Ching Lee,
     U.S. Environmental Protection Agency	
Metal Value Recovery from Metal Hydroxide Sludges
     L. Twidwell, A. Mehta, and G. Hughes, Montana College of
     Mineral Science & Technology. .  .	 
Operations  at the U.S.  EPA Combustion Research Facility
      Richard Carnes, U.S. Environmental Protection Agency.  .  .

Sampling Methods for Emissions  from Hazardous Waste Combustion
      Larry  Johnson, U.S.  Environmental Protection Agency  .  .  .

Incineration of Hazardous Waste in Power Boilers:  Emissions
Performance Study Rationale  and Test Site Matrix
      Robert Olexsey, U.S. Environmental Protection Agency.  .  .
124



129


143


155



162
                                      -vii-

-------
                            CONTENTS (Cont'd)
Posters
Observation of Fluid Dynamic Effects upon High Temperature
Destruction of Organic Compounds
     Wayne Rubey, John Graham, Douglas Hall, and Barry Del linger
     University of Dayton Research Institute 	   178
Research Projects at EPA's Louisiana State University Hazardous
Waste Research Center - US EPA Center of Excellence
     D.P. Harrison, Louisiana State University 	
Hazardous Waste Control Technology Data Base
     Richard Holberger, The MITRE Corp., and
     C.C. Lee, U.S. Environmental Protection Agency.
Evaluation of an Advanced Low NOX Heavy Oil Burner for
Incineration of Nitrated Hazardous Compounds
     W.S. Lanier and Charles Sedman
     U.S. Environmental Protection Agency	
Regulatory Support Research Studies Supported by EPA/Cincinnati
MSD Cooperative Agreement
     Boyd T. Riley, RYCON	
Flame-Mode Hazardous Waste Destruction Research
     W. Randall Seeker, John Kramlich, Michael P.  Heap, and
     Rachel K. Nihart, Energy and Environmental Research Corp.,
     and Gary Samuelson, University of California	  .
Evaluation and Development of Innovative Incineration Testina
Procedures
     Jerome Strauss, Versar, Inc	
179
180
181
182
183
184
                                     -viii-

-------
                    ENVIRONMENTAL CHARACTERIZATION OF THE COMBUSTION
                        OF WASTE OIL IN SMALL COMMERCIAL BOILERS

                       Paul F. Fennelly, Mark McCabe, Joanna Hall
                          GCA Corporation, Technology Division
                                   Bedford, Ma. 01730
                                        ABSTRACT

     Tests were conducted on seven boilers in the size range of 0.4 to 20 X 106 British
thermal units per hour (Btu/hr) which were burning a waste oil fuel spiked with a series
of organic chemical compounds commonly found in waste oil.  Compounds tested included
chlorinated solvents such as chloroform, trichloroethylene, trichloroethane, tetrachloro-
ethylene and several chlorinated benzenes.  Test results are presented here in terms of
stack gas emission concentrations and destruction efficiencies for the various compounds.
INTRODUCTION AND PURPOSE OF THE PROJECT

     In recent years the environmental
impact of the disposal of used oils has
been an area of growing concern.  Numerous
studies conducted by state and Federal
agencies have documented the presence of
contaminants such as polynuclear aromatics
(PNAs), chlorinated hydrocarbons and heavy
metals in samples of used motor oils.  One
of the more common and widespread prac-
tices for disposing of used oils is burning
as a supplemental fuel.  In some cases,
waste oil is burned directly, in others, it
is blended with other fuel feedstocks.

     The disposal of waste materials in
boilers is of particular interest because
to date, there has been little documenta-
tion of the extent to which chemical con-
taminants- in waste oil are destroyed during
the combustion-process.

     In this project, tests were conducted
on boilers in the size range of 0.4 to
20 X 106 Btu/hr.  These are commonly clas-
sified as commercial sources, as opposed to
industrial or electric utility sources.
Commercial boilers are of particular inter-
est with regard to waste oil disposal for
several reasons.  These units generally
would use untreated or poorly characterized
waste fuels.  They could be expected to
provide less efficient combustion because
of the'generally intermittent mode of
operation.  In addition, their widespread
distribution and their low stack heights
make their emissions more proximate to the
general population.

APPROACH

     Seven boilers were designated for
testing in the program.  The units were
selected so as to provide a representative
cross section of the commercial boiler
population.  A 4000 gallon lot of used
automotive oil was obtained in order to
maintain a consistent supply of waste fuel
for the program.  Portions of. the base
stock oil were spiked with predetermined
amounts of selected organic compounds which
are typically found in waste oil and in
some cases are considered hazardous waste
materials.  The selected compounds were as
follows:
     e  chloroform
     c  1,1,1-trichloroethane
     e  trichloroethylene
     
-------
      These compounds were  chosen  to simu-
 late a typical waste oil,  based on data
 from a cross section of representative
 waste oils.1

      Measurements were conducted  at each
 of the sites to determine  the atmospheric
 emissions of particulate,  inorganic com-
 pounds (principally lead an.d hydrogen
 chloride (HC1), and volatile organic and
 semivolatile organic material.  The
 destruction and removal efficiencies (ORE)
 for each of the spiked components were
 also determined.

 TEST RESULTS

 Destruction and Removal Efficiencies

      The flue gas emissions of the organic
 compounds of interest correspond to 
 destruction and removal -efficiencies of
 99.4 to 99.99 percent as  indicated in
 Table 1.   There were no strong correla-
 tions between destruction  efficiency and
 boiler size or firing technique'.   One
 trend that is apparent- from the  data is
 that the  destruction efficiencies  for  the
 semivolatile  compounds  are  consistently
 higher than those of the volatile  com-
 pounds.   The  fact that  generally higher
 ORE'S  were  achieved  for the semivolatile
 components, trichlorobenzene,  dichloro-
 naphthalene and trichlorophenol, is  con-
 sistent with  the  ranking.of the spike
 compounds on  the  EPA Hierarchy of  Waste
 Incinerability.2

      Generally  the lowest DRE-'s were
 found  in  site A,  the only boiler rated  at
 less  than 1 X  106 Btu/hr.   This unit nor-
 mally fires a No. 2  fuel oil and its
 adaption  to firing waste oil proved dif-
 ficult.   Eventually  dilution of the waste
 oil on a  1:1 basis with No. 2 oil  was
 required  for acceptable operation.  Even
 with this modification, the combustion
 efficiency and  destruction  efficiencies
were significantly lower than the  other
 units.

 Concentrations  of Contaminants in
 Combustion Gases '

     In Table 2 are shown the concentra-
tion ranges in  the stack gas of the com-
pounds studies within the program.   In
general, concentrations ranged from 40 to
400 micrograms  per cubic meter (ug/m3)
 for the volatile compounds and from about
 10 to 50 yg/m3for the semivolatile com-
 pounds.  On a volume/volume (v/v) basis,
 these are very low concentrations.  As an
 example, a range of 40 to 400 ug/m3 for a
 compound such as trichloroethylene corre-
 sponds to a concentration of 7.4  to 74
 parts per billion (ppb) on a v/v basis.
 Conducting emission tests at these low
 concentrations required extensive refine-
 ment of available emission source testing
 techniques.

 Lead and Other Metal  Emissions

      The samples of flue gas particulate
 collected at each site were analyzed for a
 total  of 27  metals by Inductively Coupled
 Argon  Emissions Plasma Spectrometry (ICAP)
 techniques.   Lead and zinc were  present at
 concentrations  substantially higher than
 any other trace metals.   The lead concen-
 trations in  the flue  gas samples  ranged
 from 5,380 ng/m3 to  72,400 yg/m3,corre-
 sponding to  an  average emission  rate of
 0.12 Ib/hr.   Calculations  based  on simpli-
 fied models  have shown in  some cases,  lead
 emissions  at these levels  could  cause  vio-
 lations  of ambient air quality standards
 for lead.  The  concentrations  of  zinc  in
 the flue gas  ranged from 3,100 to  34,000
 ug/m3,  corresponding  to  an average emission
 rate of  0.06  Ib/hr.   The ratio between  lead
 and zinc emissions was generally  2:1,  con-
 sistent  with  their concentration  in  base
 stock  oil, which  was  1,550 ppm and 760  ppm
 by  weight, respectively.   Lead and zinc
 compounds  are commonly found in waste auto-
 motive oil and  result  from both gasoline
 and oil  additives.

     Mass  flow  calculations  indicated that
 50-60 percent of  the  lead  and  chloride
 introduced into the boilers exits  from the
 system in  the flue gas.  An additional
 10-15 percent appears  in ash collected in
 the firetubes of  the boilers.

     Other metals which received special
 attention were arsenic, cadmium and chro-
mium.  These were generally at low enough
 concentrations in the stack gas so that
when diluted in the atmosphere they should
not cause major problems, but the  situa-
 tion is still of some concern as  the con-
centration of these metals could  be sub-
stantially higher in other waste  oil base
stocks.

-------
   ,  The results of metal emissions are
summarized in Table 3.

Particulate and Chloride Emissions

     Particulate emission rates at the six
sites ranged from 0.07 to 1.2 Ib/hr with an
average value of 0.73 Ib/hr (0.34 lb/106Btu
heat input).  This is significantly higher
than the literature value of 0.09 lb/106Btu
for commercial boilers firing residual
oil,3 but the higher value-is consistent
with the much higher ash content of waste
oil, which can range from 0.15 to 1.5 per-
cent.  Particulate sizing measurements con-
ducted at four of the test sites indicated
that 80 to 90 percent of the particulates
containing lead are submicron in nature and
would be readily inhalable.

     The flue gas emissions  of HC1 from the
six boilers averaged 2.6 Ib/hr.  This is a
relatively high emission rate for such
small units, but it is below the 4.0 Ib/hr
air emission standard established for
hazardous waste incinerators, which would
typically burn large quantities of chlori-
nated compounds similar to those used in
this program.

Products of Incomplete Combustion

     The flue gas samples from each site
were screened by Gas Chromatograph/Mass
Spectrometer(GC/MS) for additional organic
components, considered to be potential
products of incomplete combustion.  The
types of compound which were-identified are
shown in Table 4.  In general, the compo-
nents were nonchlorinated in nature and
were representative of the types of com-
pounds that result from the  combustion of
traditional fossil fuels.3>4  These com-
pounds were also very typical of contami-
nants sometimes found on the blank sample
adsorbing medium, XAD-2 resin.  The extent
to which these compounds, when detected,
resulted from combustion byproducts or from
resin contaminants could not be determined;
hence, the concentrations in Table 4 could
be viewed as upper limits for many of the
nonchlorinated products of incomplete com-
bustion.  During the course  of this program
there were some baseline runs done on con-
ventional No. 4 fuel oil.  As expected, no
chlorinated hydrocarbons were detected in
the stack gas, with detection limits being
8 yg/m3.  With conventional  No. 4 oil,
combustion products such as naphthalene
and similar polyaromatic hydrocarbon (PAH)
compounds were 100 yg/m3 or less.

     Chlorinated dibenzofuran (PCDF) or
chlorinated dioxin (PCDD) species were
detected in 15 of the 25 samples analyzed.
The concentration of these compounds
ranged from 0.07 to 17 yg/m3.  On a v/v
basis, this corresponds to a range of 7 to
470 parts per trillion (ppt).

     Bulk samples of firetube ash collec-
ted at one of the sites contained parts
per billion levels of 11 PCDF and PCDD.
isomers on a weight by weight basis.
Because chlorinated dioxins and chlorina-
ted dibenzofurans were found in the flue
gas, tests were also conducted on the
waste oil base stock, both spiked and
unspiked to determine the extent to which
these types of compounds might be present
in the oil.  No chlorinated dioxins or
chlorinated dibenzofurans were found in
either the spiked or unspiked oil at
detection limits which were 0.2 milligrams
per kilogram (mg/kg) for the tetrachloro-
dibenzodioxin and as low as 0.04 mg/kg for
the monochlorodibenzodioxin.  The same
detection limits were achieved for the
chlorinated dibenzofurans.

CONCLUSIONS

     Although a sample population of six
boilers is very limited, there do seem to
be several general conclusions which can
be reached regarding the combustion of
waste automotive fuels in boilers in this
size range.

1.  It is possible to achieve combustion
efficiencies greater than 99.9 percent for
small commercial boilers firing waste
oils.

2.  Destruction and removal efficiencies
of greater than 99.9 percent can be ob-.
tanned for chlorinated organic contami-
nants typically present in waste oils.
For the volatile compounds studies
(Chloroform, trichloroethylene, trichloro-
ethane and perchloroethylene), destruction
and removal efficiencies were on the order
of 99.9 percent.  For the semi volatile
compounds, )trichlorobenzene, dichloro-
naphthalene, trichlorophenol), destruction
and removal efficiencies were on the order
of 99.95 percent.         '

-------
3.  For boilers above 1 X TO6 Btu/hr input,
there were no apparent correlations between
boiler size or firing method and destruc-
tion efficiency of organic contaminants.

4.  Inorganic components, as opposed to
organic components of waste oil, have sub-
stantially greater mass emission rates to
the atmosphere, as a result of the combus-
tion of automotive waste oils.  The princi-
pal inorganic components of concern are
lead, hydrochloric acid and total particu-
late.  'Also of potential concern are
arsenic, cadmium and chromium.  The parti-
al! ate lead emissions from a source may,
during the peak heating season, affect the
compliance with the primary ambient air
standard for lead.  A significant percen-
tage of the particulate .lead emissions is
submicron in nature and would be readily
inhalable.

5.  Detectable levels of emissions of poly-
chlorinated dibenzofurans and polychlori-
nated dibenzodioxin compounds' were found in
some of the boilers tested.  These com-
pounds, when present, were usually at
levels less than 5 yg/m3, which is less
than 0.5 part per billion by volume in the
stack gas.  The extent to which these com-
pounds pose a hazard at these low levels is
undetermined.

    Tests were done on the base .stock waste
oil, with and without the spiked contami-
nants, to determine the extent to which the
oil may have contained trace levels of
dioxin.  No dioxin or dibenzofuran com-
pounds were detected in any of the oil
samples; detection limits were 200 ppb by
weight for TCDD and TCDF.  If dioxin com-
pounds were present at or below their
detection limits, such a quantity would
not be large enough to account for the ob-
served levels in the stack gas even with
zero percent destruction.  The conclusion
is that dioxin and dibenzofuran found in
the stack gas most probably was formed
during the combustion process.

6.  The fly ash deposited in the firetubes
of the boilers may contain percent levels
of lead and parts per billion levels of  ,
chlorinated dibenzofuran and dioxin com-
pounds.  The ash has the potential for
being classified as hazardous on this
basis, and may be subject to hazardous
waste regulations for disposal.
ACKNOWLEDGEMENTS

     This work was done under contract to
the U.S. Environmental Protection Agency
by GCA Corporation under Contract No.
68-02-3168.  The authors would-like to
thank our Project Officers, Mr. Harry
Freeman of EPA's Industrial and Environ-
mental Research Laboratory in Cincinnati
and Mr. Mike Petnjska of EPA's Office of
Solid Waste in Washington, DC for their
direction and support.  Mr. George
Huffman of EPA's Industrial and Environ-
mental Research Laboratory also provided
important technical direction in the early
stages of the program.

-------
(





h-
p?
LU
CsC
LU
O_
*  *

CO
LU
O
^y
LU
i  i
C_3
i  i
u_
u_
LU
r _ j
o
LU
Q:
o
<
ZL
o
1 1
I-
  1
o:
1
LU
O
Q
LU
1 
3
 i
5
.

LU
	 I
CO
 o
 I
to -C
i  O
o
^>

o
CO

CD
CD


CO
CD

CD
CD



CD
CO

CD
CD
n
CD
CD


r-.
01
01
01


LO
CD
CD
CD

00
^
CD
CD







OJ
to
-E
+J
a>
o
S-
o
1 

CJ
 r
S-
h-



CM
CO

CD
CD


CO

CD
CD



CM
CD

CD
CD
LO
CO
CD
CD


CD
00
CD
CD


CM
CD
CD
CD

LO
**
CD
CD





0)
E
a>
_c
4-^
O
0
i 

O
I
S-
H-



^J-
co

CD
CD


10
01

01
01



CO

CD
CD
r-.
CD
01


00
CD
01


,_
CD
CD
CD

^.
r^
CD
CD


to
o
E
CD rs
E O
O) Q.
>5 O
-E O
4J
OJ CU
O i 
S- -r-.
O 4->
r1- tO

u o
S- >
0) .,-
 O
cu o. !E
-Q re o
O E 'r-
S- 0 S-
0 S- I
r O 1
-El  LO
CJ .E 
i O ^d"

1 Q CM



-------
TABLE 2. AVERAGE FLUE GAS CONCENTRATIONS OF THE VOLATILE
            AND SEMIVOLATILE SPIKE COMPONENTS


Site
Volatile Components
Chi prof orm
Trichloroethane
Tri chl oroethyl ene
Tetrachl oroethyl ene
Semi -Volatile Components
Trichlorobenzene
Di chl oronaphthal ene
Trichlorophenol

A

no
90
160
160

48
16
10

C

120
88
94
200

20
12
9

D

44
22
66
180

75
22

(yg/m3)
E

53
580
62
220

51
23
16

F

88
400
70
170

48
36
6

- G

153
140
200
100

89
89


TABLE
3. CONCENTRATIONS
OF METALS IN FLUE GAS
(yg/m3)
Site
Arsenic
Cadmi urn
Chromium
Lead 9
Zinc 5
A
11.2
31.2
62.2
,680
,150
C
655
102
166
72,400
33,700
D
26.
8.
112
5,390
3,134
E
1 106
3 182
230
20,300
12,100
F
251
350
205
49,800
26,800
G
286
81
263
51,000
27,000

-------
                TABLE 4. EXAMPLES OF CONCENTRATION RANGES FOR POSSIBLE
                           PRODUCTS OF INCOMPLETE COMBUSTION
(yg/m3)
Site A
Naphthalene,
cl
Benzaldehyde 10-60
a
Chlorobenzene isomers
Dibenzofuran 70
Ethyl benzene
Chlorotoluene 9,
Benzole acid 70
a
Alkyl benzene, 60
a
C D
30 70
20-70
20
5 2
2-20
70,
20-200
10-100
E
10
30
10
10

--
200
100-600
F
40
40
40
20
--
__
300

G
50
20
--
5
--

90
300
   This compound is a known contaminant of XAD-2 resin  and could be  a  sampling
   and analysis artifact as well  as a product of incomplete combustion.

,   At sites A and D, chlorotoluene was not spiked into  the waste oil;  hence,
   its appearance here is as a product of incomplete  combustion.

Note:

   The symbol means the compound was not detected  by mass  spectrometn'c
   computer matching scan.  In such cases,  the detection  limit  is  generally
   100 yg/m3 or less.

-------
REFERENCES

1.  Surprenant, Norman F., William H.
    Battye, Paul F. Fennelly, 1983.
    The Fate of Hazardous and Non-
    Hazardous Wastes in Used Oil
    Disposal and Recycling, report
    prepared by GCA Corporation,
    Bedford, MA for U.S. Dept of Energy,
    U.S. DOE Report No. DOE/BC/10376-6.

2.  EPA Hierarchy of Haste Incinerability-
    Organic Hazardous Constituents Under "
    Incineration Rules, released June 21,
    1982, Environmental Reporter,
    June 25, 1983.

3.  Surprenant, Norman F., et al., 1980.
    Emissions Assessment-of Conventional
    Stationary Combustion Systems:
    Volume IV:  Commercial/Institutional
    Combustion Sources, EPA-600/7-79-029e,
    U.S. Environmental Protection Agency,
    Industrial Environmental Protection
    Agency, Research Triangle Park, NC.

4.  Hansen, R. L., et al., 1975.
    Chemical Characterization of Poly-
    nuclear Aromatic Hydrocarbons in
    Airborne Effluents from an Experi-
    mental Fluidized Combustor, presented
    at the Fourth International Symposium
    on Polynuclear Aromatic Hydrocarbons,
    Battelle Columbus Laboratories.

-------
                  DEMONSTRATION OF WET AIR OXIDATION OF HAZARDOUS WASTE

                                   William Copa, Ph.D.
                                     James Heimbuch
                                    Phillip Schaefer
                                   Zimpro Incorporated
                                  Rothschild, WI  5W

                                        ABSTRACT

     The purpose of this paper is to suimarize the full scale demonstration of Wet Air
Oxidation of hazardous waste.  This work is being done at Casmalia Resources Management,
a commercial waste treater in California.

     The report will sunmarize bench scale waste screening tests completed to date, as
well as report the results of full scale oxidation of six specific wastes.  Detailed
analysis including Gas Chromatograph/Mass Spec (GCMS) scans of feed and effluent will be
given.  Wastes tested include cyanide wastes, phenolic wastes, sulfide wastes, non-
halogenated pesticides, solvent still bottoms and a general organic wastewater.

     Preliminary screening tests and the full scale work will be collated such that
predictions might be made for the various groups of compounds.
INTRODUCTION

     Wet Air Oxidation is a process
which has been used to oxidize dissolved
or suspended organic substances at
elevated temperatures and pressures.
The process is thermally self-sustaining
with relatively low organic feed
concentrations arid is therefore most
useful for wastes which are too dilute
to incinerate economically yet too toxic
to treat biologically.

     The process has been used to treat
various wastes over the last thirty to
forty years.  With the recent attention
being focused on hazardous waste, much
interest has been expressed in the use
of Wet Air Oxidation as a means of
destroying and/or detoxifying these
hazardous wastes.  During the recent
years much bench scale (5X4) and pilot
plant (2) testing has been performed by
various companies to demonstrate the
applicability of Wet Air Oxidation on
various hazardous organic wastes.  Wet
Air Oxidation units currently detoxify
specific waste streams at several
generators (1-).
     In this project, a 10 gallon per
minute (gpm) skid mounted Wet Air
Oxidation is used.  This unit is located
at Casmalia Resources, a commercial
waste treater in California.  Casmalia
Resources is using the unit to process
those acutely hazardous wastes which
have been banned from landfilling by
order of the State of California.

     Wastes selected from classified
groups of organic wastes have been
detoxified.  These groups include
cyanide wastes, phenolic wastes, sulfide
wastes, non-halogenated pesticides,
solvent still bottoms and general
organic wastewaters.  Wastes supplied by
generators in the South Vfest United
States are being tested.

     Preliminary autoclave Wet Air
Oxidation tests are run to insure
compatibility and treatability with the
full scale unit.  Operating conditions
are also established with these tests.
A continuous test of a minimum of eight
hours is completed with the full scale
unit.

-------
 PURPOSE

     The purpose of this test is to
 demonstrate Wet Air Oxidation of toxic
 and hazardous wastes at a full scale
 installation. "Data has been developed
 on a commercial scale, continuous unit.
 Wastes  tested were those actually
 produced by industrial clients in the
 South West portion of the United States.
 Results will be collated such that
 predictions might be made for all the
 compounds of particular classes.

 APPROACH

     The Zimpro Wet Air Oxidation unit
 (see Figure 1), processes aqueous wastes
 at a designed reactor temperature of
 280 C,  a designed reactor pressure of
 136 atmosphere (atm) and a liquid waste
 flow rate of 2.3 cubic meters per hour
 (nr/hr) (10 gpm).  Waste is pressurized,
 mixed with compressed air, and directed
 through the cold, heat-up side of the
 heat exchanger.  The incoming waste-air
 mixture exits from the heat-up side of
 the heat exchanger and enters the
 reactor v&ere exothermic reactions
 increase the temperature of the mixture
 to a desired value.  The waste-air
mixture exits the reactor and enters the
 hot, cool-down side of the heat
 exchanger and, after passage through the
 system  pressure control valves, is
 directed to the separator.  In the
 separator, the spent process vapors
 (non-condensible gases) are separated
 from the oxidized liquid phase and are
 directed into a two-stage water
 scrubber-carbon bed adsorber, vapor
 treatment system.

     In the Vfet Air Oxidation process,
 organic substances can be completely
 oxidized to yield highly oxygenated
 products and water.  For exanple,
organic carbon-hydrogen compounds can be
 oxidized to carbon dioxide and water,
while reduced oraanic  sulfur compounds
 (sulfides, mercaptans,  etc.J ana
 inqrqanic sulfides are easily oxidized.
It'should be noted that oxides of
nitrogen such as NO or N02  are not
formed  in Wet Air Oxidation because
reaction temperatures  are not suffi-
ciently high to form them.
     When incomplete oxidation of
organic substances occurs, the easily
oxidized sulfur and cyanide compounds
are usually still oxidized to sulfate
and carbon dioxide-ammonia provided a
sufficient degree of oxidation is
accomplished.  However, incomplete
oxidation of other organic compounds
results in the formation of low mole-
cular weight compounds such as acetal-
dehyde, acetone, and acetic acid.  These
low molecular weight compounds are
volatile and are distributed between the
process off-gas phase and the oxidized
liquid phase.  The concentration of
these low molecular weight compounds
(measured as total hydrocarbons (THC)
expressed as methane) in the process
off-gas is, dependent on their concen-
tration in the oxidized liquid phase,
which is determined by the degree of
oxidation accomplished; the waste being
oxidized; and the influent organic
concentration of the waste.

Autoclave Screening

     In this project, wastes were pre-
screened using autoclave (Photo 1) Wet
Air Oxidation tests.  These bench scale
tests generally demonstrate the level of
destruction or detoxification expected.
Corrosion tests were also conducted to
insure compatibility with the full scale
unit.

     Raw and oxidized samples were
analyzed for chemical oxygen demand
(COD), biological oxygen demand (BOD),
pH, total solids, ash, soluble chloride,
soluble fluoride and specific component,
e.g. cyanide, phenol, sulfide, pesticide
or solvent.  The condensible off-gas
from the autoclave was analyzed for
oxygen, nitrogen, carbon dioxide, total
hydro-carbon and methane.  The autoclave
and oxidized liquid effluent was
exanined for any residue which might
cause heat exchanger fouling or plugging
of the full scale unit.

Full Scale Testing

     Each waste was tested during a one
(1) day "demonstration period", to
determine the effectiveness of the Wet
Air Oxidation unit.  During each
"demonstration period", the sampling and
testing progran outlined asfollows was
used.
                                           10

-------
     Upon arrival of a truckload of a
screened and acceptable waste, a sample
of the waste was obtained and analyzed
for COD, pH, and specific component.

     During the Wet Air Oxidation of
each class of waste, liquid composites
of the influent raw waste and the
effluent oxidized waste were obtained.
Liquid samples of these two streams were
made on an hourly interval and the
liquid samples were composited over a 24
hour period or throughout the steady-
state operation of the Wet Air Oxidation
unit, if operated less than 24 hours per
day.  Split samples were made with one
sample going to the State laboratory and
the other going to Zimpro's laboratory
for analyses.  The raw waste and
oxidized waste were each analyzed for
COD, BOD, pH, total solids, ash, soluble
chloride, soluble fluoride and specific
component, e.g., cyanide, phenols, sul-
fide, chlorinated aliphatic compounds,
or non-halogenated pesticide.  In
addition, the oxidized waste were also
analyzed for dissolved organic carbon
(DOC) and a GCMS scan was made.

     Grab samples of off-gas from the
Wet Air Oxidation unit, sampled after
treatment but prior to discharge to the
atmosphere, were made in Tedlar gas
sampling bags.

     The sample of Wet Air Oxidation
process off-gas was analyzed for oxygen,
nitrogen, carbon dioxide, carbon
monoxide, total hydrocarbon, and
methane.  Gas sampling was conducted
when the Wet Air Oxidation unit was
operating at steady-state.  Two
replicate grab samples were obtained
each day that the waste were processed
in the Wet Air Oxidation unit during the
"demonstration period".  The samples
were analyzed by the State.

PROBLEMS ENCOUNTERED

     Several problems were encountered
during this project.  The most serious
were the delays in obtaining necessary
permits to operate the unit.  This
resulted in the obvious delays in
start-up.  Once running, it was
necessary to process the available
wastes being brought to the site.  The
owner did not wish to run specific
   wastes with the added sampling proce-
   dures, when wastes were available for
   commercial treatment.

        A major portion of the wastes pro-
   cessed has been plating wastes.  With
   the destruction of the cyanide and other
   organic complexing agents,  metal oxides
   and hydroxides tend to precipitate.
   Adjustments in the operating procedures
   and conditions were necessary to prevent
   excess fouling and solids deposition.

        All classes of wastes  to be tested
   have been actual wastes except for the
   pesticide waste.  After searching for
   over a year, it has been necessary to
   synthesize that waste.

   RESULTS

   Autoclave Screening Tests

        Table 1 lists results  of autoclave
   screening tests.  Corrosion tests showed
   that all wastes listed were shown to be
   compatible with the materials of
   construction of the full scale units.
   Most wastes rejected in the prescreening
   process were rejected due to a high
   corrosion rate.  Note that  prescreening
   of wastes through review of chemical
   analysis kept the number of unsuccessful
   autoclave screenings to a minimum.

   Full Scale Testing (3)

        Tables 2, 3, and 4 show the results
   for full scale testing.  Each table
   lists the results of a particular class
   of waste.  At the time of submission,
   complete off-gas data was not available
   but Tables 5, 6, and 7 show results for
   those tests completed.

        Destruction of cyanide was greater
   than 99.756.  Effluent cyanide was 82
   mg/1 with a feed of 25,390  mg/1.  The
   phenol waste oxidation resulted in
   99.77%, destruction with an effluent of
   36 mg/1 and a feed of 15,510 mg/1.
   Organic sulfur destruction  reached 94?
   with feed and effluent concentrations of
   3010 mg/1 and 180 mg/1.  In the general
   organics wastewater test dissolved
   organic carbon (DOC) destruction was
   96.7% with feed and effluent concen-
   trations of 20,830 mg/1 and 685 mg/1.
11

-------
      While  all waste'processing  runs
 have  been completed, complete analytical
 data  is not available at time of
 publication.

 REFERENCES

 1.  Canney, P.J.; and Schaefer,  P.T.;
    "Detoxification of Industrial
    Wastewaters by Wet Air Oxidation"
    1983 Spring National AIChE Meeting,
    Houston, March 1983

 2.  Chowdhury, A.K.; Wilhelmi, A.R.;
    "Treatment of Spent Caustic  Liquors
    by Wet Oxidation," 8th Annual
    Industrial Pollution Conference,
    Houston, June 1980.

 3.  Dietrich, M.J.; "Commercial
    Demonstration of Wet Air Oxidation
    of Hazardous Wastes" Interim Reports
    1, 2 and 3, October 1983.

 1.  Randall, T.L.; "Wet Oxidation of
    Toxic and Hazardous Compounds",
    Mid-Atlantic Industrial Waste
    Disposal Conference,  University of
    Delaware, 1981.

5.  Randall, T.L.; Knopp,  P.V.;
    "Detoxification of Specific Organic
    Substances by Wet Oxidation," 51st
    Annual Conference;  WPCF,  Anaheim,
    October 1973.
                                          12

-------
                       TABLE  1.   AUTOCLAVE TESTS - CASMALIA RESOURCES
                   Test Conditions - 280 C,  1  Hr.  Reactor Residence Time
Waste
Spent Caustic
Acid Distillate
Alkaline Solvent
Cyanide Plating
Metal Finishing
Cyanide
Cyanide Plating
Pesticide Rinsate
Cadmium Plating
Solvent Still
Bottoms
Rocket Fuel Waste
Mixture
Cyanide Wastewater
Pesticide Wastes
Pesticide
Parameter
COD, g/1
Total Phenols, mg/1
Sulfide, mg/1
COD, g/1
COD, g/1
Total Phenols, mg/1
COD, g/1
Cyanide, mg/1
COD, g/1
Cyanide, mg/1
COD, g/1
Cyanide, mg/1
COD, g/1
Organic N, mg/1
COD, g/1
Cyanide, mg/1
COD, g/1
BOD, 'g/1
BOD/COD
COD, g/1
COD, g/1
Cyanide, mg/1
COD, mg/1
BOD, mg/1
BOD/COD
COD, g/1
BOD, g/1
BOD/COD
Feed
40.0
5350
382
40.1
39.8
840
17.1
6910
40. 4
5090.0
40.2
20960.0
20.0
701.0
11.3
12980.0
43.9
21.3
.49
45.1
29.6
33160.0
1640.0
15.0
0.01
5.38
1.81
Effluent
7.5
66
< 1
5.8
7.5
20
3.2
120
6.0
36.0
8.6
234.0*
4.1
156.0
3.6
52.0
8.6
6.3-
.74
1.8
10.4
185.0
450.0
208.0
.46
.85
.725
% Removal
81.3
98.8
>99.7
85.5
81.2
97.6
81.3
98.3
85.1
99-3
78.6
98.9
79.5
77-7
68.1
99.6
80.4
70.4
96.0
64.9
99.4
72.6
84.2
53.0
* During continuous operation a full scale
  7 ppm cyanide or >99.96% destruction.
unit, pxidized product contained less than
                                          13

-------
                        TABLE 2.  WET AIR OXIDATION DEMONSTRATION
                          OF GULF OIL SPENT CAUSTIC WASTEWATER
 Wet Air Oxidation Conditions:
          Oxidation Temperature
          Nominal Residence Time
          Waste Flow Rate
          Compressed Air Flow Rate
          Reactor Pressure
          Residual Oxygen Concentration
              515 F (268C)
              113 min.
              5.3 GPM
              190 SCFM
              1610 PSIG
              3.7%
   Sample Description
COD, g/1
COD Reduction, %
Total Phenols, mg/1
Total Phenols Reduction, %
Total Sulfur, mg/1
Sulfate Sulfur, mg/1
"Organic  Sulfur, mg/1
Organic Sulfur Reduction, %
Sulfide Sulfur, mg/1
pH
Total Solids, g/1
Total Ash, g/1
Volatile Solids, g/1
DOC, mg/1
Soluble Chloride, mg/1
Soluble Fluoride, mg/1
Composite Influent
    Raw Waste	
      108.1
   15,510

     3580
      570
     3010
       13.0
       88.6
       57.1
       31.5

     1510
       4.4
Composite Effluent
  Oxidized Waste
        11.6
        89.3
        36
        99.77
      3090
      2910
       180
        94.0
        <1.0
         8.3
        59.7
        50.2
         9.5
      3680
      550
         1.3
               "Organic Sulfur - Total Sulfur Minus Sulfate Sulfur
                                          14

-------
                       TABLE 3.  WET AIR OXIDATION DEMONSTRATION .
                               CYANIDE WASTEWATER CLASS
Wet Air Oxidation Conditions:
         Oxidation Temperature
         Nominal Residence Time
         Waste Flow Rate
         Compressed Air Flow Rate
         Reactor Pressure
         Residual Oxygen Concentration
              495 F (275 C)
              80 min.
              7.5 GPM
              190 SCFM
              1220 PSIG
              7.1%
   Sample Description
COD, g/1
COD Reduction, %
Cyanide, mg/1
Cyanide Reduction, %
DOC, mg/1
DOC Reduction, %
PH
Total Solids, g/1
Total Ash, g/1
Volatile Solids, g/1
BOD,-, mg/1
Soluble Chloride, mg/1
Soluble Fluoride, mg/1
Composite Influent
    Raw Waste	
        37.4
    25,390

    14,710

        12.6
    '   135.3
       112.9
        22.4


       30
Composite Effluent
  Oxidized Waste
         4.2
        82
        99.7
      1710
        88.4
         9.0
        91.2
        77.4
        13-8
       603
       773
        29
                                          15

-------
                       TABLE 4. WET AIR OXIDATION DEMONSTRATION
                           GENERAL ORGANIC WASTEWATER CLASS
Wet Air Oxidation Conditions:
         Oxidation Temperature
         Nominal Residence Time
         Waste Flow Rate
         Compressed Air Flow Rate
         Reactor Pressure
         Residual Oxygen Concentration
              531 F (277O
              120 min.
              5.0 GPM
              190 SCFM
              1515 FSIG
              M.156
   Sample Description
COD, g/1
COD Reduction, %
DOC, mg/1
DOC Reduction, %
PH
Total Solids, g/1
Total Ash, g/1
Volatile Solids, g/1
BOD5> mg/1
BOD5/COD, g/g
Soluble Chloride, mg/1
Soluble Fluoride, mg/1
Composite Influent
    Raw Waste	
       76.0
   20,830

        1.9
       10.0
        0.7
        9.3
   29,680
        0.39
      212
        0.7
Composite Effluent
  Oxidized Waste
         2.5
        96.7
       685
        96.7
         3.3
         3.1
         1.1
         2.0
       880
         0.35
       866
        <0.5
                                          16

-------
         TABLE 5.  WET AIR OXIDATION DEMONSTRATION
       SPENT CAUSTIC WASTEWATER (PHENOL & SULFIDES)
                 Process Off-Gas Analysis
       Carbon Dioxide                   11.3%
       Oxygen                           H.3%
       Nitrogen                         83-256
       Carbon Monoxide                  Not Detected
       Methane                          12.1 ppm
       Total Hydrocarbons               85.4 ppm as methane
         TABLE 6.  WET AIR OXIDATION DEMONSTRATION
                 CYANIDE WASTEWATER CLASS
                 Process Off-Gas Analysis
Carbon Dioxide
Oxygen
Nitrogen
Carbon Monoxide
Methane
Total Hydrocarbons
1.55&
8.556
82.8?
Not Detected
9.0 ppm
61.1 ppm as methane
                            17

-------
                    TABLE 7.  WET AIR OXIDATION DEMONSTRATION
                       GENERAL ORGANIC WASTEWATER CLASS
                          Process Off-Gas Analysis
            Carbon Dioxide

            Oxygen

            Nitrogen

            Carbon Monixide

            Methane

            Total Hydrocarbons
              12.9%

              5.9%

              81.2%

              0.3%

              10.0 ppm

              29.1 ppm as methane
    FIGURE  I. WAO ELEMENTARY FLOW SHEET
               FEED HEAT
              EXCHANGER
   HIGH
PRESSURE
  PUMP
                                                  REACTOR
                                                        -START-UP STEAM
                                                        TO SCRUBBER
                                                        SEPARATOR
              AIR
         COMPRESSOR
COOLER
                                    18

-------
                         EVALUATION OF PILOT-SCALE  APCDs  FOR HAZARDOUS
                                   WASTE COMBUSTION SOURCES
                                        C..  W.  Westbrook
                                Research Triangle Institute
                       Research Triangle Park, North Carolina

                                         ABSTRACT
               27709
     Conventional  air pollution control  devices used at hazardous waste combustion facili-
ties may not adequately control fine particulate matter and some gaseous species.   This
project sought to identify alternate technologies that might provide a higher degree of
control.  The performance characteristics of several alternate technolgies is to be
tested at a full-scale incineration facility.  This paper is an interim report on the
project.
INTRODUCTION

    Combustion of hazardous waste compounds
can potentially generate particulate and
gaseous emissions that may have adverse
health effects on exposed populations.
Conventional control devices are of varying
effectiveness in controlling these emis-
sions.  Wet scrubbers are often used to
remove both particulate and gaseous
contaminants in the same device.  Wet
scrubbers, however, tend to remove large
particles more efficiently than they
remove small particles, and at high col-
lection efficiencies they require more
energy than do electrostatic precipitators
(ESPs) or fabric filters.  These performance
characteristics allow the escape of noxious
fine  particulate matter into the atmosphere
from  hazardous waste combustion sources.
    Advanced devices are expected to be
marketed with higher collection efficiency,
lower energy consumption, or lower cost
than  current conventional types of equip-
ment.  It is likely that some of these
devices are capable of providing better
emission control than is now available for
hazardous waste combustion products.
     In  Phase 'I of  this project, RTI
assessed the current state of development
of devices  that have the potential to
control emissions  from hazardous waste
combustion  sources and recommended a
number  for  field  testing.  Phase  II  of  the
project consists  of testing  several  of
 these units  on  a  slip-stream from  a
hazardous waste  incineration facility.
    Currently, Phase I of the project has
been completed.  Phase II, Field Testing
of the Units Selected, has just begun.
This paper describes the work completed to
date.

PURPOSE

    The purpose of this project is to
identify technologies that are potentially
superior to air pollution control devices
(APCDs) that are in current use for con-
trolling emissions from hazardous waste
combustion facilities and to test the per-
formance of these technologies on
commercial-scale combustion facilities.

APPROACH

    The project is being conducted in
several phases.  In Phase I, the currently
 used technologies and the potentially
superior technologies under development
were identified.  The potentially superior
technologies were identified by contacts
with, vendors,  research organizations,
literature reviews, and communication with
experts in the field.  Both technologies
that have been used on other combustion
sources, but not; widely used at hazardous
waste combustion facilities, and technolo-
gies at various stages of development from
bench-scale to pilot-scale were evaluated.
The primary factor considered in the eval-
uation was potential  applicability for
                                              19

-------
controlling submicron particulate matter
and/or HC1 emissions.  Secondary factors
considered were the potential for reduced
operating costs and reduced environmental
impacts of the discharge from the devices.
    After identification of the technolo-
gies that seemed to have the greatest
potential for controlling emissions from
hazardous waste combustion facilities,
efforts were undertaken, through direct
contact with the developer/marketer of
the technology, to ascertain if pilot
units of suitable size and construction
either existed or could be fabricated.
If the unit existed, discussions were
initiated to determine if the developer/
marketer would be interested in partici-
pating in testing of the unit at a hazard-
ous waste combustion facility.
    Concurrent with this latter activity,
discussions were initiated with several
commercial hazardous waste incineration
facilities to solicit their interest and
cooperation in allowing tests of the
selected pilot units at their facility.
Scheduling was coordinated with all
involved parties to pr-ovide sufficient
time for testing, yet not unduly inconven-
ience the host incineration facility.

PROBLEMS ENCOUNTERED

    Few problems were encountered in
Phase I of this project.  However, RTI
recognizes that some innovative devices
in current commercial use may not have
been identified.  A detailed investigation
of all facilities was not undertaken and
some facilities having custom-designed
equipment may have been unwilling to
discuss what might be considered competi-
tive advantages.  RTI also recognizes
that it is possible that not all innovative
new technologies were discovered in this
investigation.  Technologies that had not
been publicized or that were not reasonably
well-known to researchers in this field
Would have been overlooked.
    More significant problems were encoun-
tered in Phase II, Field Testing, that
had an impact on timely completion of the
project.  These difficulties arose both
in obtaining units to test and in finding
a suitable host plant at which to conduct
the tests.
    In Phase I, we had identified a
number of technologies meriting demonstra-
tion testing.  Our study required that
mobile pilot units having capacities of
1,000 to 4,000 cubic feet per minute of
stack gas be available.  These were not
available for all technologies.  Secondly,
the study would compare the performance
of all tested APCDs on the same source.
Some APCD developers were unwilling to
participate under these constraints,
perhaps because they did not feel their
test unit had yet achieved its potential.
Among the APCD developers who did wish to
participate, a few more mundane problems
arose.  These included negotiating the
financial arrangement and coordination of
test schedules.
    Delays were also encountered in
locating a suitable host plant for the
study.  Our first choice, who had expressed
a strong interest in participating,
encountered difficulties with regulatory
authorities and, after much delay, had
to withdraw.  An alternative facility was
sought which met the following criteria:
regulatory authorities will allow the test,
the facility is burning the appropriate
type of waste (contained significant
amounts of ash and chlorine), the facility
design is such that a slip-stream of gas
could safely be taken prior to the plant's
emission control system, the facility
management is willing to allow the testing,
acceptable schedules, and financial
arrangements could be made.
    As a result of the difficulties cited
above, a significant schedule delay was
encountered.  However, field testing of
the first pilot unit did commence in March
1984.

RESULTS

    In this section, we describe some of
the results of this project.  However,
since the field testing phase has just
begun, no test results can be provided.

Current Technologies

    As might be expected, the principal
generic types of APCDs in current use are
wet scrubbers, baghouses, and electro-
static precipitators.  Wet scrubbers,
which have the capability of controlling
both particulate and gaseous components,
predominate.  Fine particulate matter is
generally captured poorly, and many
facilities have a difficult time consis-
tently meeting regulatory requirements.

Advanced Devices

    A partial listing of the advanced
                                            20

-------
APCDs identified in Phase I is given in
Table 1.  The three principal types (ESPs,
wet scrubbers, dry scrubbers) are listed
by the company marketing the device.
Where appropriate, the type of device is
given.

     TABLE 1.  ADVANCED DEVICES
ESPs

  Beltran Inc. .and Micropul, Inc.
    (Wet Type)

Wet Scrubbers

  Lone Star Steel
    (Steam Hydro and Free Jet)
  Ceil cote, Inc.
    (Ionizing)
  University of Washington
    (Electrostatic Spray)

Dry Scrubbers

  Micropul/Koch
  Niro Atomizer, Inc.
  ETS, Inc.
  Rockwell International
    Wet ESPs have been around for sometime
but have not been widely used on hazardous
waste combustion facilities.   In these
devices, particulate collection is by
electrostatic forces.  Collected particu-
lates are removed with a water spray.     \
This reduces re-entrainment problems from
high resistivity.particles.
    The Lone Star Steel wet scrubbers
shown in Figure 1 use steam or air to
produce very finely atomized water droplets
which contact the particulate matter,
resulting in an increase in the apparent
diameter.  Removal can then be accom-
plished easily.with a cyclone separator.
    The Ceil cote ionizing jet scrubber
shown in Figure 2 precharges the particu-
late entering the scrubber.  Fine particu-
late adheres to the packing materials by
induced polarization and is removed by the
water spray.  The electrostatic spray
scrubber developed at the University of
Washington carries this concept one step
further by also applying a charge to the
water spray.
    The discharge from all of the dry
scrubbers is a dry powder.  Each differs
in the technology for introducing the gas
scrubbing medium, however.  Most of these
units atomize a caustic (lime or sodium1
carbonate) slurry into the gas stream and
use the heat content of the gas to evapo-
rate the moisture.  The ETS, Inc. unit
shown in Figure 3, however, injects a dry
powder (either lime or nacholite).

Devices Selected for Testing

    At this point, arrangements have been
made to test four pilot APCDs.  These are
the two versions of the Lone Star Steel
scrubbers, the Ceilcote, Inc. unit, the
ETS, Inc. unit, and the Hi-Tac unit devel-
oped by Vulcan Engineering Company.  Th
Hi-Tac unit, which has not been mentioned
previously, is a specially designed
baghouse that can operate at temperatures
up to 1300F.
    Each of these are mobile pilot units
that have capacities within the range of
1,000 to 4,000 cubic feet per minute of
stack gas.

Host Plant for Testing

    ENSCO, Inc. agreed to participate in
testing the selected pilot units at their
El Dorado, Arkansas incinerators complex.
At this facility, oils and capacitors
containing polychlorinated biphenyls
(PCBs) are incinerated.  Duct work has
been installed to remove a slip-stream of
gas from the system immediately upstream
of the emission control system.  Provision
has been made to cool the gas to the
temperatures required by the various pilot
units.

Test Methods

    Simultaneous inlet and outlet testing
is scheduled for each unit.  The testing
is comprised of particle size'determina-
tions (by impactor), total particulate
concentration (by EPA Method 5), and HC1
concentration (by adsorption is back half
of Method 5 train).   Three Method 5
particulate/HCl tests are conducted each
day.  One particle size determination per
day is made.  A 2-week test period is
allotted for each unit.

Test Results

    Field testing has just been completed
on the first pilot .unit (provided by Lone
Star Steel).  Preliminary data indicate
the unit achieved excellent removal  of
                                            21

-------
both fine particulate and HC1.   The  data,
hoy/ever, have  not been subjected to  a
thorough review at this time.   In our
opinion, it is inadvisable to  publish data
under these conditions.

ACKNOWLEDGMENTS

    This project is being conducted  under
EPA Contract No. 68-03-3149, Work Assign-
ments 12-1, 12-4, and 12-9. Harry Freeman,
lERL/Cincinnati, is the EPA Project  Monitor.
Engineering Science, Inc., is  conducting
the onsite testing under a subcontract to
RTI.
                           Steam or
                           Compressed Air
                             Water
                   Polluted Gas
Turbulent Mixing
Particulant Wetted
and
Vapors Absorbed
                                                     :'-Agglomeration
                                       Supersonic Ejector Nozzle
                        Figure 1.   Ejector Drive for Steam-Hydro Scrubber
                      Electrode -
                      Wires
                                           Spray Header     Packing
                  Grounded Plates
                                                                   >0ut
                         Figure  2.  Ceilcote Ionizing Wet Scrubber
                                          22

-------
  Flue Gas
Dry Reactant  ,-
                            Induced
                            Draft Fan
                             Baghouse
Dry Stable
Byproduct
                             Dry Stable
                             Byproduct
                                                  Controlled
                                                    Stack
   Figure 3.   ETS Dry Reactor
                   23

-------
                     DISPOSAL OF HAZARDOUS WASTE IN AGGREGATE KILNS

                                     James A. Peters
                              Environmental Sciences Center
                                    Monsanto Company
                                   St. Louis, MO 63167
                                           and
                                      Duane R. Day
                              Monsanto Research Corporation
                                    Dayton, OH 45407
                                           and
                                  Robert E. Mournighan
                              Incineration Research Branch
                           U.S. Environmental Protection Agency
                                  Cincinnati, OH 45268
                                        ABSTRACT

     Incineration of chlorinated liquid organic waste in an aggregate kiln was investiga-
ted in a one-week program at Florida Solite Company.  POHCs (toluene, tetrachloroethylene,
methyl ethyl ketone, and methyl isobutyl ketone) were monitored in the waste and stack
emissions.  In addition, stack emissions were monitored for particulate matter, particu-
late trace metals, HC1, S02, and NOX.  Process samples were collected and analyzed for
trace metals and chlorine.  The destruction and removal efficiency of POHCs and the fate
of trace metals and chlorine in the kiln process were determined.

     Consistent achievement of greater than 99.99% DRE was demonstrated for each POHC.
Emissions of other pollutants ranged as follows:  particulates - 4.4 to 6.5 kg/hr; HCL -
0.008 to 0.034 kg/hr; S02 - 72.2 to 99.6 kg/hr; NOX - 1.9 to 11.7 kg/hr.  Between 60 and
90% of the chlorine is fed to the kiln from the waste fuel and scrubber influent water.
Over 95% of the chlorine is removed from the kiln in the scrubber effluent water.
INTRODUCTION

     Cofiring of hazardous wastes in high
temperature industrial processes is an
attractive alternative to hazardous waste
incineration because it utilizes the
waste's heat content.  Many cofiring pro-
cesses, which include cement and dolomite
kilns, glass furnaces, steel furnaces, and
some industrial boilers, provide tempera-
tures and residence times similar to those
required for dedicated hazardous waste
incinerators.  In addition to the savings
derived from the heat value, the use of
existing industrial equipment does not
require the capital investment required for
a separate or new incinerator to dispose of
a hazardous waste, and it may provide an
environmentally acceptable alternative to
conventional hazardous waste land disposal
options.

     Aggregate kilns, because of their high
energy use, are an excellent example of
this concept.   Such kilns typically oper-
ate at temperatures over 1093C (2000F),
have gas residence times in excess of 1.5
seconds, and have a highly turbulent
                                            24

-------
combustion zone.   However, the need exists
for data that shows the environmental
effect of cofiring hazardous waste on  the
emissions and effluents from the aggregate
process.

     Currently, the State of Florida
Department of Environmental Regulation (DER),
in implementing RCRA regulations, allows
exemptions from hazardous waste incinera-
tor emission standards if the facility
uses the waste as a fuel for the purpose
of recdvering useable energy.  The Florida
Solite Corporation located in Green Cove
Springs, Florida, petitioned the Florida
DER for a change in their air pollution
permit to allow the burning of organic
wastes as supplemental kiln fuel, thereby
recovering energy from these wastes and
offsetting their use of pulverized coal.
A sampling and analysis program was con-
ducted at Florida Solite Corporation from
February 20-25, 1983, when burning coal
and hazardous waste as fuel.  Program re-
sults will allow the Florida DER to decide
whether the existing regulatory exemptions
are environmentally sound and they will
provide additional hazardous waste incin-
eration data for the EPA Industrial
Environmental Research Laboratory - Cincin-
nati.

     The sampling and analysis program  in-
cluded evaluation of:   (1) the destruction
and removal efficiency  (ORE) of principal
organic  hazardous constituents (POHCs)
when cofiring coal and waste fuel; (2)  the
concentrations of particulate matter, S02,
NOX, HC1, and metals  in stack emission,  and
(3) the  concentration  and  fate of metals
and chlorine in  the process  streams.

FACILITY AND PROCESS  DESCRIPTION

     The Florida Solite  Company operates
an  aggregate kiln  in  Green  Cove Springs,
Florida, which  is  located  approximately 20
miles  south of Jacksonville.  Annual  pro-
duction  of  the expanded lightweight inor-
ganic  material  used as  aggregate  in a ce-
ment mix is  approximately  5.45  x  107  kg
 (60,000 tons)  per year.

      This  industrial  process involves the
heating of clay to  a  temperature  of 2,000
 F to  2,100F in a horizontal  rotary kiln
to prepare an expanded  lightweight  inorgan-
ic material used as aggregate  in cement mix.

     The  kiln, with refractory linings, is
2.7 m(9 ft) in diameter and 45.7 m  (150 ft)
long.  The kiln rotates slowly (90  revolu-
tions  per hour) and has a gentle slope
(6.25  cm/m) to allow material  to pass
through by gravity.  The kiln  operates in
a counter current  flow  pattern; i.e., solid
materials travel in one direction and hot
gases  and dust travel in the opposite direc-
tion,  as  shown in  Figure 1.  Clay is fed
into the  kiln at the upper end at a rate of
approximately 12,260 kg/hr (27,000  Ib/hr).
At the opposite end of  the kiln, a  mixture
of coal and waste  fuel  is burned at rates
of approximately 700 kg/hr (1,540 Ib/hr)
and 0.87  m3/hr (230 gal/hr), respectively,
to provide a heat  input of approximately
22,000 kw (74 million Btu/hr).  As  the clay
feed travels down  the inclined rotating
kiln,  it  passes through various temperature
ranges which cause transformation of the
clay into the lightweight aggregate product
The lightweight aggregate is produced at a
rate of approximately 9,080 kg/hr (20,000
Ib/hr).   After heating  and transformation
in the kiln, the aggregate is  graded and
large  clumps are crushed for sizing.  The
final  product is stored in large piles
until  sold.

     The  kiln exhaust gases pass through a
pair of mechanical dust collectors, whose
dust is recycled into the kiln, then into
a horizontal cross-flow water  scrubber of
,fiber-reinforced-plastic (FRP) construc-
tion.  The series  of water sprays clean
the particulate matter  and reduces  the gas
temperature from about  370 to 70C (700CF
to 160F) before the gases reach the knock-
out chamber and fiberglass stack.   The
scrubber  discharge released from the knock-
out chamber is a mixture of raw steam and
water  with the entrapped particulate
matter.'   This discharge stream is released
to an  open ditch which  drains  to a  pond.
There  fs  no recycle of  the scrubber water.

     The  fuel used to fire the kiln is an
unblended combination of crushed coal and
waste  organic liquids.  The liquid  wastes,
which  are trucked  directly from the gener-
ators, consist primarily of solvents,
alcohols, ethers,  still bottoms, and a
                                             25

-------
small fraction of chlorinated hydrocarbons.
Any manifested wasteload that contains
pesticides, PCBs, acids, caustics, cyani-
des, sulfides, mercaptans, electroplating
wastes, or metal finishing wastes is re-
jected and returned to the generator.  The
organic waste mixture makes up from 50% to
100% of the fuel used.  During the test
period, the waste fuel comprised approxi-
mately 54% of the total heat input.

EXPERIMENTAL PROGRAM

       The sampling and analytical program
was designed to identify the major pollu-
tants from burning waste fuel in an aggre-
gate kiln, quantify their respective
emission rates, determine the destruction
and removal efficiency (ORE) of the POHCs,
and provide information fora mass balance
around the process for metals and chlorine.
Measured stack pollutants include POHCs
(toluene, tetrachloroethylene, methyl ethyl
ketone, and methyl isobutyl ketone),  par-
ti cul ate matter, particulate trace metals,
carbon dioxide, hydrogen chloride, sulfur
dioxide, and nitrogen oxides.  In addition,
the distribution of the metals and chlorine
were measured in all of the process input
and output streams; i.e., the coal feed,
waste feed, clay feed, scrubber influent
water, aggregate product, and scrubber
effluent water.  Waste fuel and coal sam-
ples were submitted for analyses of sulfur,
ash, and Btu content.  Waste fuel and
scrubber effluent water also were analyzed
for principal organics.  Table 1 summarizes
the overall test program and lists each
sampling and analytical method use.

       Sampling was conducted for five full
test days.  Because the kiln rarely oper-
ates on 100% coal fuel feed, the test was
conducted on normal fuel operation (i.e.,
coal and waste fuel) and no baseline (coal
only) testing was performed.  A Quality
Assurance Project Plan was reviewed and
approved prior to the test program.  A
full description of the QA/QC results
involving replicates, splits, blanks,
spikes, and reference standards is provided
in the final report.

RESULTS AND DISCUSSIONS

            WASTE FUEL

A detailed summary of the waste fuel
samples collected as shown in Table II.
Table III and IV show the concentration
of each POHC and other properties for the
five waste fuel samples (one sample per
day, Runs 1-5).

POHC Destruction and Removal Efficiencies

    The complex combustion, chemistry for
organic materials becomes perplexing when
a broad range of organic compounds present
in a liquid waste are burned.  On a weight
basis, most of the organic carbon in the
waste is oxidized to C02 in the combustion
process, but trace amounts of organic
chemicals survive the oxidation process
and are partially reacted.  Accordingly,
the test burn investigated the amount of
destruction of the organic compound in the
hazardous waste.

    Destruction and removal efficiency
(DRE) for an incineration/air pollution
control  system is defined by the following
equation:
    DRE = Uin - Mout  (100)
     W,
             Win
Where DRE = Destruction and removal
      w     efficiency, %
       in = Mass feed rate of principal
            organic hazardous constituent
            (s) (POHCs) fed to the incin-
            erator
     out = Mass emission rate of princi-
            pal organic hazardous
            constituent(s) (POHCs) to the
            atmosphere (as measured in
            stack prior to discharge).

    DRE calculations are based on combined
efficiencies of the destruction of the
POHC in the incinerator, or the aggregate
kiln in this case, and the removal of the
POHC from the gas stream in the air pol-
lution control system.  The presence of
POHCs in liquid/solid discharges from the
air pollution control devices is not
accounted for in the DRE calculation as
currently defined by the EPA.  RCRA Part
264, Subpart 0 regulations for hazardous
waste incinerators require a DRE of 99.99%
for all principal organic hazardous
constituents of a waste unless it can be
demonstrated that a higher or lower DRE
is more appropriate based on human health
                                            26

-------
                 INPUTS
                                                         OUTPUTS
        Coal

        Waste fuel

        Scrubber  influent water

        Clay  feed
Aggregate product

Scrubber effluent  water

Stack  gas
                                                                            STACK
                                                                            GASES
     COAL
WASTE FUEL
      AIR
            AGGREGATE
            PRODUCT
              WATER   WATER
              IN     OUT
         Figure  1.   Schematic  diagram of aggregate  kiln process,
 criteria.   Specification of the POHCs in a
 waste is subject to  best engineering judg-
 ment, considering the  toxicity, thermal
 stability,  and quantity of each organic
 waste constituent.   For the selected POHCs,
 toluene, MEK,  and Perc are RCRA Part 261,
 Appendix VIII  compounds; MIBK  is not list-
 ed.  ORE requirements  in the Subpart 0
 regulations do not apply to metals or other
 noncombustible materials.  Currently, ag-
 gregate kilns  are exempt from  trial burn
 testing.

        Toluene, MEK, and J>1IBK  were present
 in high concentration  for organic compounds
 (see Table II).  Perc  was selected because
 the chlorinated hydrocarbons are, in gen-
 eral, difficult to thermally destroy and
 Perc was the only chlorinated  organic
 measured in the top twenty constituents of
 the wastes fuel.

        The four POHCs  were sampled in the
 exhaust gas by the volatile  organic sam-
 pling train (VOST) and analyzed  by gas
 chromatography/mass spectrometry  (GC/MS).
 The number of acceptable VOST  runs made
 each day were as follows:   day 1  - 0  runs;
 day 2-6 runs; day 3 - 6  runs;  day 4 -  8
 runs; day 5 - 5 runs.   The average  ORE
 and  range obtained  is shown  in Table  V.
Figure 2 illustrates  the  frequency of
obtaining a certain  "number of nines" of
ORE for each POHC.

    Methyl  ethyl  ketone was destroyed and
removed to at least  99.99% efficiency.
Only three runs  showed DREs less than
99.999%:  Runs 4A, 3A3 and 3B3 were side-
by-side runs (with 3A and 3B) that were
split with the EPA QA contractor.  Runs
3A3 and 3B3 do not show good  comparison
with Runs 3A and 3B  for MEK,  possibly
because of blank contamination problems
in the QA contractor field blanks.  Run
3B3 is an outlier and is  not  considered a
significant part of  the data.  The overall
ORE average for  MEK  for all five days was
99.998%  0.006% (95% confidence limits).

    DREs for methyl  isobutyl  ketone (MIBK)
ranged from 99.986%  to >99.999%.  The
99.986% value was the only ORE less than
99.992%.  The overall average for MIBK
was 99.998%  0.006% (95% confidence
1 i mi ts).

    DREs for tetrachloroethylene (Perc)
ranged from 99.993%  to ^99.999%.  Excel-
lent consistency was found for each day
damp!ing.  Split samples  on  Day 3  (Runs
3A3, 3B3, and 3D1)  all showed low
                                            27

-------
         TABLE 1.   SUMMARY  OF FLORIDA SOLITE AGGREGATE KILN
                     SAMPLING AND  ANALYTICAL PROGRAM
          Parameter
      Sampling method"
                                                             Analytical  method
Stack Gas

 POHCs (tetrachloroethylene,
  toluene, MEK, MIBK)
Volatile organic sampling
  train (VOST)
GC/MS, thermal desorption
  and SIM
 Particulate matter
Metals on particulate
 Hydrogen chloride
 C02 and O2
 Nitrogen oxides
 Sulfur dioxide
Waste Fuel
Principal organics
Metals
Chlorine, sulfur
Btu content
Ash content
Scrubber Discharge8
POHCs
Metals
Lead
Hexavalent chromium
Chlorine
Aggregate Product
 Metals
 Chlorine
Clay Feed
 Metals
 Chlorine
Coal
 Metals
 Chlorine, sulfur
 Btu and ash content
Scrubber Influent
 Metals
 Chlorine
EPA 5
EPA 5
Impinger absorption in 0.5 M
NaoAc (back half of EPA 5)
EPA 3
EPA 7
EPA 6

Grab  composite
Grab > composite
Grab  composite
Grab -> composite
Grab * composite

Grab * composite
Grab  composite
Grab  composite
Grab  composite
Grab * composite

Grab  composite
Grab -> composite

Grab * composite
Grab > composite

Grab * composite
Grab  composite
Grab > composite

Grab  composite
Grab * composite
EPA 5
ICP
Specific ion electrode
Fyrite
EPA 7
EPA 6

GC/MS
ICP
XRF
ASTM D240-64
ASTM D482-IP4

GC/MS
ICP
AAS
APHA312B
XRF

ICP
XRF

ICP
XRF

ICP
XRF
ASTM D240-64

ICP
XRF
aThe scrubber discharge was split into sludge and supernatant fractions and was
 analyzed separately where applicable.
                                      28

-------
       TABLE II.   RESULTS OF  CAPILLARY GC/MS ANALYSIS
                  OF MAJOR COMPONENTS  OF WASTE FUELS
                  NUMBER 1 AND NUMBER  4

                                                    Concentration,
                                                        wt %
Waste fuel component
Ethanol
2-Propanol
1-Butanol
Ethyl acetate
Methyl ethyl ketone (POHC)
Methyl isobutyl ketone (POHC)
Toluene (POHC)
Tetrachloroethylene (POHC)
Ethylbenzene
Xylene (isomer No. 1)
Styrene
Xylene (isomer No. 2)
2-Ethoxyethyl acetate ,
Cs-Benzene (isomer No. 1),
Cs-'Benzene (isomer No. 2)
C10-Alkane (isomer),
Cn-Alkane (isomer)
n-Propyl acetate
2-Propanol, l-(2-methoxy-l-methylethoxy)-isomer No. 1
2-Propanol, l-(2-methoxy-l-methylethoxy) -isomer No. 2
2-Cyclohex 4-l-one,trimethyl (isomer)
Number
la
1.55
4.55
1.78
0.68
2.03
1.52
8.40
0.19
1.23
4.47
0.71
1.29
2.03
0.47
0.57
0.83
0.72
1.50
0.46
0.49
1.28
Number
4a
1.83
1.97
0.77
0.72
2.81
1.12
8.06
0.07
2.28
7.89
0.28
2.52
1.20
0.33
0.35
0.76
0.60
1.00
0.14
0.16
0.54
 Average of split sample.
 Compounds containing three carbons associated with a benzene ring.
cCompounds containing ten carbons associated with an alkane.
 Compounds containing eleven carbons associated with an alkane.
                                 29

-------






















tJ
M
5
fa

w
EH
VI
f^
15

a
n
w
u
a
CH


I-H
t *

w
LjJ
CQ
2
EH
























G
H
g
\
cn
..
-~^
C
H
3j_

0)
4J
rei

U)
CO
<0
g

rH
a)
3



<^ in ^^ r*^
in en CM co
CM H i-H
rH


O CO CO CO
O^ C"*1* ^J^ CO
CO CO O

CM

CM in rH P~
co in co en
CO CM CO
^
rH



in o co ^
CM en o in
   
^* CO O !*"

t-4 CM iO en
CO rH O en
CM rH O r--


CO rH r- rH
CO * i-H CM
 *  
rH H O CO


oo o  r~-
C-- f~ rH CM
   
rH rH O Cn


en co en co
en in r-j co
H rH O CO
0) OJ
C G
O (U
P -H
Q) rH >l
 ^i G
P *p
i-H 3 i A 0
10 (H
-P W O
 \
ft >H
W tP


K
f~{
w 5^
f^


^
OJ
rH f.
rd i 
>\
-P -P
ffl PC
0)



fc.
ffl g
o a
P4 a


.
M
3
4H ^S
3

^
(D
G
-H
M ^^
O
rH
f{
U

M
0)


c




O O O O O
r~ co m <> r~
vo en oo ^o CM
VO rH CO rH CO
rH CM rH rH rH





>.Q "4* ^ CO CO
in co f- co 10
<* in co CM co





<> CM CO vO
*o CM c** r^ ^0
cr en en o en
O O O rH O



'st* oo r^ co
r- CM * in rH
    
r~ c-~ c~- in vo
rH



o o o o o
m in <* co r~-
in ^j< r- in *a
CM rH CM Cf> CM
rH rH rH rH





It)
Q O Q Q O
J? rH Z z 2




rH rH CD VO CM
^ ^ CO CM *
O O O O O




co oo <* in in
o o o in in
 *   
rH rH rH O O






rH CM CO * in



































,
g
ft

rH


II

4J
H
g
H
rH

G
O
-H
P
U
0)
-P
a

(U
-P
u

-------
            Table V.  DESTRUCTION AND REMOVAL EFFICIENCIES  OF POHCs

DRE, %
Run
number
Day 2
Range
Average
POHC 1
(MEK)
99.999-
>99.999
99.999
POHC 2
(MIBK)
99.999-
>99.999
99.999
POHC 3
(Perc)
99.999-
>99.999
99.999
POHC 4
(Toluene)
99.999-
>99.999
99.999
             Day  3
             Range'

             Average

             Day  4
             Range

             Average

             Day  5
             Range

             Average
 99.968-
 99.999
 99.992
 99.998-
 99.999
 99.999
 99.999-
>99.999
 99.999
99.998-
99.999
99.999
99.986-
99.998
99.995
99.999-
99.999
99.999
99.998-
99.999
99.999
99.993-
99.998
99.997
99.991-
99.997
99.995
99.999-
99.999
99.999
99.995-
99.999
99.998
99.998-
99.999
99.999
Overall
Average
99.998 99.998 99.997
99.999

relative difference.  The overall DRE
average for Perc was 99.997%   0.004%
(95% confidence limits).

     DREs toluene ranged from 99.995% to
>99.999%.   The overall average  for toluene
was 99.999%  0.002% (95% confidence
limits), making toluene the easiest POHC
to destroy and remove.

        VOST S & A Split Results

     Because VOST sample tubes  are therm-
ally desorbed, analytical splits for QA/
QC purposes are extremely difficult and
typically not attempted.   This  disadvan-
tage of using a conventional  thermal
desorption technique is that once the
sample is desorbed into the analytical
system and is expended, there is no way
of reanalyzing the sample.  Therefore,
one cannot run duplicates or adjust for
scaling or instrument sensitivity prob-
lems if the sample is not in an appropri-
ate concentration for the analytical
conditions.
                        To overcome this analytical deficiency
                    and  provide an estimate  of VOST precision,
                    a test run was specified in the QA Plan in
                    which two VOSTs were sampled simultaneously
                    through the'sample port  with the probes
                    tied together.  While this was not a true
                    split, it nevertheless represented a whole
                    method (S&A) split rather than only an
                    analytical split.

                        Only one "within lab" split was speci-
                    fied; however, while in  the field, two
                    additional "between lab" split runs were
                    made and these samples were split with the
                    EPA  QA contractor who was conducting a
                    systems audit.  The "within lab" split at
                    MRC  shows excellent duplication and pre-
                    cision for all POHCs except MEK (relative
                    difference = 152%).  However, even a dif-
                    ference of 152% for MEK  on Runs 3D and 3D1
                    did  not affect the final  destruction effi-
                    ciency of 99.999% for each run, as shown
                    in Table VI.  VOST samples split between
                    labs (MRC and QA contractor) showed good
                    duplication for all POHCs except MEK.   In
                                          31

-------
  30 -
  20 -
  10 -
       Methyl ethyl ketone
            12345
               Number of Nines ORE
  30 -
  20 -
  10 -
       Methyl isobutyl ketone
                     _L
            1     2     3     4.5
               Number of Nines DRE
                                               (continued)

Figure 2.  Destruction and removal effeciencies.
                         32

-------
30 -
     Tetrachloroethylene
20 -
10 -
          12345
             Number of Nines DRE
30  -
     Toluene
 20  -
 10  -
           12345
              Number of Nines  DRE
             Figure 2 (continued)
                       33

-------

















ca
3
J3
en
JU
a
u
ex
<
>j
CQ
^J
CO

i
CO
IK
O
ca



S-|
J
<
<
p
g



55
M
J
o


03J
ca
u
K
0


,
H-l
"^

C

H









H
to v
10 3
 10
C1-IU4
"^ fl C
0,0
1 ^n 
o o
u z
ml































10
s
H
3
U
41


0)
Q

to

4J
H
*o




















41 O 
J4 01 O
2 S3
 1

S C
-I 10 CO
a, -
S-O n
S C 0
CO 10 O

a co
*M H B
O 10 r-
a e
 e E
O fl u
ca

u
O 0) 
M C
HE
O O.S
Z ca ca

ca
MJ 01
O CiH
10 C
-1 
OjQ n
Z 0


O)
jlative
fferenc
OS-H
O

3
1
-O
01 M
!i?M
Se
<*e
D




- CM
31
00
S 1-1

4J }^
p^
a e
10 -H i-l
K tfj



1
04



U
01
1
"i


CO
o o
01 03
A

Z.H
d
CO
ta
M
a
f-i
en
V.
in
-_-
M
ca


O
in

M

*

^
O





"*>
p>|




cn o in co
cMTfvacM co4*inio co in M co
in co co CM CQ coin co





r- 1-1 u3
s< vo iH CM m inm
^JCOCOCM CMn^in ^oesr^
rt -i o CM cnmocn t- CM o c
H iH < ^




CO CO CO
co *H c^ in a co in
in co ro o -i
O-IOC4 fHPIOtA CN^Or**
- IH


S S S
*.go ^gal ^ss
S>- aj o ut-*ft)o u*-(uo
C O* t^ X K CU E^ X X O* t^


0 TJ TJ
C - C rt  .^
fO Q (3 < fd CQ
n n fo
a < 03
co co co

* ~*
? II
egg
H 41 ft)
 ?, *,
4* 4J 4J
H J> t>
* J3 .Q
*^ H^ ^^
1 CM CO






to

a

ca
4H
15

i
0)
ca
n
i)
2
5

o
0
3
H
o
. !
M >
Bca
M
o c
n n
is 2
I.S-
< *J
01 u
So ca
H 
. c S JS
"So -o e
-< CO
< 5 -H

g CO (0
4-> A 4J
c 3 e
HO) 4J U
J? CI O
+3 o -PC
H CO
 o u
U l-l
^ g IM 0
in -H
co o CM ^
3  4->
M 14 -a ia
H M
t) 10 me
HO. -03
a* 01 a
a e 4J m
CO O CQ TJ
0) *H 10
a o . 10 -o
-H J-< 10 C C
ca -H 01 ki <3
i a 1 1 "
I "a n -rt <-i
O  J2 p
H -H CO 43 U
ca IM IH -H i
to ow * M
3 * O tJ 0
O +! il
o -t ^ o
C CO  ^ 33
 oi 4J a> 5
4J 'Q n M ^^
-* 3 4J
3 r-4 41 M
B 0 14 4 3

I/I 1-1 IM H p4
<0 J3 O tT
34

-------
this case, the DREs for MEK were affected
(I.e., Run 3A ORE = 99.999%, Run 3A3  ORE =
99.985%).  Causes for this difference in
MEK results are uncertain but may include
contamination of the tubes with MEK and
analysis performed by different labs.

            Stack Samples

     Results for particulate, hydrogen
chloride, sulfur dioxide, and nitrogen
oxides are summarized in Table VII.  Par-
ticulate emissions of 5.3 kg/hr (11.7 Ibs/
hr) were less than air permit regulations
for this site issued by the Florida Depart-
ment of Environmental Regulations (DER) of
8.82 kg/hr (19.43 Ib/hr).  The first S02
test result had a low value of 270 ppm and
is considered an outlier when compared to
the remaining seven S02 test results
which ranged 1,030 to 1,470 ppm.  The low
NOX value of 40 ppm was expected as it
occurred on Day 3 during start-up of the
kiln.
        Chlorine and Metals Balance

     As shown in Table VIII, the amount of
chlorine contributed by the waste fuel is
61-85% of the total chlorine to the kiln.
The scrubber influent water contributes
the remaining 15-39% of the chlorine to the
kiln.  Virtually all the chlorine leaving
the kiln does so in the scrubber effluent.

     Table VIII shows material balance for
selected metals.  The clay feed and waste
fuel contribute the largest percentage of
metals to the kiln process, while the
aggregate product and scrubber effluent
filtrate contain the majority of metals
leaving the kiln.  Difficulties in achiev-
ing good material balance closure include
the crude estimates for coal usage, clay
input, and product output rates (not
metered) and the analytical variability
inherent in ICP analyses.  The goal of
determining the major contributers of cer-
tain elements and their ultimate fate were
met, however.
            TABLE VII.  AVERAGE  RESULTS  FOR STACK GAS,  PARTICULATES,
                        HC1, S025 AND  NOX  EMISSIONS

Parameter and unit
Stack flow rate, m3/min
Stack moisture, %
Stack velocity, m/sec
Parti culates
mg/dscm
kg/hr
HC1 , ppm
SO 2, ppm
NOX, ppm
Range
623
21.5
16.6

163
4.4
0.15
270
40
- 673
- 28.8
- 17.1

- 273
- 6.5
- 0.68
- 1,470
- 227
Average
652
26.2
16.8

215
5.3
0.46
1,130
162
Standard
deviation
19
3.3
0.2

48
1.0
0.22
380
67

CONCLUSIONS

    The results of the program were as
fol1ows:

      Constant achievement of at least
      99.99% destruction and removal ef-
      feciency was demonstrated for each
      -POHC  (MEK, MIBK, Perc, and toluene)
      in the aggregate kiln process.

      Emission of pollutants were determined
      and ranged as follows:  particulates
      -  4.4 to 6.5 kg/hr; HC1  - 0.008 to
     0.034 kg/hr; S02 - 72.2 to 99.6 kg/hr;
     NOX.- 1.9 to 11.7 kg/hr.

     Approximately 60-90% of the chlorine
     is fed to the kiln from the waste
     fuel, while virtually all the chlor-
     ine exits the .kiln from the scrubber
     effluent water.

     The major percentage of any metal is
     fed to the kiln from the clay feed
     and waste fuel, while the major per-
     centage of any metal  exits the kiln  in
     the aggregate product and scrubber
     effluent filtrate.
                                            35

-------

















W
o

3
i 1
<
CQ
i_*j
*3
i_{
&
V}
EH
(rf
g
n
p*
O
J
8
J-H
^

U
<
EH
















-



i-H
(0 (1) (U
HUM
M C 3
0) <0 W 5-5
4-> rH O
(0 (tf iH
S XI U
rH -- M
(0 4-> X!

O O X)
EH rH


S-S

rH
H
K^

S
o

4H
M
H
OS
H

(U

(C
S

U CO
fd (d
p &
CO
M -P
0) C

XI -3
^J rH
M ^H
U s
c
rH
H
v|

O
4J
U
rH
It!
H
M
OJ
P
rd
S



fC Q)
rH (U
u MH

0 C
xi oj
XI 3
SrH
1 1 'j [
u d
CO-H

a)

W  F^ ^i
IO VO CT> rH 
rH



CM <* O 00 rH
    
00 **D CM *3* O
rH CM CM rH rH

rH rH rH rH rH
V V V V V




CT O O CT CTv
CT> O O CT* CT
H rH







rH rH rH rH rH
V V V V V



CTI in CTI c*^ oo
    
r- oo CM CM in
CM CO CM rH rH
rH rH H rH rH
V V V V V




in CT c^ co CTt
rH rH rH CO CO






in rH CO t~- rH
00 00 00 ^D >




rH rH rH rH rH
V V V V V



rH CM CO <# in
36

-------














(3
W
O

s
,_q
^
CQ

^
n
p*j
w
EH
S

<
EH
W
^
W
U
$
EH
.J
U
I-H
^
1 1
1 	 1
W
, 1
PQ
EH
























rH
re
4-J
O
EH
1
4-
C/J



C
H L,
M <1>
it
S-. Li
W-l O
W

c
u
OJ
CL,







rH
ID
4-J
O




C
i-H
H
O
*
4-






4-J
C
D
H
tw
4-1
(U




~H ^5*
CM n


o
**H rH r  1^)
co o OT i i ^ m

c--cvio<*onr~ocoo
CM t-i 1-1 m
r-H

V V V V V V





(nSincMSfoSScSS




CM i 1 CO CO ] ^ fo ^ . .^.
V i-t

'



'X>COi-l^}'\Dir)CO O CM CM C CM
'm
to m c^ r- r-i
0%ino^vDr-aioocnrHn




V V V V V V V




i | ff) i*-\ r^ CQ ^ | r-,j p^, jrt rQ
V*VCMCM rH OOUD


H^ffOCMrH^-l^rHr-li-l
V V V




f i re c) cr> LI o^ re -H xi C
rtJCQMUOSSSpHN









































Ul


OJ
V-j

O
tn

DI
re
(U
m
a)

re
in

37

-------
          EFFECTS  OF DISPOSAL OF HAZARDOUS WASTES  IN CEMENT KILNS

                    ON CONVENTIONAL POLLUTANT  EMISSIONS
                              James A. Peters
                       Environmental Sciences Center
                             Monsanto Company
                           St. Louis, MO  63167

                                    and

                           Robert E. Mournighan
                       Incineration Research Branch
                   U.S. Environmental Protection Agency
                           Cincinnati, OH  45268
INTRODUCTION

     To turn  the  old adage around,
"one man's  poison is another man's
meat" has  taken on  a  more literal
meaning as the RCR_A regulations for
hazardous  waste  incineration  have
become  promulgated   [1].   Because
many industrial waste  products can
be readily used as fuels or proces-
sed  into   supplemental  fuels,   a
waste fuel  market has  been rapidly
developing  in  the  U.S.   [2].   if
reprocessed  waste  liquids  do  not
contain  significant quantities  of
toxic metals, halogenated materi-
als,  or  any publically  sensitive
organics  (such  as PCBs),  they can
be   economically   substituted  for
coal, coke,  oil or  natural  gas  in
many  industrial   processes.   The
list of high temperature industrial
furnaces/processes  is   larger  than
one  may think,  and many  of these
already  burn  hazardous  waste  as
supplemental fuel at some facility:
cement  kilns  (both  wet  and  dry
process),  lime  and dolomite kilns,
fuller's   earth   (kitty   litter)
kilns,   steel   production   blast
furnaces, phosphate  rock  calciners
and  dryers,  iron  one dryers, brick
and tile tunnel kilns,  mineral wool
furnaces,  and glass melt furnaces.
     Cement  kilns,   in  particular,
have   become   the  process  heater
leader  in using  industrial  wastes
as  supplemental  fuel  for  several
reasons:   1)  the product (clinker)
cost is  extremely energy intensive
and any  fuel  cost savings gives an
economic/competitive  advantage;  2)
cement clinker quality is relative-
ly insensitive to addition of trace
impurities;  3)  cement  plants  are
typically  located near  population
and  industrial  centers;  4)  kiln
temperatures   and   gas   residence
times  are well  in  excess of  that
seen in most hazardous waste incin-
erators; and 5) regulations enacted
under  Subtitle C of RCRA provided
incentives for waste disposal  out-
side   of  conventional   hazardous
waste  incineration under the aegis
of energy recovery.

     The    Incineration   Research
Branch of EPA's  Industrial Environ-
mental  Research Laboratory-Cincin-
nati has has  been researching haz-
ardous waste incineration in indus-
trial  furnaces since 1977.   A num-
ber of test programs on waste in-
cineration  at  cement  plants  have
been conducted since 1974 in Cana-
da ,  the U.S.,  and  Western Europe.
The tests documented in the avail-
                                     38

-------
able literature include:
changes in emissions occur.
1.  St. Lawrence Cement
    Mississauga, Ontario [3,4]
      waste oil and high chlor-
      inated organic waste with
      No. 6 fuel oil a dry pro-
      cess kiln equipped with
      suspension preheaters,
      air pollution control by
      ESP.

2.  Stora Vika Test Center,'Stora
    Vika, Sweden [5,6]
      highly chlorinated organic
      waste, PCBs,  and Freon TF
      with coal in a wet process
      kiln with ESP.

3.  Marquette Cement, Oglesby,
    Illinois [7]
      low chlorine (<5%) or-
      ganic waste with coal in
      a dry process kiln with
      ESP.

4.  San Juan Cement, Dorado,
    Puerto Rico [8]
      highly chlorinated organic
      waste with Bunker C fuel
      oil in a wet process kiln
      with fabric filter.

5.  Rockwell Lime,  Manitowoc,
    Wisconsin [9]
      low chlorine organic
      waste with coke and
      natural gas in dry
      process dolomite kiln
      with fabric filter.

6.  Los Robles Cement, Lebec,
    California  [10]
      low chlorine organic
      waste with coal in
      a dry process kiln
      with fabric filter.

     This   paper  summarizes   the
available information on changes in
conventional  pollutant  emissions
when burning hazardous wastes.  The
example  test  programs  have  been
chosen because  testing was conduc-
ted  during   primary  fuel  firing
(baseline)  conditions  in  addition
to waste combined with primary fuel
firing conditions.   This comparison
testing allows the tools of statis-
tics to be used to evaluate whether
STATISTICAL   TESTING  FOR   CHANGE

     In  a  variety  of  situations,
decisions  must be  made  concerning
whether  a  new set  of  test results
represents  a   significant  change
when compared  to the old  or typi-
cal  test  results.   Sometimes  the
results of a change may be obvious.
In  other  cases,   however,   it  may
appear that a  change has occurred,
but it is not clear that the change
is enough  to  justify the claims or
problems  that  can  accompany  such
changes.   Decision-making  becomes
more difficult in such a situation,
and a  decision to  endorse  or  con-
demn a new process,  based on data
from small samples,  can  be  diffi-
cult.

     In  the context of this inves-
tigation, this means that one wants
to know  "if the  conventional  pol-
lutant emissions  from  a  kiln burn-
ing hazardous waste as supplemental
fuel are truly different from nor-
mal  emissions."  One  wants to  be
confident  that  the  results   are
correct  and  spend  no  more  money
than is  really necessary.   Statis-
tical  design  and interpretation is
the only way to meet this objective
[11] .   A  fairly simple  procedure
called   the   "one-sided   t-test"
enables  us to  decide,  to a subjec-
tively selected degree of certain-"
ty, whether  a statistically signi-
ficant  difference  exists  between
one set of data and another (assum-
ing  that  the  sample  populations
have approximately  normal  distri-
butions ).                       )   '

     If  one  takes  stack  samples
from the same stack ,on  two separ-
ate  occasions  and  analyzes  those
samples,  the measured concentration
will nearly always be different for
the two  samples.   There  are proba-
bly four principal  sources  of  this
variability:

   variability resulting from
    actual changes in pollutant
    concentration due to the ef-
    fect of two different opera-
    ting conditions (normal and
                                    39

-------
    waste firing)

    variability observed and inher-
    ent in a large-scale combustion
    process

    variability introduced as a re-
    sult of the sampling procedure

    variability in the analytical
    procedure

     To determine  whether a change
has occurred,  one  must subtract or
otherwise  take  into  account  the
variability  from  the process  and
sampling  and analysis,  and simply
compare the  actual concentrations.
Proper  application  of  the t-test
makes  this   distinction  possible
[11].

     All  that  one  needs  to use the
t-test is a computation of the mean
and standard deviation of each data
set, one-button features on modern
scientific  hand  calculators.   In
comparing two means, it is not nec-
essary that  they be made up of the
same number of  terms.   A  mean of
14 measurements made under  one test
condition can be compared with the
mean of  6 measurements  made under
another test condition.  As long as
the number of measurements, repre-
sented by nA and iig, are correctly
applied  in  the equations  and the
degrees   of  freedom  corresponding
to  n.  +  ng  - 2 are  used,  the re-
sulting  t-test will be  applicable.

     The  t-value is calculated us-
ing the following  equation:
        t =
              XA  -
              "B
where XA and Xg =
      n. and
      SA and SB
two separate means
number of values
from which XA and
Xj, are, respec-
tively, the means
standard devia-
tions of each
data set,
     This calculated t-value is com-
pared to a tabulated value of t at a
known probability  level,  usually 5%
or  1%  (equivalent  to  a 95%  or 99%
confidence level).   When the calcu-
lated t-value is less than the crit-
ical tabulated t-value, we say there
is no significant difference between
the means.  What we are really say-
ing is that with  the observed vari-
ation of the measurements  involved
we  cannot  tell  whether the observed
difference between  the means is due
to  some  real  cause  or  is  due  to
fluctuations  in  the  measurements.
If  t-calculated is greater  than t-
tabulated, then mean Xn is signifi-
                       -c\
cantly different than mean Xg.

     One  must always  remember that
statistical tests  are  based on pro-
bability,   consequently  there  is
some chance  that the  wrong conclu-
sion is  reached.   The  choice of the
probability  level  defines   the de-
gree of  chance  that the  wrong con-
clusion  was  reached.   For instance,
the  use  of  0.05  probability level
(95%  confidence  level) means  that
the  conclusion  reached from using
the t-test has  a  5% chance, or 1 in
20,  that  it  is  incorrect.   Con-
versely,  you  could give odds  of 20
to  1  that  the correct  conclusion
was reached.

PARTICULATE MATTER

     An  excellent  review  and dis-
cussion  of the  formation of partic-
ulate  emissions   when  chlorinated
wastes  are added  as  fuel  to a ce-
ment  kiln  is  provided  by  Weitzman
[12].   Data  from  the  St.  Lawrence
Cement  and  Stora  Vika tests  were
re-examined,  compared,  and plotted
as  a function of chlorine feed rate
with respect to  clinker production
rate,  which  clearly  demonstrate   a
linear increase in particulate
emissions as a  function of chlorine
input.   The  hydrogen chloride com-
bustion product reacts with the po-
tassium,  sodium,  and  calcium com-
pounds in the hot kiln to form the
respective  chlorides.   These vola-
tized chloride  salts are carried by
the hot  gas  to  the cold end of the
kiln where they  condense  at about
800-900C  to form  a fine particu-
                                     40

-------
late.   The character  of  this  new
particulate and resultant change in
dust resistivity  lowered the elec-
trostatic  precipitator  (ESP)  col-
lection  efficiency   and  produced
significantly  increased  dust emis-
sions  during  these two chlorinated
organic waste  firing tests.   Based
on his presumption of formation of
very fine  particles, he prognosti-
cated  that a  kiln using  a  fabric
filter  rather  than an ESP for par-
ticulate ,control  would  encounter
similar   changes   in   efficiency.

     Two  other tests  conducted at
kilns  equipped with  an ESPwaste
oil  at St.  Lawrence Cement and low
.chlorine  (<5%) waste  at Marquette
Cementdid not demonstrate an in-
crease  in  particulate  emissions.
Comparison  between  baseline  and
waste  firing  at  Marquette  Cement
showed no statistically significant
difference, an expected result when
one  examines   the  linear  relation-
ship   between   chlorine  input  and
particulate  emissions.   A statis-
tically significant difference was
seen,  however, for  the nonchlori-
nated  waste oil experiments at St.
Lawrence Cementa decrease in par-
ticulate  emissions.   This improve-
ment was  attributed  to  the addi-
tional water   introduced  (in  the
waste   oil),  which  been  shown to
improve cement kiln ESP  efficiency
even at gas  temperatures of 350C
 [3].

     The particulate emissions  from
the  kiln tested at San Juan Cement,
equipped  with a fabric  filter  dust
collection system,  did not exhibit
an increase with progressively  lar-
ger  chlorine   input  as   shown in
Figure 1.   The  linear  regression
curves from the St.  Lawrence  Cement
and  Stora Vika tests  are  also  dis-
played for comparison.   It can be
 seen  that,   not  only   was   more
 organic chlorine fed to the  kiln at
 San  Juan  Cement  than the previous
 highly chlorinated waste tests,  the
 fabric  filter system  was able  to
 maintain  a high  collection  effi-
 ciency  regardless  of the  changed
 character  of   dust  from  the  kiln.
 An  ancillary  part of the San  Juan
 Cement test program was Extraction
too
IS
II
                    CHLORINE FEED
               (KG CHLORINE METRIC TON CLINKER)

 Figure 1.   Variation of particulate
            emissions with  chlorine
            feed at  the three  high
            chlorine  feed  to  kiln
            tests.

 Procedure  (EP) Toxicity tests  on
 the baghouse  dust collected during
 waste firing.   Also,  the same dust
 was   extracted  and  analyzed  for
 polychlorinated  dibenzodioxin  and
 furan isomeric groups.  None of the
 compounds  tested  for  were found in
 these  analyses  and  the  baghouse
 dust  EP   Toxicity  tests  results
 matched  the  landfill  suitability
 results obtained  in a  much larger
 survey  of  cement kiln  dusts [13].

      Thus,  the addition of substan-
 tial amounts (>5%) of organic chlo-
 rine  in waste fuel  is  the  "bad
 actor"  in  lowering ESP collection
 efficiency, but no  drop in control
 effiof  fabric  filters will result.
 An ESP vendor could "fine-tune" the
 ESP to  overcome the change in dust
 resistivity as  long as  the  ESP is
 well maintained.

      Noteworthy in all the  tests
 described  above was that  particu-
 late concentrations were well below
 the  RCRA  performance   standard  of
 180  mg/m3   for a hazardous  waste
 incinerator.

 PARTICULATE LEAD

      The choicest waste streams for
 waste fuel use  in kilns are gener-
 ated  by paint,  ink,   and  coating
                                      41

-------
 users because of the high  Btu con-
 tent  of  solvents   and  alcohols.
 These waste  fuels  usually  contain
 suspended solids, with  a high con-
 centration of inorganic metals such
 as lead.  The  fate  of  lead  in the
 waste incineration  process  is  of
 particular  interest  because   the
 health effects are recognized,  the
 concentrations  in  the  waste  fuel
 are  often significant  (>1000  ppm
 Pb),and   implications  are  that  it
 may become widely dispersed  in the
 environment  [14].

      Comparison  measurements   of
 lead  emissions  were  conducted  at
 three of the kiln tests described:
 St.  Lawrence  Cement in  1974 burning
 a  waste  oil contaminated with 6,000
 ppm  (0.6  wt.  %)  lead; Marquette
 Cement burning waste solvents  con-
 taminated with 1,050  to 2,520  ppm
 lead;  and Rockwell  Lime  burning
 waste solvents   contaminated  with
 128 to 139 ppm  lead.   These  three
 tests provide distinctly different
 rates (almost  order-of-magnitude)
 of lead  input to  a kiln.

      Lead  emissions   at  Rockwell
 Lime increased slightly with waste
 firing to 1.7 g/hr,  and a statis-
 tically  significant  difference  was
 noted at the 95% confidence  level.
 The  measured  emission  rate   of
 1.7  0.25 g/hr (standard deviation
 included)  is  equivalent  to an auto-
 mobile  driven  at  30  mph   using
 leaded fuel.  An  automobile travel-
 ing at 60 mph  will  emit 3.54  g/hr
 and also  re-entrain  1.9   g/hr  of
 fugitive  lead  dust  from   a  paved
 highway  [15].

     At  Marguette Cement the  addi-
 tion of  waste  firing reduced  lead
 emissions  by almost  one-half,  but
this  was  not  a statistically signi-
ficant difference.  Apparently, the
substitution  of waste  fuel  for coal
caused the reduction in lead emit-
ted.   The lead emissions while burn-
ing  waste  fuel  contaminated  with
1,050  to  2,520  ppm lead is equiva-
lent  to  emissions from  ten automo-
biles  traveling 30 mph (neglecting
re-entrained  dust   contributions).

     A similar  reduction  in  lead
 emissions   of  almost   one-half  was
 measured  at  St.  Lawrence  Cement
 when  contaminated  waste  oil  was
 burned.    Because  more tests  were
 run   and    better   precision   was
 obtained  with-in  each   data  set
 condition,   a  statistically  signi-
 ficant  difference  was  noted  for
 these  data.   Again,   large  energy
 substitution  for  coal   may  have
 caused  the  lead  emission  reduc-
 tion.   The  waste  oil  firing emis-
 sion  rate  of  lead was  equivalent
 to   three  automobiles  traveling  at
 30  mph.

 CARBON MONOXIDE

     As   part  of   the  regulatory
 requirements   under   RCRA,   carbon
 monoxide  is  continuously  monitored
 in   the  exhaust  gas   of  hazardous
 waste  incinerators as  an  indicator
 of  combustion efficiency  (not  as  a
 surrogate    compound    for   DRE's).
 Moreover, the waste feed cutoff sys-
 tem is coupled to a maximum CO level
 specified in  each incinerator's Part
 B   permit,   established from  Trial
 Burn results.

     The   term  "combustion   effi-
 ciency"  is used as  a  measure  of  a
 combustion unit's performance.  Con-
 ceptually,  this  term  defines  the
 percentage  of  emissions  that  are
 completely  oxidized  to CO2,  and  is
 mathematically defined  as:
       % CE =
                COg
              CO., + CO
x 100
where  CO2  and  CO are  expressed  in
ppm by volume.   More rigorous defi-
nitions of combustion efficiency can
also include  denominator values for
THC  (expressed  in  ppm  volume  as
methane)  and Cp  (expressed  in ppm
volume  assuming  smoke  particulate
as  gaseous   elemental   carbon  and
ideal gas, 2.03 L/g).

     Cement production is a calcina-
tion process  in which large amounts
of CO2  are liberated from  the feed
to the exhaust  gas.   In fact,  the
CO2  from calcination  is more than
double  the  CO2  from  combustion and
confounds  a  combustion  efficiency
determination.
                                     42

-------
     CO  concentrations   in  cement
kilns  can  range  from  10 ppm  up to
occasional excursions in the percen-
tage range;  a  smooth kiln operation
is typically indicated by CO concen-
trations of 50 ppm or less.  Our ob-
servations  of  CO  from  kilns  have
convinced us that any process change
creates  significant CO  excursions,
independent of waste firing.    Some
examples which  temporairly increase
CO concentrations are changes in the
primary  air/fuel  ratio,   any change
in  clinker  cooler  air  (secondary
air),   irregular   solid   fuel  feed
rates,   and  changes  in exhaust damp-
er  setting.   These  CO  "spikes"  are
indistinguishable from the increased
CO observed for waste fuel introduc-
tion and shutoff  at San  Juan Cement
[16].  Thus, CO is more an indicator
of  process  stability in calcination
kilns  than  combustion  efficiency.

     At  two  other  kiln  testsLos
Robles  Cement  and  Rockwell  Lime
where CO was monitored continuously,
seemingly different results were ob-
served.  CO  concentrations at Rock-
well  Lime  fluctuated often during
both baseline  and waste  firing per-
iods such  that the mean concentra-
tions  were  similar and the standard
deviations large.  Naturally, the t-
test conclusion was that no signifi-
cant difference exists.

     On the  other hand,  the Los Ro-
bles Cement kiln demonstrated excel-
lent process  stability,  i.e. low CO
concentrations.   During  three waste
firing  test periods the  maximum CO
concentration  was  100 ppm,  a value
which  is easily exceeded due to any
process  change.   An  earlier series
of  baseline  tests  has a  CO maximum
of  618 ppm  with  daily  average CO
concentrations 2-6  times higher than
daily  averages  during waste firing.
Although  the   raw  data  were  not
available  for  statistical  testing,
the very narrow ranges of CO concen-
trations reported during waste fir-
ing suggest that a  t-test would show
that a significant  difference lower.
This would be  a misleading use of a
statistical  tool  given  the experi-
ences described above.
TOTAL HYDROCARBONS

     Total   hydrocarbon   analyzers
using  a  flame  ionization  detector
have   been  proposed   as  possible
regulatory   tool  for  testing  and
online   evaluation   of  incinerator
performance  [17].   The  evaluation
method  for hazardous waste inciner-
ation   involves   determining   the
total  unburned hydrocarbons  in the
combustion   gases   and  calculating
the   overall  destruction   of  the
hydrocarbons   in  the   feed.    The
theory  is that  an  incinerator pro-
ducing  significant  amounts of un-
burned hydrocarbons is not approach-
ing  thermodynamic  equilibrium  and
will  not pass the  overall  perform-
ance  test  of  99.99%  DRE  of  indi-
vidual compounds.

     This  .proposed monitoring tool
may  be  workable  for  incinerators
and  even  boilers  which  burn  waste
fuels,  but  testing  experience has
shown  that  kilns  do  not  exhibit
enough  change  in total hydrocarbon
(THC)  emissions.   The main  reason
is   that  the   primary  fuel  still
comprises  60%  or  greater  of  the
energy  input to  the kiln.  Unburned
carbons,   THC   emissions,   have   a
baseline   level   which  exists  for
this  fuel,  typically from  4  to  8
ppm  THC   (quantified  as  methane).
The  addition  of a  secondary  waste
fuel  simply does  not  cause  enough
change   in  THC  emissions,  which,
measured    continuously,    exhibit
very   smooth  concentration  curves
and  no  wide fluctuations.  A kiln
process  change  which  causes   a  CO
spike  goes  unnoticed  on  the  THC
analyzer.

     Results  from  the  Stora  Vika
tests  show that  THC  concentrations
were less  than 10 ppm for baseline,
chlorinated  aliphatics  fired,  and
Freon  TF fired,  but  rose  to 10 ppm
when   chlorinated  aromatics   were
burned.   The  data  were  not  in  a
form amenable  to statistical analy-
sis.   At San Juan Cement  a statis-
tically  significant  difference (and
increase)  was  observed when  chlor-
inated  aliphatics  were  burned  to
partially  offset fuel  oil,  presum-
ably  because  the  addition of high
                                     43

-------
levels  of  chlorine  made  complete
combustion   more   difficult.    At
Rockwell Lime  and Marquette Cement,
however,  a  significant  difference
(and  decrease)  was  observed.   At
both of these  tests,  a low chlorine
waste fuel was  burned at high ener-
gy  substitution   to  offset  solid
fuel  and  complete  combustion  was
probably enhanced.

NITROGEN OXIDES

     Nitrogen  oxides are  formed by
the  oxidation  of nitrogen  (N2) in
air or  nitrogen  compounds  in fuel.
Because nitrogen  has oxidation num-
bers ranging from  -3 to +5,  it is
possible  to  form  many  different
oxides.   Nitrogen oxides (NOx) con-
sist mainly  of NO and N02,  with a
typical ratio  of NO2 to  NO of 2:1.
Most current sampling and analytical
techniques rely on  oxidation  of NO
to  N02,  and  report concentrations
as NOx.

     In industrial  kilns  there is a
direct  functional  relationship  be-
tween NOx  emissions and secondary
air, which  is  heated air  from  the
product  cooler.   There  is  usually
little to no NOx  formed from nitro-
gen  in  fuel relative  to NOx  from
combustion  air.   Kilns  are  gener-
ally operated as  "poor"  combustors
when compared to  boilers; they have
a   reduced   primary  air/fuel  mix,
poor  secondary  air   mixing,   and
often  require  a  confined,  narrow
flame   configuration   for   product
quality.   Typical   NOx  concentra-
tions  in  the  flue  gas  can  range
from 200  to  1500 ppm,  even within
hours at the same kiln.  NOx levels
are very operator-dependent; contin-
uous monitoring  has shown  that at
some kilns   NOx " levels  vary widely
from hour  to  hour,   and  others  can
maintain steady NOx  concentrations.
The  absence  or  addition  of  waste
firing  has   little to  do  with  NOx
variations.   The  effect of excess
air  and  its  temperature  are  the
main   determinants   in  kiln  NOx
emissions.

     At  both  Marquette  Cement  and
San  Juan Cement  a significant  dif-
ference   (decrease)  was  observed
with  the  addition of  waste firing.
This  may  have  been  the result  of
lowered oxygen  input  more  than ad-
dition  of  waste  firing.   At  Los
Robles  Cement   the   NOx  emissions
decreased  slightly but  it  was not
statistically  significant.   A  sig-
nificant  increase was  observed  at
Rockwell Line,  but this was  due  to
operator intent to increase the NOx
emissions during waste firing.

     A definite relationship between
O2,   NOx,   and  SO2  in  calcination
kilns  has  been  shown  [18].   NOx
emissions  increase with increasing
O2  input  and degree  of preheating,
while  S02  emissions   decrease.   For
instance,   a  38%  reduction in NOx
due to  O2  lowering can  cause a 47%
increase in SO2.  Continuous  moni-
toring  at  Rockwell Lime illustrated
that  NOx  and SO2  change simultane-
ously,  as  shown in Figure 2.  More-
over,  the NOx/SO2 change is followed
within  5-15  minutes  by  a  CO spike,
the kiln's  indicator  of a momentary
process  instability.    As   shown  in
Figure 2,  this  event  was recurring.

ACID GASES

     Cement  kilns  are,  by  their
alkaline  nature,  built-in  control
devices for  acid gases  such as SO2
and HC1.   In a  well  operated cement
kiln  with  medium-to-high feed alka-
linity,   SO2   will   be   absorbed
  1000  1030   1100  1130  1ZOO  1300  1!30  1330  1400  KM
 Figure 2.  Illustration of inter-
            relationship between
            NOx, SO2, and CO emis-
            sions in a lime kiln.
                                     44

-------
(scrubbed)  in  the  kiln  environment
and  alkaline dust  of the APCD  to
control   efficiencies  of  greater
than 90%.   For  HC1 emissions  a re-
moval  of  efficiency  greater  than
99% was observed at San Juan Cement,
where  high chlorine  levels  were
added  to  a  relatively low alkaline
kiln.

     Waste  fuels  typically do not
contain much sulfur  and  thus the
partial substitution  of  waste fuel
for  fuel  oil or  coal will  reduce
sulfur  input to   the kiln.   This
will result in even less theoretical
SO2  to be  formed.   At  Marque tte
Cement, where a  13% energy substi-
tution  was  fired,  SO2  stack gas
concentrations  were  reduced  from
93  ppm to  18  ppm.   At  Los  Robles
Cement no significant difference in
SO2 emissions was noted.   SO2  emis-
sions  increased  significantly  at
San  Juan  Cement,   from  280 ppm to
450  ppm,  probably for two reasons:
the  NOx  emissions were  reduced
indicative  of  lower  O2  inputand
the high rate of chlorine input may
have caused a competitive  acid gas
"scrubbing"   situationHCl    was
preferentially  absorbed  over  S02.
The  kiln   operator's  choice  of O2
input  will  have  a greater  effect
on  SO2 emissions  than .addition of
a  waste  fuel.   It is possible to
operate  a  cement  kiln  with  less
than 50 ppm SO2  in the  stack gas.

     At Rockwell  Lime,  however,  a
different  mode  of  operation  was
required;   SO2 emissions  are inten-
tionally  not absorbed because the
presence of sulfur in lime is unde-
sirable.  No significant difference
was  observed  for  SO2   emissions,
which  remained  between 500-600 ppm
regardless  of waste firing.

CONCLUSIONS

     The  foregoing technical  exam-
ples  of  burning  liquid  hazardous
wastes  as   a  fuel   supplement  in
cement and  lime  kilns  in Canada,
Sweden,   Illinois,   Puerto   Rico,
California,   and   Wisconsin   have
demonstrated  that,   often  times,
emissions   of  conventional  pollu-
tants  are reduced and that process
operation   changes  have  a greater
effect on changes in emissions than
the  partial substitution  of waste
as  fuel.    It  is a tribute  to the
expertise of  the operators at each
plant tested that they were able to
make  product  while  burning  a new
and  different   fuel  mix.   A  new
hazardous waste  incinerator is per-
mitted  720  hours  of  "shakedown"
operation  before  a  trial burn is
conducted;  these  test  situations
were  not allowed  that  luxury and
the  operator's  skill  in  burning
waste  fuels   could  only  improve
with practice.

     Before  kilns  are  allowed to
routinely   burn   hazardous   waste
liquids,  monitoring devices capable
of  assuring the environmental ade-
quacy  of  the  burning  operation
should be  identified.   The  amount
of  information  now  available con-
cerning process  variables is insuf-
ficient  to  identify  a  single ade-
quate monitoring system.  Equipment
capable  of  detecting trace amounts
of  unburned hydrocarbons have been
shown  to be  inappropriate because
of  the  predominance  of  .unburned
hydrocarbons from the primary fuel.
The  emission tests which  have in-
cluded  continuous  monitoring  for
SO2,  NOx,   O2,  and CO  have  demon-
strated  that  combustion operation
can be (with training) easily opti-
mized.   Addition of these monitors
is  an excellent operating tool be-
cause of the increased feedback and
decreased  response lag  times com-
pared  to air flow, temperator, and
undergrate  pressure  sensors,  but
such   continuous  monitoring   (and
subsequent  reporting)  is excessive
as a regulatory  tool.   They require
constant  attention,   calibration,
and   maintenance.    Unfortunately,
the combination  of monitors is val-
uable and no  single  monitor  serves
the purpose.

     Further   research   should  be
conducted to  identify the simplest
monitoring  techniques  that  can be
used  to  effectively  ensure  the
safety of waste firing.   The iden-
tification  of  adequate  monitoring
systems  entails a thorough  under-
standing of the mechanism by which
waste destruction  occurs in  kilns.
                                     45

-------
REFERENCES
     Worthy,  W.   "Hazardous  waste:
     treatment  technology  grows."
     Chemical  &  Engineering  News,
     60  (10):  10-16,  March 1982.

     Hazelwood, D.  L.,  F.  J.  Smith
     and E.  M. Gartner.   "Assess-
     ment  of  Waste  Fuel  Use  in
     Cement Kilns."  EPA-600/S2-82-
     013, U.  S.  Environmental Pro-
     tection  'Agency,   Cincinnati,
     Ohio.   October 1982.

     Berry,  E. E., L. P. MacDonald,
     and D.  J. Skinner.   "Experi-
     mental Burning of Waste Oil as
     a Fuel  in Cement  Manufactur-
     ing."   Environment Canada, EPS
     4-WP-75-1.  June 1976.

     MacDonald,   L.  P.,   D.   J.
     Skinner,  F.   J.  Hopton,  and
     G. H.  Thomas.   "Burning Waste
     Chlorinated  Hydrocarbons  in a
     Cement  Kiln."   Fisheries  and
     Environment  Canada,  EPS  4-WP-
     77-2  (also  EPA/530/SW-147C).
     March 1977.

     Ahling,  B.   "Combustion  Test
     With Chlorinated  Hydrocarbons
     in a Cement Kiln at Stora Vika
     Test Center."   Swedish  Water
     and Air Pollution Research In-
     stitute,   Stockholm,   Sweden.
     March 1978.

     Ahling,  B.    "Destruction  of
     Chlorinated  Hydrocarbons  in  a
     Cement   Kiln."   Environmental
     Science  &  Technology,  13  (11):
     1377-1379, November 1979.

     Higgins,  G.   M.,   and  A.  J.
     Helmstetter.     "Evaluation  of
     hazardous  waste   incineration
     in  a dry process  cement kiln."
     Proceedings of the Eight Annual
     Research  Symposium:   Incinera-
     tion and Treatment of Hazardous
     Waste.   EPA-600/9-83-003. April
     1983.   pp. 243-252.

     Peters,  J.   A.,  T. W.  Hughes,
     J.  R.   McKendree,   L.  A.  Cox,
     and B.  M.  Hughes.   "Industrial
     Kilns    Processing .  Hazardous
     Wastes-San Juan Cement Company
     Demonstration  Program."   EPA-
     600/S2-84-, U. S. Environmental
     Protection  Agency,  Cincinnati,
     Ohio.

 9.  Day,  D.  R.,   and  L.  A.  Cox.
     "Evaluation of Hazardous Waste
     Incinceration  in Lime Kilns at
     Rockwell  Lime  Company."   Mon-
     santo    Research   Corporation
     draft  report  on EPA  Contract
     68-03-3025,    Work    Directive
     SDM-05.  October 1983.

10.  Jenkins,  H.  C.,  G.  Murchison,
     R.  C.  Adrian,  R. D.  Fletcher,
     and D. C. Simeroth.   "Emissions
     from  Los Robles  Kiln."   State
     of   California  Air  Resources
     Board,  Engineering  Evaluation
     Report  C-82-080.   August 1983.

11.  Duvel, W.  A.  "Practical inter-
     pretation  of  groundwater moni-
     toring results."  In:  Proceed-
     ings of the National Conference
     on  Management  of  Uncontrolled
     Hazardous Waste Sites.  Hazard-
     ous  Material   Control  Research
     Institute, Silver Spring, Mary-
     land, 1982. pp. 86-90.

12.  Weitzman,  L.  "Cement kilns  as
     hazardous  waste incinerators."
     Environmental  Progress,  2  (1):
     10-14, February 1983.

13.  Haynes, B. W. and G. W. Kramer.
     "Characterization    .'of    U.S.
     Cement  Kiln Dust."   Bureau  of
     Mines  1C 8885, Avondale, Mary-
     land.  1982.

14.  Mix,  T.  W. and B.   L.  Murphy.
     "Risks  associated with  waste-
     fuel  use   in  cement  kilns."
     Environmental  Progress,  3  (1):
     64-70, February 1974.

15.  Compilation  of  Air  Pollutant
     Emission  Factors, "Third  Edi-
     tion.  U. S. Environmental  Pro-
     tection  Agency,  Research  Tri-
     angle  Park,  N.C.  Publication
     No. AP-42, July 1979.

16.  Peters,   J.  A.,  T.  W.  Hughes,
     and  R.  E. Mournighan  "Evalua-
     tion of  hazardous waste  incin-
                                    46

-------
     eration  in  a cement  kiln . at
     San  Juan  Cement."   Presented
     at  the  Ninth Annual  Research
     Symposium  on  Land  Disposal,
     Incineration  and Treatment  of
     Hazardous Waste,  May 2-4, 1983,
     Ft. Mitchell, Kentucky.

17.  Monroe, E. S. "Quicker, cheaper
     testing  of  incinerator perfor-
     mance."   Chemical  Engineering,
     90  (4):   69-71,  February 1983.

18.  Tidona,  R.   J. ,  W. A.  Carter,
     H.  J.  Buening,  S. S.  Cherry,
     and M.  N. Mansour.   "Evaluation
     of  Combustion  Variable Effects
     on  NOx Emissions  from Mineral
     Kilns."      EPA-600/S7-83-045.
     U.S.  Environmental  Protection
     Agency, Research Triangle Park,
     N.C.  November 1983.
                                     47

-------
                     EVALUATION OF HAZARDOUS WASTE
                      INCINERATION IN A LIME KILN
                              Duane R. Day
                            Monsanto Company
                              Dayton,  Ohio

                                   and

                          Robert E. Mournighan
                  U.S. Environmental Protection Agency
                            C inc innati,  Ohio
                                 ABSTRACT
      During a one-week test burn, hazardous waste was used as supple-
 mental fuel and co-fired with petroleum coke in a lime kiln in eastern
 Wisconsin.  Detailed sampling and analysis was conducted on the stack
 gas for principal organic hazardous constituents (POHCs), particulates,
 particulate metals,  HC1, SO2, NOX, CO,  and THC and on process
 streams for metals and chlorine.   POHCs were also analyzed in the waste
 fuel.  Sampling was  conducted during three baseline and five waste fuel
 test burn days.  The program objectives were to determine the destruc-
 tion and removal efficiency (DRE) for each POHC, determine concentration
 of stack gas pollutants under baseline  and waste fuel test burn condi-
 tions, determine the fate of chlorine,  sulfur, and trace metals in the
 kiln process, and evaluate kiln performance when operating with
 hazardous waste as supplemental fuel.

      Results show average DRE's greater than 99.99% for each POHC and
 little change in pollutant emissions from baseline to waste fuel test
 conditions.  In addition, material balance results show that 95% of
 chlorine enters the  process from  the limestone feed and the chlorine
 exits the kiln in the baghouse dust and lime product at 61% and 38%,
 respectively.
INTRODUCTION

     Cofiring  of hazardous  wastes
in   high  temperature   industrial
processes is an attractive alterna-
tive  to  hazardous  waste incinera-
tion.  The alternative makes use of
the  waste' s  heat content  and pro-
vides  temperatures  and  residence
times similar to those required for
incinerators dedicated to incinera-
tion of hazardous wastes.  In addi-
tion  to  the  savings  derived from
the heat value, the use of existing
industrial  equipment does  not re-
quire  the  capital  required  if  a
separate  incinerator to  process  a
given  amount of hazardous waste is
to be built,  and  it may provide an
environmentally acceptable alterna-
tive   to    conventional   hazardous
waste disposal.

     Lime  kilns,  because of their
high enengy use,   are  an excellent
example  of  this  concept.    Such
                                    48

-------
kilns typically operate' at tempera-
tures  over  1093C  (2000F),  have
gas  residence times  in  excess  of
1.5  seconds,  and  have  a  highly
turbulent  combustion  zone.   How-
ever,  the  need  exists  for  data
that  shows  the effect  of cofiring
hazardous  waste  on the  emissions
from the lime process.

     The State of Wisconsin Depart-
ment of Natural Resources (DNR) and
EPA  Region V,  in  implementing RCRA
regulations,    issued  a  temporary
permit to Rockwell Lime Company' to
conduct  a   hazardous   waste  test
burn.  This   test would  allow the
burning  of  hazardous  liquid waste
as  supplemental  fuel  along  with
petroleum  coke.   The  waste  fuel
would replace natural gas as a fuel
component.

     Monsanto Company (MC) was con-
tracted  by  the U.S.  Environmental
Protection    Agency,    Office   of
Research and Development  (ORD-IERL)
in   Cincinnati,   Ohio   to  perform
sampling  and  analysis  of  stack
gases  and. process  samples  during
the  test  burn  conducted  at  the
Rockwell Lime Company  in Rockwood,
Wisconsin on April 15  and April 29
through  May  6,  1983 under baseline
(no  waste fuel  burned)  and waste
fuel  conditions   (coke   and  waste
fuel burned).

     The primary  objectives of the
sampling  and  analytical  program
were  to  (1)  determine  the  effects
of   cofiring  petroleum  coke  and
hazardous  waste  on the  emissions
from  the  kiln,   (2) determine the
fate   of  the  principal  organic
hazardous constituents  (POHCs) and
determine  destruction  and  removal
effeciencies   (DRE),  (3) determine
the  fate  of chlorine   and  trace
metals   in   the   kiln   process,
(4) determine  the   concentration
of   SO2,  NO  ,  particulates,  HCl,
            X
metals,   total  hydrocarbons,  and
carbon monoxide in the  stack  gas at
baseline and waste fuel  test burn
conditions,   and  (5) evaluate kiln
operation  during  hazardous  waste
fuel burning condition.   This  test-
ing  would provide to the Wisconsin
DNR  and the  EPA  Region V the data
necessary  to  determine whether  a
permit  can be  issued  to  Rockewll
Lime  Company   to  burn  hazardous
waste.  The  testing also will pro-
vide  the  EPA-ORD  with additional
data  in their  research  on  the in-
cineration of  hazardous waste and
the environmental  problems  associ-
ated with incineration.

FACILITY AND 'PROCESS DESCRIPTION

     The   Rockwell  Lime   Company
operates  a lime  kiln  in Rockwood,
Wisconsin,   which    is   located
approximately  10  miles  north  of
Manitowac  in   eastern  Wisconsin.
Production of  lime, which  is con-
verted  into  various forms  such as
hydrated and barn lime,  is approxi-
mately   1.3  x  106  kg   (1,430 tons)
per week,  which  varies based upon
product demand.

     The   industrial  process  in-
volves  the heating of limestone to
a  'temperature   of    approximatly
1,093C  (2,000F)   in  a horizontal
rotary  kiln to  prepare the lime
product.  The calcining  is achieved
by  interfacing the  hot gases with
the limestone  which drives  off the
CO2 from the limestone,  leaving the
lime product (CaO).

     The kiln,  with refractory lin-
ings,  is  2.4 m  (8  ft) in diameter
and 67.1 m (220 ft)  long.  The kiln
rotates  slowly  (approximately one
revolution  per minute) and  has  a
gentle  slope (6 cm/m of length) to
allow  material to  pass through by
gravity.   The,  kiln  also-  has  a
countercurrent  flow pattern,  i.e.,
solid   materials    travel   in  one
direction  and  hot  gases  and dust
emissions  travel  in  the  opposite
direction  as  shown   in Figure 1.
The coke  and natural gas provide a
heat  input of approximately  14,700
kw  (50  million Btu/hr)  or approxi-
mately  6.5 million Btu/ton of lime
product.   As  the   limestone  feed
travels  down the inclined rotating
kiln,   it  passes   through  various
temperature  zones,  and  the  hot
gases  calcine  the  limestone into
the  lime product.   The product is
produced  at  a  rate of  approximate-
ly    7,720 kg/hr    (17,000 Ib/hr).
                                     49

-------
After  heating  and  transformation
in  the kiln,  the lime  product is
cooled by  air  and either stored in
silos  or hydrated prior to storage
in  silos   where . it  is  held  until
sold.
                         3

                         e
                                 co
                                 en
                                 
-------
                TABLE  1.   SUMMARY OF  ROCKWELL LIME KILN
                            SAMPLING AND ANALYTICAL PROGRAM
       Parameter
     Sampling method
                        Analytical method
STACK GAS

 1.   POHCs*
 2.  Particulate matter
     Metals on
       particulate

 3.  Hydrogen chloride
 4.  Carbon dioxide
       and oxygen

 5.  Nitrogen oxides
 6.  Sulfur dioxide


 7.  Carbon monoxide

 8.  Total hydrocarbons

 9.  Oxygen

10.  Waste fuel
     POHCs
     Metals
     Chlorine,  sulfur
     Btu content

11.  Baghouse dust
     Metals
     Chlorine,  sulfur

12.  Lime product
     Metals
     Chlorine,  sulfur

13.  Dry limestone feed
     Metals
     Chlorine

14.  Primary fuel coke
     Metals
     Chlorine,  sulfur
     Btu content
Volatile organic sampling  GC/MS, thermal desorption
  train (VOST)               and SIM
EPA Method 5
EPA Method 5
                    EPA Method 5
                    ICP
Impinger absorption in     Specific ion electrode
  0.5 M NaoAc (back half
  of EPA Method 5)
EPA Method 3


Continuous


Continuous


Continuous

Continuous

Continuous
Grab  composite
Grab - composite
Grab * composite
Grab * composite
Grab.- composite
Grab - composite
Grab
Grab
composite
composite
Grab - composite
Grab * composite
Grab -> composite
Grab - composite
Grab * composite
                    Fyrite


                    Chemiluminescence photo-
                      metric analyzer

                    Pulsed fluorescence
                      TECO analyzer

                    Infrared-EPA Method 10

                    Flame ionization detector

                    Teledyne's micro-fuel cell
                    GC/MS
                    ICP
                    ASTM D240-64
                    ASTM D482-IP4
                    ICP
                    XRF
ICP
XRF
                    ICP
                    XRF
                    ICP
                    XRF
                    ASTM D240-64
 Tetrachloroethylene,  trichloroethylene, methylene chloride,
 1,1,1-trichloroethane,  methyl ethyl ketone,  and toluene.
                                       51

-------
 RESULTS AND DISCUSSION

 Waste  Fuel

     A_ summary  of the  waste  fuel
 composition   for  two  samples   is
 shown  in Table  2.   Tables 3 and 4
 show the concentration of  each  POHC
 and  other properties  averaged  for
 the  five waste   fuel  samples  (one
 sample per day,  Runs 4-8).

 POHC Destruction and
 Removal Effeciencies

     The complex combustion  chemis-
 try  for organic materials  becomes
 perplexing  when  a broad  range  of
 organic   compounds  present  in  a
 liquid waste  are  burned.   On  a
 weight basis,  most  of  the  organic
 carbon in the  waste is oxidized  to
 CO2 in the  combustion process, but
 trace  amounts  of organic  chemicals
 survive  the  oxidation process and
 are    only    partially    reacted.
Accordingly,  the test burn investi-
 gated  the amount of destruction  of
 the organic  compounds in  the haz-
 ardous waste.

     Destruction and  removal effi-
ciency t (DRE)  for  an incineration/
 air pollution control system is de-
 fined  by the  following   equation:
      DRE = Win " Wout
         (100)
(1)
where   DRE =
        Win =
       W
        out
destruction and
removal efficiency, %
mass feed rate of
principal organic haz-
ardous constituent(s)
(POHCs) fed to the
incinerator
mass emission rate of
principal organic haz-
ardous constituent^ s)
(POHCs) to the atmos-
phere (as measured in
stack prior to
discharge).
      DRE calculations  are based on
 combined  efficiencies  of the  de-
 struction of the POHC in the incin-
 erator, or  the lime kiln,  and the
removal  of the  POHC  from  the  gas
stream in the air pollution control
system.   The  presence of 'POHCs  in
solid  discharges  from the  air pol-
lution   control   devices   is   not
accounted  for in the  DRE  calcula-
tion  as   currently  defined by  the
EPA.  RCRA Part 264, Subpart 0 reg-
ulations for hazardous waste incin-
erators require a DRE of 99.99% for
all  principal   organic   hazardous
constituents  of   a  waste  during
trial burns unless it can be demon-
strated that  a higher or lower DRE
is  more  appropriate based  on human
health criteria.   Specification  of
the POHCs  in  a waste  is  subject to
best engineering  judgment,  consid-
ering the toxicity,  thermal stabil-
ity,  and quantity  of  each organic
waste constituent.

     Approximately  six VOST  samp-
ling runs were made each day (Runs
1-8).    Destruction   and   removal
efficiencies,  calculated  for waste
fuel  runs 4-8,   are  summarized  in
Figure 2.   In general, DREs ranged
from  99.60%  to   >99.999%   for  all
compounds  and  averaged  99.9989%.
Only four  runs  had DREs  less than
99.99%;  three of  these  were  for
methylene  chloride, the  fourth for
1,1,1,-trichloroethane.

     DREs  for  methylene  chloride
ranged from 99.60%  to <99.999% and
averaged  99.983%.   Three runs at a
DRE less  than 99.99%  (Run Nos. 5A,
5F,  and  4E).   In Run No.  5A,  the
tube  mass collected  was  1,066  ng
compared to a maximum of 55 ng for
all of the other runs.  This high
mass was  most likely  caused by lab
contamination.   Run No.  5A  is  not
included  in  the  averages,  due  to
the  lab  contamination,   thus  the
Day 5  average is 99.995%  and  the
overall   methylene    chloride   DRE
average is 99.997%.

     Methyl  ethyl  ketone  had  an
average  DRE of  99.999%  and .-ranged
from   99.998%   to    greater   than
99.999%.    These  high  destruction
efficiencies     were     consistent
throughout  the   waste  fuel  test
runs.

     DREs for 1,1,1-trichloroethane
                                     52

-------
              TABLE 2.   RESULTS OF CAPILLARY GC/MS ANALYSIS
                        OF MAJOR COMPONENTS OF WASTE FUELS

                                             Concentration,
                                                  wt %
Waste fuel component
Acetone
Methyl ethyl ketone (POHC)
1,1, 1-Trichloroethane ( POHC )
1-Butanol
Trichloroethylene (POHC)
2 -Ethoxyethanol
Methyl isobutyl ketone
Toluene (POHC)
Tetrachloroethylene (POHC)
Butyl acetone
Ethylbenzene
Xylene (isomer No. 1)
Xylene (isomer No. 2)
2-Butoxyethyanol
2-Ethoxyethyl acetate
C3 -Benzene (isomer No. 1)
C3 -Benzene (isomer No. 2)
C10 -Alkane
Alkane >C8
Alkane
C1:l- Alkane
2 - Cyc 1 ohexen- 1 - one
Alkane >C6
Number
4s-
0.23
2.48
0.24
0.32
1.73
0.85
1.06
11.0
2.17
0.27
1.42
4.92
1.43
1.99
5.91
0.28
0.46
0.80
0.24
0.14
1.26
0.15
0.27
Number
0.22
3.17
0.22
0.37
2.16
0.92
1.16
12.5
2.49
0.32
1.58
5.58
1.60
2.07
6.37
0.32
0.57
0.94
0.28
0.18
1.48
0.18
0.24

              Average of split sample.
TABLE  3.   CONCENTRATION  OF  POHCs
   TABLE 4.  WASTE FUEL CONDITIONS
                   Average
                concentrati on,
POHC
MeCl2
MEK
CH3CC13
TCE
Perc
Toluene
wt %
0.11
2.69
0.24
1.84
2.26
11.6

                          Average
	Parameter	concentration

Chlorine, % vol              3.1
Sulfur, %                   0.07
PCB, ppm                    <1.0

Heat value, Btu/lb        12,870

Specific gravity,
  g/cc                      1.01

Feed rate, gpm Run 4        0.64
                   5        1.21
                   6        2.05
                   7A       0.78
                   7B       2.90
                   8        2.88
                                     53

-------
   30-
   20H
o  10-
Methylene  Chloride
   30-
   20-
o  10-
          12345

           Number of Nines ORE
Methylethyl ketone
          1     23     4     5

          Number of Nines  ORE
                                                    30-
                                                    2CH
                                                 0  10H
                                                  
                                                 o
                                                  Trlchloroethylene
                                                    30-
                                                    20-
                                         0  ioH
                                         d
                                                   1234

                                                     Nunber  of Nines ORE
                                                 Tetrachloroethylene
                                                              i     i
                                                   12345

                                                     Number of  Nines ORE
   30-
   20-]
    1
o  10-
1, 1, 1  Trlchloroethane
 1     i   ^     4     !;
  Number of Nines ORE
                                                    30-
                                            20-
                                          .  10-
Toluene
                                                           1234

                                                             Number of Nines ORE
             Fiqure 2.   Destruction and removal efficiencies.
                                          54

-------
 ranged  from 99.989% to  99.999% and
 averaged  99.997%.  Only Run No. 4E
 had  a DRE  less  than 99.99%.

      DREs    for   trichloroethylene
 (TCE) were greater than 99.999% for
 all  runs.    TCE was  spiked to  the
 waste  fuel prior to the waste fuel
 testing  to increase its concentra-
 tion and allow easier  detection of
 TCE  in the stack  gas.

      DREs   for  tetrachloroethylene
 (Perc)   also   were  greater   than
 99.999%  for  all  runs.   Like  TCE,
 Perc was spiked  to the  waste  fuel
 to the maximum  allowable concentra-
 tion ' described in the  test  burn
 permit prior to the test.

      Toluene was the POHC  of high-
 est  concentration in the waste fuel
 (average   11.6% by weight).   DREs
 for  toluene were above  99.999% for
 all  runs.   Data   for  toluene  was
 very consistent  during all  waste
 fuelvtest  runs.

 Stack Samples

      Results  for  stack  conditions
 and  particulate,  HC1,  SO2,  NO ,  CO,
 and  THC  emissions  for baseline and
 waste  fuel  runs  are summarized in
                     Table 5.  The  dry stack rate aver-
                     aged 487 dscm/min (17,210 dscf/min).
                     As evidenced  by  the  high standard
                     deviations the CO and,  to a lesser
                     degree,  the SO2  fluctuated.  Minor
                     kiln upsets (i.e.,  coke feed chute
                     cleaning,  clumps of coke falling to
                     kiln, change in process conditions)
                     created  high   CO  excurtions.   A
                     trend which  occurred  quite  often
                     was an increase in SO2 by ~200 ppm,
                     followed by a reduction  in  NO  by
                     ~50  ppm,  following by  an increase
                     in CO by ~500  ppm over a 15 minute
                     period.   These  trends are expected
                     when a lower intensity flame occurs
                     (or  kiln  upset).  However,  as re-
                     vealed by Figure 2, the kiln upsets
                     had little or  no effect on the DRE
                     results.

                     Chlorine,  Sulfur and Metals Balance

                          Chlorine   and sulfur  macerial
                     balances are  summarized  for  base-
                     line and  waste  fuel  conditions in
                     Table 6.  The  majority of chlorine
                     (for either baseline  or waste fuel
                     conditions)  enters the  kiln in the
                     limestone  feed and exits  the kiln
                     in the  lime  product  and baghouse
                     dust.  Sulfur  (for either baseline
                     or waste  fuel  conditions)  enters
                     the kiln in the  petroleum coke and
                      TABLE 5.  AVERAGE  STACK EMISSIONS

Parameter and unit
Stack rate, ms/miri
Stack velocity, m/sec


Range
805
5.0
- 975
- 6.0
Baseline
Average
917
5.7
Waste fuel
Standard
deviation
76
0.4
Range
791
4.9
- 938
- 5.8
Average
847
5.2
Standard
deviation
52
0.3
Particulates
  mg/dscm
  kg/hr


HCl, ppm


S02, ppm

NO , ppm
  A

CO, ppm


THC, ppm
24.0 - 35.0
0.66 - 1.1

0.74 - 3.9

 123 - 730

 306 - 460

  10 - 4,900

 6.7 - 12.7
28.7
 0.9

 2.0

 553

 386

 477

 8.2
4.7
0.2

1.4

110

 49

966

1.9
24.9 - 48.7
0.68 - 1.4

 2.5 - 6.0

 183 - 1,924

 288 - 552

  10 - 4,540

 1.5 - 10.0
35.3
 1.0

 4.4

 596

 446

 599

 3.5
  8.0
  0.3

  1.2

  240

  64

1,409

  1.1
                                      55

-------
exits  the kiln  distributed in the
lime  product (~9%),  baghouse dust
(~27%), and stack gas (~64%).
     An elemental  material balance
was  performed   for  21  elements.
Balance results  show no difference
for baseline  and waste fuel condi-
tions for distribution of metals in
the kiln  process.   Also the major-
ity of mass  entering the  kiln is
contributed  by the  limestone  feed
(typically   80-100%)  except   for
zinc  which has  42%  in the  lime-
stone  and 53% in the waste  fuel.
The mass  exiting the kiln  is  dis-
tributed  between the lime  product
and  baghouse  dust  (approximately
50-90%  in product  and  10-45%  in
dust).  Magnesium and calcium had
mass  rates  (to  or  from kiln)  of
approximately 2,500  kg/hr.   Sodium
and  iron mass  rate were  approxi-
mately  20 kg/hr.  All  other  ele-
ments  had  mass  rates  less  than
5 kg/hr.

Baseline  vs. Waste Fuel
and Kiln  Operation

      Emissions    were    evaluated
under baseline  and waste fuel con-
ditions.     For    the   pollutants
listed in  Table 5,   HC1,   NO   and
TEC  showed  a significant  differ-
ence   in  stack  emissions  under
baseline  and  waste  fuel  condit-
        ions .   For the POHC's, only methy-
        lene chloride and toluene showed a
        minor  difference  (increase)  from
        baseline to waste fuel conditions.
         All  remaining  POHC's  showed  no
        significant  difference  in  base-
        line  vs.  waste  fuel  emissions.

             As  described previously,  the
        kiln operation  fluctuated as indi-
        cated by CO  and SO2  emission vari-
        ations  during waste  fuel  burning.
        Kiln  fluctuations  were caused by
        several items including nonconstant
        fuel rates,  product rushes, clumps
        of  coke fed  to  kiln accidentally,
        and   operator   inexperience   with
        burning  waste fuel.   The  fluctua-
        tion  resulted  in occasional  kiln
        O2 increases  and stack gas SO2 de-
        crease  causing  a  poorer  quality
        lime  product  most  likely due to
        excess  sulfur in the product.  The
        following  items  were identified as
        ways to improve  operation at this
        kiln under waste  fuel conditions:

        1.  Change waste fuel burner con-
            figuration such that at low
            waste fuel rates the waste
            fuel is mixed with the coke
            to maintain a flame.
        2.  Decrease the fan speed  (i.e.,
            reduce the draft) to lower the
            O2 in the kiln resulting in
            lower sulfur in product and
            higher sulfur in the stack.
              TABLE 6.  CHLORINE AND  SULFUR MATERIAL BALANCE
             Percent to kiln
Total
Percent from kiln
                                                              Total
                                      Material
                                      balance
Run
number
Chlorine
Baseline3
Waste fuelb
Sulfur
Baseline
Waste fuelb
Coke
4
5
100
99
Waste
fuel
0
2
0
1
Limestone
feed
96
93
NA
NA
mass in,
kq/hr
23
21
59
54
Lime
product
47
34
8
10
Baghouse
dust
52
65
29
25
Stack
gas
1
1
63
65
mass out, closure,
kg/hr %
13 55
20 105
91 73
70 77
Average values of baseline Runs 1-3.

Average values of waste fuel Runs 4-8.
                                     56

-------
               FIELD TESTS OF INDUSTRIAL BOILERS COFIRING HAZARDOUS WASTES

          Carlo Castaldini, Howard B. Mason, Robert J. DeRosier, Stefan Unnasch
                                   Acurex Corporation
                            Mountain View, California  94039


                                        ABSTRACT

     Thirteen field emission tests have been performed on eight industrial boilers
cofired with conventional fuels and hazardous wastes.  The U.S. Environmental Protection
Agency (EPA) sponsored the tests to evaluate the effectiveness of industrial boilers for
hazardous waste thermal destruction while recovering useful energy.  Five initial tests
were made on a wide range of waste type/boiler design combinations to quantify emissions
representative of current industry practices.  For the second series of tests, boilers
firing waste streams spiked with a mixture of carbon tetrachloride, chlorobenzene, and
trichlorethylene were tested to extend the data base for chlorinated wastes and
facilitate intrasource comparisons.  Principal organic hazardous constituents (POHC's) in
the flue gas were quantified using two extractive sampling trains.  Volatile compounds
were absorbed on two Tenax traps in the volatile organic sampling train (VOST).
Semi volatile compounds were absorbed on XAD-2 sorbent in a modified Method 5 train.  The
sampling protocol also included continuous monitor measurement of stack 0%, CO, C02, NOX
and total hydrocarbons; extractive train samples of HC1; and grab samples of the
conventional and waste fuels and ash streams.  Post-test analyses of the inlet fuels and
the outlet stream samples were done with gas chromatography/mass spectrometry to quantify
the destruction and removal efficiency of POHC's.  Results for the thirteen tests are
discussed in this paper.  Specie destruction and removal efficiency (DRE) ranged from
99.5 to >99.999 percent.  The mass weighted destruction efficiency for all RCRA
Appendix VIII compounds was -99.998 percent.  Products of incomplete combustion (PIC's)
were observed in concentrations one order of magnitude greater than measured breakthrough
POHC's.
INTRODUCTION

     Many hazardous wastes have
sufficiently high heat of combustion to be
candidates for cofiring with conventional
fuels in industrial boilers.  Cofiring can
be economically and environmentally
advantageous to other disposal/destruction
options while meeting the Resource
Conservation and Recovery Act (RCRA) goal
of resource recovery.  The benefits of
cofiring include partial displacement of
conventional fuels, and use in existing
industrial boilers where incinerators are
not available.  Although boiler cofiring
is not currently regulated, the need for
inclusion in RCRA provisions is currently
being studied by EPA.  EPA sponsored the
tests discussed in this paper to quantify
the waste destruction potential of
industrial boilers.  To obtain results
representative of current or potential
industrial practice, a broad range of
boiler designs and waste types was
selected for testing (1).  Eight sites
were tested encompassing thirteen tests of
which the initial three were reported at
the Ninth Symposium(2).

APPROACH

     Candidate sites were screened based
on representativeness of the boiler design
and wastes being fired, and the
availability and accessibility'for
cofiring tests.  Within this context,
preference was given to sites which were
regarded as more challenging in attaining
                                           57

-------
a high level of waste destruction.
Table 1 summarizes the eight test sites
and thirteen test series.  Sites A and H
were solid-fuel-fired and were equipped
with a cyclone and electrostatic
predpitator, respectively.  Site G was
fired with chlorinated hydrocarbons
without auxiliary conventional fuel and
was equipped with two scrubber columns for
HC1 recovery and cleanup.  The selected
sites spanned a broad range of design and
operating conditions:  firetube and
watertube designs; capacities from
1.1 kg/s (8,500 Ib/hr) to 32 kg/s
(250,000 Ib/hr); gas, oil, coal arid wood
firing; loads from 25 to 100 percent of
rated capacity; wastes ranging over an
order of magnitude in heat of combustion;
and residence time and heat release
variations over an order of magnitude.  To
extend the range of waste destruction
characteristics tested, the wastes at
site E-H were spiked with carbon
tetrachloride and, in most cases,
monochlorobenzene and trichloroethylene.

     The nominal test series involved an
initial conventional fuel baseline test to
characterize unit operation and emissions
in the absence of waste firing, and
          VV.isle
          Day
          T.mK
triplicate cofiring runs.  For these runs,
the unit load was held constant to enable
replicate comparisons of results.  In most
other respects, routine operational
variations such as excess air levels and
waste flowrates were tolerated to obtain
results representative of normal
operation.

     The major inlet and outlet streams
were sampled and analyzed as shown in
Figure 1, for the case of a coal-fired
unit, and boiler operational data was
taken to characterize performance with and
without waste firing.  Retails on the
protocol are given in (1).  Waste and fuel
grab samples were taken approximately
every hour, composited, and analyzed in
the laboratory for ultimate and proximate
analyses, chloride content, and POHC
concentration.  Bottom and hopper ash
composite samples were analyzed for
chlorides, POHC's, and carbon content.
The major sampling was conducted at the
stack where the following samples were
taken:

        Continuous monitor analyses of
         02, CO, C02, NOX, and UHC
            A  Liquid Waste Grab Samples (Composite)        D,F  Stack Emissions
            B  Fuel Grab Samples (Composite)               E   Particulate Collector
            C  Boiler Bottom Ash Grab Samples (Composite)          Hopper Ash


                               Figure 1.   Sampling schematic.
                                             58

-------
                  TABLE 1.  BOILER DESIGN AND OPERATING CHARACTERISTICS
                                                      Operation
Boiler^ Type
and Capacity
Site kg/s (103 Ib/hr)
A WT Stoker
1.26
(10)
B FT 1 Burner
1.1
(8.5)
C WT 6 Burner
29
(230)
D WT 4 Burner
ii
(90)


E WT 1 Burner
14
(110)








F WT 2 Burner
7.6
'(60)
G FT 1 Burner
4.0
(32)
H WT T-Fired
32
(250)
Waste S
Primary (percent heat input)
Fuel POHC's* (
Wood (40)
waste creosote

Gas Al kyd
wastewater

Gas (38)
phenol /cumene

Oil (26)
xylene, PCE

(48)
toluene, BCEE
Oil (28)
MMA, CC14, Cl, TCE

(42)
MMA, CC14, Cl<|,, TCE
(19)
' MMA, CC14, Cl, TCE
(52)
MMA, toluene
Oil (10)
solvent, CC14,
C1,j>, TCE
NA (100)
CC14, chlorinated
hydrocarbons
Coal (3)
MeA, CC14, Cl$,
1,1,1-TCA
:team Load Residence
kg/s Percent time
103 Ib/hr) 02 (s) (:
1.26 6-16 1.2
(10)

0.26 4-6 0.8
(2)

7.4 10 2
(59)

8.8 4-6 1.1
(70)

5.7 6 1.3
(45)
6.9 7 0.7
(55)

5.0 6  1.0
(40)
10 7 0.5
(80)
7.2 6 0.8
(57)
5.6 7 1.1
(44)
4.0 7-11 2
(32)

1.8 8 0.4
(14)

32 6 2
(250)

Heat
release
kW/m3
LO3 Rtu/ft3h)
300
(29)

750
(72)

78
(7.5)

400
(39)

240
(23)
520
(50)

350
(34)
770
(74)
470
(45)
370
(36)
110
(H)

820
(79)

180
(17)

*PCE = tetrachloroethylene; BCEE = bis(2-chloroethyl)ether; MMA = nethylmethacrylate;
 Cl<(, = monochlorobenzene; TCE = trichloroethylene;  MeA  = methyl acetate;
       1,1,1-TCA  =  1,1,1-trichloroethane
1"WT = watertube;  FT = firetube
                                              59

-------
        Volatile organics extractive
         samples by the Volatile Organic
         Sampling Train (VOST)  (3)

        Semi volatile organics  and
         particulates by the modified
         Method 5 extractive sampling
         train (4)

     t   Chlorides by a Method  6
         extractive sampling train

        GJ-CS hydrocarbons by  a gas bomb
         grab sample and gas chromatograph
         analyses

Each test required approximately 6 hours
of run time.  Post-test analyses of the
volatile and semi volatile samples
collected on resin traps were done by gas
chromatography/mass spectroscopy (GC/MS).
Waste destruction and removal efficiencies
were computed based on feed and stack mass
flowrates quantified in the samples:

   ORE = (Mfeed - Mstack)/Mfeed x 100

Constituents appearing in the stack which
were not identified in the feed or other
inlet stream were identified as products
of incomplete combustion (PIC's).
   RESULTS

        Table 2 summarizes POHC-specific ORE
   results.  Calculated DRE's are based on
   blank corrected emission rates measured
   during cofiring.  Therefore,, credit is not
   given for emission of POHC's measured
   during baseline tests.  Results indicate a
   wide range in DRE's from 99.5 to
   >99.999 percent.  The mass weighted
   average ORE for all POHC's was
   99.998 percent.  Although the average ORE
   for each POHC was generally greater than
   99.99 percent, some emission measurements
   resulted in DRE's below this level, the
   current RCRA incinerator standard.  These
   low DRE's often coincided with unsteady
   boiler operation and burner combustion
   instability.  For example, low DRE's for
   carbon tetrachloride, chlorobenzene, and
   trichloroethylene are generally
   attributable to site F.  Improper burner
   settings at this site resulted in coking
   of the burner nozzle, fuel impingement on
   burner throat, and occasionally high
   levels of combustible CO and soot
   emissions.

         Low DRE's for site E methylmetha-
   crylate were measured during few tests
   characterized by instability in waste
                          TABLE 2.  SUMMARY OF POHC ORE RESULTS

POHC
Carbon tetrachloride
Trichloroethylene
1,1,1 Trichloroethane
Chlorobenzene
Phenol
Pentachlorophenol
Tetrachloroethylene
B1s(2-chloroethyl )ether
Toluene
Fluorene
Naphthalene
Methyl nethacry late
2-4 Dlraethylphenol
Epichlorohydrin
B1s(2-chloro1sopropyl Jether
Inlet
(ppra) |Hwl x 10-
IR?J
1.0-50
0.78-12
0.82-0.96
0.13-7.5
0.23-21
0.88-2.4
8.0-90
18-20
0.46-450
1.8-3.0
2.2-7.6
7.9-62
0.12-0.52
160-210
400-540
Emission
3 rate
Mg/s
1-460
28-470
<0.3-17
3.6-190
<0. 6-190
<3.2-44
510-1,400
<3-6.8
29-8,800
<0.6-34
25-150
17-8,300
<3.2-13
<2
<2
Mass
average
ORE
percent
99.998
99.998
99.994
99.998
99.999
99.998
99.998
>99.999
99.998
99.996
99.983
99.991
>99.983
>99.999
>99.999
ORE
range
percent
99.97-99.999
99.98-99.999
99.98-99.999
99.97-99.999
99.5-99.999
99.985-99.998
99.994-99.999
>99.999
99.90-99.999
99.986->99.999
99.946-99.988
99.95-99.999
99.96->99.99
>99.999
>99.999

Site
E, F, G,
E, F
H
E, F, H
A, C
A
D
n
B, D, E,
A
A
E
A
G
G


H







F






                                0.12-540
<0.6-8,800   99.998    99.5->99.999
                                            60

-------
feedrates and insufficient combustion air.
These operating conditions led to several
high CO and smoke emission episodes.
Wood-fired stokers, such as the site A
boiler, are typically high excess air and
high CO emitters.  These conditions result
from the physical properties of wood waste
(e.g., wood chip size and high moisture
content), combustion cooling by very high
excess air levels and inefficient
fuel-air/mixing of combustion on a fuel
bed.  Site A DRE's were generally below
99.99 percent.  Table 3 illustrates this
site dependence more clearly.  An attempt
to correlate DRE's with CO emissions was
not successful, however.  It is likely
that this was in part due to insufficient
emission measurements during periods of
high CO or smoke emissions.

     Figure 2 illustrates an apparent ORE
dependence on POHC concentration in the
waste fuels.  The data are based on mass
average DRE's for all POHC's at each site.
Therefore, included are ORE  results for
other organics not listed in Table 3.  The
data indicate that higher DRE's were
obtained with higher POHC firing rates.
This result suggests that the amount of
POHC breakthrough is independent of the
POHC concentration in the waste fuel.  One
factor that contributes to this trend is
the high DRE's for semi volatile organics
(phenol, BCEE, epichlorohydrin, and
bis(2-chloroisopropyl)ether) that were
present at high concentrations in waste
fuels of sites C, D, and G.  Results for
other nonhazardous organics present at the
high concentrations in site E waste fuel
also showed high DRE's.  A second factor,
is the background contamination from
sampling and analysis techniques which
tends to be more significant at low POHC
emissions.  Also important is the
contribution of volatile organics from
fossil fuel combustion.

     Quantisation of PIC emissions was
based on the detection of organic priority
pollutants not found in the waste fuel.
Figure 3 illustrates the relationship
between the sum of total PIC's,
halogenated and nonhalogenated organics,
versus the test average ORE for site D
through H.  The data suggest that the
lower the quantity of  POHC breakthrough
(i.e., higher ORE) the lower the total
quantity of PIC's.  Therefore, higher POHC
DRE's are also likely  to result in lower
PIC emissions.  On the average, PIC
emissions were one to  two orders of
magnitude greater than the total quantity
                           TABLE 3.   SUMMARY OF SITE ORE  RESULTS


Site
A
B
C
D
E
F
G
H



<
ORE range
99.5-99.999
99.991
99.998->99.999
99.994-99.9996
99.95-99.9995
99.90-99.999
99.995->99.999
99.97-99.999
99.5->99.999


CC14
99.999
99.98
99.998
99.988
99.998
POHC DRE
percent
TCE* Cl<|) PCP1"
99.988
99.998 99.998
99.996 99.98
99.992
99.998 99.998 99.988


PCEt1" Toluene
99.98
99.998 99.9994
99.997
99.95

99.998 99.998


Average
DRE
99.98
99.98
99.9990
99.9990
99.995
99.98
99.9996
99.991
99.998
   *TCE = trichloroethylene
   i"PCP = pentachlorophenol
  t1"PCE = tetrachloroethylene
                                            61

-------
                   FIELD TESTS OF  INDUSTRIAL BOILERS AND  INDUSTRIAL PROCESSES
                                DISPOSING OF HAZARDOUS WASTES

                   Radford C. Adams, Michael W. Hartman and Denny E. Wagoner
                                     Radian Corporation
                        Research  Triangle Park, North Carolina  27709


                                          ABSTRACT

     Two industrial boilers were  tested to determine their hazardous waste destruction
performance.  The  tests are part  of a research program conducted by the EPA to evaluate a
spectrum of representative boiler types that are currently being used for hazardous waste
disposal.  These boiler tests provide destruction efficiency (DE) results as the primary
performance parameter.  Waste materials typically burned in the boilers were spiked with
a standard test mixture containing components expected to vary in incinerability.  The
standard test mixture provides the EPA with a means of comparing the thermal destruction
performance among  several boilers.

     The boilers were operated at a single steady state load of 80% of design capacity
with operator settings of excess  air and feed rate controlled by normal operating pro-
cedures.  A series of three background runs and three waste feed runs were conducted at
the prescribed boiler load.  Additional sampling runs were required during Boiler No. 1
testing to evaluate NOX control with and without staged combustion.

     The destruction efficiencies of ten POHCs (principal organic hazardous constituents)
were determined at low and high waste to fuel  ratios.  No correlation was found of
destruction efficiency with potenital incinerability indicators (carbon monoxide and
total hydrocarbons).  Only two products of incomplete combustion (PICs) were identified
during the two bciler tests.  Numerous organic species not detected in the feed were
detected in the emissions but these are believed degradation products of the sample
collection media.  Staged combustion NOX control  was used when burning nitrogenous waste
and was found to reduce NOX emissions.  Heavy metals concentrations in the emissions when
burning wastes were compared with concentrations when burning the baseline fuel.
INTRODUCTION

     Two industrial boilers were tested to
determine their hazardous waste destruction
performance.  The tests are part of a
research program conducted by the EPA to
evaluate a spectrum of representative
boiler types that are currently being used
for hazardous waste disposal.
     These boiler tests provide destruction
efficiency (DE) results as the primary
performance parameter.  Waste materials
typically burned in the boilers were spiked
with a standard test mixture containing
components expected to vary in inciner-
ability.  The standard test mixture pro-
vides the EPA with a means of comparing the
thermal destruction performance among
several boilers.
                                            62

-------
     The two boilers epitomize considerable
diversity in waste disposal goals and
waste feed characteristics.  Wastes are
burned in Boiler No. 1 primarily to dispose
of a hazardous waste and not for apprecia-
ble fuel savings.  Waste to fuel ratios
determined on a heat input basis were
0.09 to 1, meaning that only 8% of the
energy was supplied by the waste.  There
is no clearly identifiable waste to fuel
ratio to report for Boiler No. 2 because a
substantial amount of the waste burned is
off specification fuel.  During the test,
No. 6 fuel oil contributed less than 50%
of the input energy and it is clearly the
intent to obtain substantial fuel savings
by burning either wastes or off specifica-
tion fuels.  Table 1 summarizes significant
differences in the design and waste disposal
characteristics of the two boilers.
Boiler No. 1 represents a class of boilers
disposing of a specific waste that is well
characterized with respect to knowledge of
the composition and of the expected volume
of waste.  Boiler No. 1 also represents
cofiring of a liquid waste with a gaseous
fuel.  Boiler No. 2 represents a class of
boilers that are not limited to a specific
waste stream.  Variation .in fuel and waste
composition is the expected norm for
Boiler No. 2.  Boiler No. 2 cofires liquid
wastes and off specification fuels with
No. 6 fuel oil.  Specific components of
the liquid wastes are not always known.
PURPOSE

     The two boilers were tested by similar
procedures.  The principal goals were:

          the identification of principal
          organic hazardous constituents
          (POHCs) in the feed and the
          extent of their destruction,

          the correlation of destruction
          efficiency with boiler operating
          settings with the boiler
          operating'at steady state,

          the fate of other feed components
          such as chlorine, heavy metals,
          and fuel bound nitrogen,

          the identification of products
          of incomplete combustion (PICs),
          and

          the correlation of candidate
          incinerability indicators with
          destruction efficiency.

The objectives of this report are to
(1) report and compare the results of the
two boiler tests and (2) identify and
highlight specific thermal destruction
measurement limitations that were encoun-
tered during the test programs.
                         TABLE 1.  BOILER DESIGN AND OPERATING DATA
                                         Boiler No. 1
              Boiler No. 2
      Manufacturer                            FW
      Model number                           AG252

      Heat exchanger
      Design fuel .                            Gas
      Design excess air                       3%
      Design steam rate, Ib/hr              68,000
      Steam temperature,  F                    353
      Steam pressure, psig                     250
      Design fuel heating value, Btu/lb     23,000
      Waste heating value, Btu/lb           10,600
                   CE
                  VU-10
    -Water Tube-
                No. 6 oil

                   3%

                 60,000

                    406

                    125

                 18,500

                 17,700
      FW - Foster Wheeler
      CE - Combustion Engineering
                                            63

-------
 APPROACH

      The boilers  were operated at a single
 steady state load of 80% of design capacity
 with operator settings of excess  air and
 feed rate controlled by normal  operating
 procedures.   A series of three background
 runs and three waste feed runs were con-
 ducted at the prescribed boiler load.
 Additional  sampling  runs were  required
 during Boiler No.  1  testing to evaluate
 NQx  control  with  and without staged
 combustion.

      Characteristics of the boiler feed
 streams  and  procedures for sampling and
 analysis are described in the  following
 sections.

 Feed  Characteristics

      Boiler  No. 1  was fired with  fuel gas
 and  a liquid nitrogenous waste.   The fuel
 gas was  predominantly natural  gas  with
 small  amounts of a waste gas not  substan-
 tially different from natural  gas.   The
 nitrogenous  waste  was predominantly
 nitrobenzene with  the balance  consisting
 of benzene,  aniline  and  water.  Waste to
 fuel  ratio on a weight basis averaged
 0.2  to 1 at  a liquid waste feed rate of
 0.6 gpm  (362 Ib/hour).

      Initial  boiler  design included  two
 vertically oriented  ring burners.   The
 burner system had  been modified by the
 addition of  steam  atomized liquid  burners.
 The boiler controls  had  been modified to
 achieve  staged combustion  during periods
 of liquid waste firing by  restricting air
 flow  to  the  bottom set of  liquid and gas
 burners.  Air not  supplied  to the  lower
 burners  was  directed to  the  upper  burner
 to maintain  the total  air  flow necessary
 for boiler excess  air requirements.  Two
 series of sampling runs  were conducted to
 compare  NOX  emissions  with and without
 combustion control.   To  do this, the
 automatic controls for restricting air to
 the lower burners was  disabled.  Balanced
 air flow to  each set of  burners was  thus
maintained to establish  the  "unstaged"
 combustion effect.

      Boiler  No. 2 normally burns one of
 two fuels that are classified according to
 viscosity.   Generic  designations are
 "heavy oil"  for the  high viscosity liquids
 and "light oil" for  the  low  viscosity
 liquids.  Various  blends can be created of
 No.  6  fuel oil with various combustible
 liquids  such as recycled solvents, alco-
 hols,  mineral spirits and substandard fuel
 oils.  The blends vary in composition
 according to availability.  Boiler No. 2
 is capable of burning two different blends
 simultaneously and this option was chosen
 for  testing.  The boiler has four burners
 oriented horizontally on the front face of
 the  boiler.  The heavy oil and light oil
 were not substantially different in heating
 value.  Therefore, approximately equal
 amounts of each feed stream could be
 burned and a set of two burners each were
 set  aside for each feed type.  Heavy oil
 was  steam atomized and light oil was air
 atomized.

     Waste liquids were spiked with a
 standard test mixture:  carbon tetra-
 chloride, 1,1,2-trichloroethylene, mono-
 chlorobenzene, and toluene.  To protect
 the  boiler surfaces from chloride corrosion
 damage, quantities were limited to one
 weight percent as chlorine of the total
 fuel and waste.   Selection of test mixture
 components is an attempt to provide a
 broad range of incinerability of hazardous
 compounds as postulated from relative
 heats of combustion (1).   Actual total
 chlorine in the boiler feed was 0.7 weight
 percent for both boiler tests.   The
 standard test mixture was added to the
 light oil of Boiler No.  2.   Table 2 shows
 typical feed stream compositions for the
 two boilers.

 Sampling

     Feed and emissions  samples were
collected simultaneously.   Boiler No.  2
was operated with two liquid feeds during
testing and samples of each were collected,
composited and analyzed  separately.   One
sample per run was collected of Boiler
No. 1 feed gas.

     Emissions were sampled with three
sampling trains:

          the volatile organic  sampling
          train  (VOST)  for  volatile
          organics,

          the modified  EPA  Method 5  train
          (MM5)  for semi-volatile organics
          and metals,  and

          the EPA Method  5  train for  HC1
          and metals.
                                            64

-------
                                TABLE 2.  FEED COMPOSITIONS
                                              Weight % of fuel + waste
                                         Boiler No. 1
                 Boiler No. 2
      Fuel gas                               83.8

      Standard test mixture

           Carbon tetrachloride               0.3
           Trichloroethylene                  0.3
           Monochlorobenzene                  0.5
           Toluene                            0.5

      Liquid waste
           Nitrobenzene                      13.4
           Phenol                        not detected
           Aniline                            0.4
           Benzene                            0.3
           m&p-Xylene                    not detected
           o-Xylene                      not detected
           Water                              0.5
           Small quantities of             not used
            typical fuel oil
            components and solvents

      Total POHCsdetected                   15.7
                   not used
                      0.5
                      0.4
                      0.4
                      2.0
                 not detected
                     11.8
                 not detected
                      0.2
                      4.3
                      0.6
                      1.1
                    balance
                     20.2
     The VOST train collects the more
volatile organics from a filtered, pre-
cooled gas stream by adsorption on Tenax
resin.  The, gas was passed through two
sorbent traps in series.  The first trap
contained Tenax and the second trap con-
tained a layer of Tenax followed by a
layer of charcoal.  Protection from sample
contamination by trace organics dictates
strict protocols for cleaning glassware,
for sealing glass joints without the
use of grease, and air tight sealing of
sorbent traps during storage and shipment.
Minimum leakage at glass joints was assured
by strict adherence to leak check proce-
dures and guidelines.  The samples were
recovered from the Tenax by thermal
desorption and analyzed by gas chromato-
graphy/mass spectrometry procedures.

     The modified Method 5 train collects
organics with boiling points above a
nominal 100 C level by adsorption on
XAD-2 resin.  An EPA Method 5 train was
modified by addition of a sorbent trap
ahead of the impinger train.  Temperature
of the sorbent traps was controlled so as
not to exceed a 20 C exit temperature.
Organics collected in traps were recovered
by solvent extraction from which the
sample-containing solvent was analyzed  by
GC/MS after appropriate concentration or
dilution.  Some samples were prescreened
by GC/FID to determine expected calibration
ranges for the GC/MS.

     The EPA Method 5 train was used to
collect HC1 samples at a single point on
the duct traverse.  The HC1 was absorbed
in 1.0 molar solution of sodium acetate
and the sodium acetate combined with
filter extract was analyzed by ion chro-
matography.  Metals were also collected by
this sampling train during Boiler No. 1
testing.  Boiler No. 2 testing utilized a
second EPA Method 5 train to collect
particles for metals analysis.  Method  5
procedures were employed.
                                            65

-------
 PROBLEMS ENCOUNTERED
                                                Benzene Interference
 Feed Preparation

      Specific quantities of the standard
 test mixture (carbon tetrachloride,  tri-
 chloroethylene,  monochlorobenzene,  and
 toluene)  had to  be added to the plant
 wastes and mixed well  enough to attain
 consistent feed  composition throughout  the
 test period.  A  premixed batch  was  prepared
 in enough quantity to  provide all of the
 feed mixture needed.   Examination of
 sample mixtures  prepared prior  to testing
 showed that only one liquid phase was
 present.   However,  variation in density at
 different tank levels  was observed  for'the
 Boiler No.  2 liquid feed mixture.   Hence,
 constant  circulation with rapid turnover
 of the stored liquid was required to
 minimize  stratification.

 Ambient Air Contamination

      The  VOST sampling train is equipped
 with a charcoal  filter to trap  organics of
 the ambient air  when the train  is pres-
 surized for leak testing.   It was found
 during Boiler No.  1 testing that organics
 in  the ambient air  exceeded organic
 emissions.   The  modified  Method 5 train
 was not equipped with a  similar air
 cleaning  filter  until Boiler No. 2 testing.
 Evidently,  there were no  POHCs  or Pics  in
 the ambient air  in  detectable quantities
 during Boiler No.  1 testing that would
 contaminate modified Method 5 samples.
 Nevertheless,  as a  precaution,  all ambient
 air to the  organic  sampling trains will be
 filtered  during  future test programs.

 Identification of Products  of Incomplete
 Combustion  (PICs)"

      Comparisons  of analytical  results of
 all  test  runs, both waste feed  and baseline
 runs,  consistently  show the  same compounds
 in  the  extracts  from XAD-2  emission
 sampling  traps.   The same compounds are
 found  in  the  extracts from  field blanks.
 These  compounds  are attributed  to degrada-
 tion/leaching of XAD-2 resin and not to
 PIC  formation.  The collection method
 (XAD-2  resin) followed by fused  silica
 capillary chromatography with mass spectro-
meter detection  allows for  excellent
determination of  the destruction efficiency
of  POHCs, but the identification and
 quantisation of  PICs must  be approached
with caution.
     Benzene analysis indicates that
destruction efficiencies were less than
99.99% during both Boiler No. 1 and Boiler
No. 2 testing.  Statistical treatment of
Boiler No. 1 test data showed that benzene
emission levels during the waste feed runs
were not significantly different from the
baseline runs.  We thus attributed benzene
concentrations in the emissions to a
benzene background in the Tenax sampling
trap.  Benzene levels during Boiler No. 2
testing were an order of magnitude greater
in the baseline runs,than in the waste
feed runs despite the fact that no benzene
was detected in the baseline fuel  (No. 6
fuel oil).  This is a different set of
circumstances than the Boiler No.  1 case
and a phenomena we are unable to explain.
Unfortunately, destruction efficiency of
benzene was not measureable during the two
boiler tests due to the apparent benzene
interferences.

RESULTS

     Program accomplishments were  as
fol1ows:

          Destruction efficiencies of ten
          POHCs (principal  organic hazard-
          ous constitutents) were  deter-
          mi ned.

          Hazardous waste disposal
          performance at low and high
          waste/fuel  ratios were compared.

          Data for correlation  of  destruc-
          tion efficiency with  operating
          parameters  were provided.

          Very few PICs  (products  of
          incomplete  combustion) were
          detected and identified.

          NOX emissions  when burning  a
          nitrogenous waste was  compared
          with and without  combustion
          control.

          Heavy metals emission  data  were
          collected.

          Chloride mass  balance  calcula-
          tions showed greater  than  100%
          closures  (127% for Boiler  No.  1
          and  108% for Boiler No.  2).
                                            66

-------
Boiler No. 1, which is used to solve a
specific waste disposal problem without
large fuel savings, was successful in
destroying the plant's wastes at the
concentrations fed to the boiler.  Boiler
No. 2, which capitalizes on various waste
disposal opportunities to accomplish
significant fuel savings, realized accept-
able DEs except for xylenes for relatively
high concentrations of POHCs .in the feed.

Destruction Efficiency

     Destruction efficiency (DE) was
determined by analyzing for POHCs of
interest in the feed and in the emissions,
calculated as follows:

     DE = 100 (POHCin - POHCout)/POHCin

The POHCs include the components of the
standard test mixture and the major
hazardous compounds in the waste feed.  DE
requirements of 99.99% were exceeded for
all compounds except xylenes (Table 3).
Better than 99.99% DE was accomplished at
POHC levels in the feed of 16% (Boiler
No. 1) and 20% (Boiler No. 2).
                                         Correlation  with  Boiler  Operation

                                              The test programs provide  an
                                         opportunity  to look for  a  correlation  of
                                         easily measured boiler parameters  with
                                         destruction  efficiencies.   The  purpose is
                                         to develop indicators that could  sub-
                                         stitute for  the time consuming  and costly
                                         destruction  efficiency tests.   Two of  the
                                         indicators often  mentioned are  carbon
                                         monoxide (CO) and total  hydrocarbon con-
                                         centrations  in the flue  gas.   CO  and total
                                         hydrocarbon  (THC) concentration did not
                                         correlate with destruction efficiency  at
                                         the low levels of emissions concentrations
                                         encountered.  A possible explanation is
                                         that even the low levels of CO  and THC
                                         represent orders  of magnitude  higher
                                         concentrations than the  concentrations of
                                         POHCs in the, emissions.   For example;  the
                                         highest CO concentration observed, 242 ppmv,
                                         reflects a feed carbon to  unoxidized
                                         carbon ratio of 430 to 1 whereas  a 99.99%
                                         destruction  efficiency of  a POHC  represents
                                         a 10,000 to  1 ratio of feed POHC  to
                                         emissions POHC.
                       TABLE 3.   DESTRUCTION EFFICIENCY
         POHC
                                  Boiler No.  1
                                   Staged
                                 combustion
 Unstaged
combustion
Boiler No. 2
Standard test mixture
  Carbon tetrachloride       99.997
  Trichloroethylene          99.999
  Monochlorobenzene          99.998
  Toluene                    99.997
                                                   99.997
                                                   99.999
                                                   99.998
                                                   99.998
                  99.999
                  99.999
                  99.999
                  99.999
Liquid waste
Nitrobenzene
Phenol
Aniline
o-Xylene
m&p-Xylene

99.999
not detected
99.998
not detected
not detected

99.999
not detected
99.998
not detected
not detected

not detected
99.999
not detected
99.958
99.947

Values beyond three decimal places were ignored.
                                      67

-------
Products of  Incomplete Combustion  (PICs)

     PICs were found  in the Boiler No. 2
emissions and consisted of substituted
phenols (methyl butyl phenol, dimethyl
phenol).  Their presence  is probably due
to the high  concentration of phenol in the
boiler feed.  The substituted phenols are
not the XAD-2 artifacts that were  discussed
in the Problems Encountered section.

NOx Control  When Burning Nitrogenous Haste

     Gas fired Boiler No. 1 had been
modified to  accommodate cofiring of a
mixture of nitrobenzene, aniline and
benzene with fuel gas.  Included in the
modifications were changes in air  distri-
bution to the burner to provide staged
combustion NOx control.  Testing confirmed
that staged  combustion was effective in
controlling  NOx emissions (Figure  1).  NOX
emissions were 73% less than the equivalent
organic nitrogen present in the liquid
waste.  Organic nitrogen compounds were
not detected in the fuel gas.  Corrected
for background emissions, which are all
derived from combustion air, estimated NOx
emissions were 78% less than the organic
nitrogen in  the feed.  Without staged
combustion,  organic nitrogen was lowered
by only 24%  or by an estimated 28% if
corrected for baseline emissions.

Fate of Metals

     Particle catches were analyzed for
specific heavy metals.  Particulate
matter emissions from Boiler No. 1 were
quite low and specific metal concentrations
exceeded minimum detectable limits only
slightly if  at all.  Also, differences
between liquid waste runs and background
runs v/ere not distinguishable for most
metals.  Mercury and lead were the only
exceptions.  It was concluded that heavy
metals emissions from Boiler No. 1 were
not a significant problem.

     Boiler  No. 2 feed streams, both the
No. 6 fuel oil and the waste solvents,
contained significant amounts of several
of the heavy metals.  Strainers were used
to remove bulk quantities of solids from
the boiler feed streams.  This is usual
practice with fuels of the type used.
Thus, metals concentrations in the filtered
feed streams emulate typical practice.
The two feed streams were examined for
fourteen metals.  Nine metals were found
in excess of one ppmw (Table 4).  Signif-
icant levels of vanadium, nickel, zinc,
lead and cobalt were found in the baseline
fuel (No. 6 fuel oil).  Substantial
quantities of vanadium, zinc and lead in
the light oil added to the potential for
having significant concentrations of these
metals emitted to the atmosphere.  Also,
the light oil contained a significant
level of manganese.  Since about one third
of the total boiler feed during the waste
disposal runs was baseline No. 6 fuel oil,
it was of interest to know whether there
were significant differences between
baseline emissions and waste firing
emissions.  Waste firing resulted in
higher emissions of zinc, lead, cadmium,
copper, manganese, chromium and cobalt.
Vanadium and nickel were the only metals
in significant concentrations in No. 6
fuel oil that did not show an increase.
Of the metals showing an increase, con-
centration ranges during waste firing were
as follows:
     zinc:

     lead:

     cadmium:

     chromium

     copper:

     manganese:

     chromium:

     cobalt:

ACKNOWLEDGEMENTS
>1000 yg/m3

>1000 yg/m3

100 - 1000 yg/m3

10 - 100 yg/m3

10 - 100 yg/m3

10 - 100 pg/m3

10 - 100 yg/m3

10 - 100 yg/m3
     The authors gratefully acknowledge
the cooperation of the management of the
host facilities for providing boilers to
test and for donating employee time to
help coordinate the effort.  This project
is sponsored by the U.S.  EPA Industrial
Environmental  Research Laboratory -
Cincinnati.  Technical project monitor for
EPA is Mr.  Robert Olexsey.

REFERENCES

1.   Olexsey,  R.A., 1983.   Incineration of
     Hazardous Waste in Power Boilers:
     Emissions Performance  Study Rationale
     and Test Site Matrix.   U.S. Environ-
     mental Protection Agency, Cincinnati,
     Ohio.
                                            68

-------
2.0-
.c
o
.a
c
0)
CO
'1- 1.0-
UJ

u
vvwwywww
MB
BwfflsBS
mm
mm
mm*
nstage
^  Fuel Bound N^
NOX
Emissions
iJ&J&tefiS*
V* '&**<[ t W
sd Combustion
Figure 1. Boiler No. 1 f
TABLE 4. TRACE [*
mm,
w
Staged
JO con
A
1ETALS
I
NOX
Emissions
Emissions
Combustion Background
trol performance.
ANALYZED

Greater than Greater than
one p'pm in , one ppm in
feed? feed?
Zinc Yes Chromium Yes
Lead Yes Cobalt Yes
Vanadium Yes Arsenic No
Nickel Yes Mercury No
Cadmium Yes Antimony Mo
Copper Yes Selenium Mo
Manganese Yes Strontium Mo
69

-------
                       SUMMARY OF FIELD TESTS FOR AN INDUSTRIAL BOILER
                                DISPOSING OF HAZARDOUS WASTES

                                     John T. Chehaske
                                   Engineering-Science
                                  10521 Rosehaven Street
                                 Fairfax, Virginia 22030
                                           and
                                    Gregory M. Higgins
                                   SYSTECH Corporation
                                  245 North Valley Road
                                    Xenia, Ohio 45385
                                         ABSTRACT

     This paper presents the results of field tests of a North American Model 3200X
package fire-tube boiler burning a mixture of toluene and chlorinated solvents.  Two
fuel blends were tested.  Blend #1 contained 98 wt % toluene, 1 wt % trichloroethylene,
and 0.5 wt % each of chlorobenzene and carbon tetrachloride.  Blend #2 differed in that
the trichloroethylene concentration was increased to 2 wt % and the toluene concentration
was reduced to 97 wt %.  Flue gas samples were collected with the VOST, Modified Method
5, and Method 23 sampling systems.  The VOST and Modified Method 5 samples were analyzed
by GO/MS while the Method 23 samples were analyzed by on-site GC.  Continuous analyzers
were used to measure the oxygen, carbon monoxide, and nitrogen oxides content of the
flue gas.
INTRODUCTION

    The U.S. Environmental Protection
Agency (EPA) regulates the destruction of
hazardous wastes in incinerators requir-
ing a Destruction and Removal Efficiency
(ORE) of 99.99 percent (%) for the Prin-
cipal Organic Hazardous Constituents
(POHCs) that are present in the waste.
Particulate matter emissions are limited
to 0.08 grains/dry standard cubic foot
(gr/dscf) corrected to 12% carbon diox-
ide (C02), and hydrogen chloride (HC1)
emissions are limited to 4 pounds per
hour (Ib/hr) or 1% of the uncontrolled
HC1 whichever is greater.  Processes
which primarily produce energy and coin-
cidentally burn hazardous wastes are cur-
rently exempt from hazardous waste des-
truction regulation.  This includes pro-
cesses such as industrial boilers, kilns,
blast furnaces, process heaters, and
ovens.
    The EPA is investigating the need for
regulating hazardous waste combustion in
thermal industrial processes.  A Regula-
tory Impact Analysis (RIA) is being con-
ducted to determine the need for and ef-
fects of regulating these processes.
Part of the input needed for the RIA is
information on the ability of these pro-
cesses to destroy hazardous wastes.

    This paper presents data obtained
from tests of a fire-tube boiler burning
100% hazardous waste.  The project was
sponsored by the EPA Office of Solid
Waste (OSW).  The OSW Project Manager was
Mr. Marc Turgeon.  The project was con-
ducted under contract from the EPA Indus-
trial Environmental Research Laboratory
(IERL) in Cincinnati, Ohio, who provided
the technical direction.  Mr. Robert A.
Olexsey was the IERL Technical Manager.
Engineering-Science (ES) was the prime
contractor for the project and was
                                          70

-------
 responsible for overall project coordina-
 tion;  collection of the Volatile Organic
 Sampling Train (VOST),  Modified Method 5
 (MM5),  and HC1 samples;  and report prepa-
 ration.

     ES  was assisted by  SYSTECH Corpora-
 tion, who was  responsible  for  obtaining
 the test site,  coordinating-with the
 plant,  and conducting EPA  Proposed Method
 23  (M23)  sampling.   Analyses of the VOST
 and MM5 samples were performed by Re-
 search  Triangle Institute.
PURPOSE

    This  test was one of a matrix of
about  14  tests  conducted by IERL to ob-
tain DRE  data for industrial boilers.
The matrix  included a wide range of
boiler sizes and types.  This particular
test was  to provide DRE data for a small
fire-tube boiler.  Most of the other
tests involved  co-firing of hazardous
waste with  the  normal boiler fuel.  This
test, however,  was designed to determine
DRE for 100% hazardous waste burning, a
practice  that is known to be used by some
facilities.

    There was a secondary purpose of this
test program:   to compare the results ob-
tained with VOST and MM5 to those ob-
tained with M23.

APPROACH

Boiler Description

    All field testing was accomplished on
a North American Model 3200X package fire-
tube boiler.  This unit is referred to as
an "Ohio  Special" because it has only 358
ft2 of flue gas to water heat exchange
surface,  and therefore does not require a
full-time licensed operator.   "Ohio Spe-
cials" probably represent a worst-case
situation for destroying hazardous wastes
in fire-tube boilers.  With a high heat
release of  92,000 Btu/hr ft2 and a high
liberation  rate of 164,400 Btu/hr ft3,
it is expected  that high combustion tem-
peratures and very short retention times
occur in  the furnace of the test unit.
However, ..the nonconservative  design of
an "Ohio Special" may also have some
 advantages when well controlled combus-
 tion is  desirable.   It was  discovered
 that the unit used  in these tests  has a
 very narrow range of excess air levels
 within which it could operate.   Below the
 minimum  used in the tests  (17%  excess
 air),  combustion was incomplete resulting
 in smoke formation.   Above  the  maximum
 excess air used in  the tests (53%)  the
 fire would blow out the back of the com-
 bustion  chamber and the system  would au-
 tomatically shut down.   Furthermore,
 since this type of  boiler is kept  at a
 positive furnace pressure,  a large  change
 in excess air between minimum firing irate
 and maximum firing  rate will not occur as
 it would in a negative pressure furnace.

     The  burner of the test  unit is  de-
 signed to fire on either gas or distil-
 late oil.   For normal operation, natural
 gas is fired.   Therefore, the fuel  oil
 firing system was available to  be
 cleaned,  adjusted, and  used to  fire test
 blends of liquid chemicals  without  inter-
 rupting  normal operation of the boiler.
 The fuel oil system  consists of a 1000-
 gal cylindrical tank, a pump with mecha-
 nical  Teflon seals,  an atomizing air
 compressor,  and a spray nozzle.  The
 storage  tank is  plumbed to  a nearby tank
 farm to  permit filling  the  1000-gal tank
 from long-term storage.  Mixing is  pro-
 vided  in the storage  tank by a  recycle
 loop from the  boiler.

 Test Fuel

     To provide  control  of fuel  composi-
 tion and assure  consistent purity in the
 materials burned for  all test conditions,
 virgin chemicals  were purchased and
 blended  to simulate a hypothetical hazar-
 dous waste.  Reagent  grade carbon tetra-
 chloride, technical grade monochloroben-
 zene,  and technical trichloroethylene
 were carefully weighed to the desired
 amounts and added into the fuel tank.
 The  tank was then topped off with nitra-
 tion-grade toluene (>99% pure).   Fuel was
usually blended  the  evening before each
 test day so that the chemicals could mix
 overnight by recirculating through the
fuel pump circuit.  Table 1  shows the
composition of the two test fuels fired
in this program.
                                           71

-------
  TABLE 1.  AVERAGE POHCs COMPOSITION OF
                TEST FUEL
               Carbon
Fuel           tetra-  Trichloro- Chloro-
blend Toluene chloride  ethylene  benzene
 no.  (wt %)   (wt %)    (wt %)    (wt %)
1
2
98
97
.5
.5
1
2
.5
.5
Test Matrix

    Testing was conducted at six differ-
ent boiler operating conditions.  Table
2 summarizes the test matrix of fuel
blend and boiler load/excess air (EA).
Triplicate samples were collected with
each test method at each test condition,
except condition 6 for which only dupli-
cate samples were collected.

          TABLE 2.  TEST MATRIX
Test
condition
no.
1
2
3
4

5

6

Fuel
blend
no.
1
1
2
1

1

2

Boiler
conditions
half load
full load
half load
full load
(high EA)
half load
(high EA)
full load
(load EA)
Procedures

    The two volatile organic components
of the fuelr carbon tetrachloride  (CC14)
and tricoloroethylene  (TCE) were sampled
and analyzed according to the VOST pro-
tocol  (1).  Each test  run consisted of
three pairs  (one Tenax-GC and one
Tenax-GC/charcoal tube) except test 16
which consisted of six pairs and test 17
which consisted of four pairs.  Sampling
was conducted  at 0.5 liters per minute
(1/m).  The volume sampled by each pair
of tubes was 20 liters (1).  VOST  ana-
lytical results were subjected to  the
following acceptance criteria:
    o  Both surrogate compounds  must have
       been detected.
    o  The fractional recovery of both
       surrogates must have been between
       0.5 and 1.5.
                  OR
    o  The fractional recoveries of both
       surrogates must have been less
       than 5.0 and the two fractional
       recoveries must have agreed within
       25%.

      (recovery 1 divided by recovery 2
          is between .75 and 1.25)

    The semi-volatile organic compounds,
toluene and monochlorobenzene (MCB), were
sampled and analyzed by the MM5  procedure
(2).  Analysis of the resultant  extracts
was by GC/FID.

    In addition to the VOST and  MM5 sam-
pling, the POHCs in the flue gas were de-
termined by M23 (3).  This method, shown
schematically in Figure 1, employs inte-
grated bag sampling followed by  gas chro-
ma tography.
        FLOW
        METER
           \
                                     PUMP
  Figure 1
Schematic Diagram, Method 23
  Sampling Equipment
                                           72

-------
    The gas samples were collected in
Tedlar bags at a flow rate of 0.5 1/min
over a 40-min interval simultaneously
with the first leg of VOST sampling.
Prior to testing, all Tedlar bags were
assembled, filled with ultra-pure nitro-
gen, leak checked, and tested by GC for
background contamination.  Toluene and
chlorobenzene were analyzed by FID, while
BCD was used for the remaining chlori-
nated POHCs.  The sampling systems were
also assembled and leak checked.  Field
sampling for POHCs consisted of two ba-
sic functions: preparation and analysis
of calibration standards and flue gas
sampling.  Both functions were performed
as described in M23.  Spikes and repli-
cate analyses were included in the flue
gas analyses.
PROBLEMS ENCOUNTERED

VOST

    An unusually large number of the VOST
tubes were broken in shipment from the
laboratory to the field.  The breaks oc-
curred at the Swaglok fittings, indicat-
ing the fittings had been overtightened,
had bottomed out on the tube, or the tube
diameter was too large.  Two tubes were
broken on return to the laboratory, re-
sulting in lost samples.  There were pro-
blems with the GC/MS that resulted in two
more lost samples, and there were four
tube pairs that were lost by the labora-
tory.  These lost samples combined with
the data that were excluded because of
poor surrogate recoveries resulted in a
total of 21 tube pairs that were ex-
cluded out of a total of 55 tube pairs
collected.
enough to verify an average DRE of 95.64%
when MCB was not detected, and MCB was
not detected in any of the 18 test runs.
RESULTS

    The boiler performed very well
throughout the testing program.  Only a
few minor delays were encountered when
steam production exceeded demand.  Table
3 lists the average carbon monoxide (CO)
and nitrogen oxides (NOX) concentrations
for each of the six test conditions.
With the exception of condition 3 and 6
the CO results track boiler conditions as
expected.  For condition 3 and 6 the CO
concentrations were lower than expected.
These were the two runs with the highest
fuel chlorine content, but it does not
seem likely that the additional chlorine
would have depressed the CO concentra-
tions.  The NOX concentrations tracked
boiler conditions as expected, although
the NOX concentration for condition 6
was lower than expected.
       TABLE 3.  CO AND NOX RESULTS


Condi-
tion
no.
1
5
3
2
4
6
Boiler
load/EA (%)
half/38
half/50
half/31
full/38
full/41
f ull/1 7
Chlori-
nated
Solvent
(%)
2
2
3
2
2
3


CO
(ppm)
108
89
10
142
86
20


NOX
(ppm)
172
130
158
82
75
84
Modified Method 5

    The analytical detection limit for
MCB was four times higher than had been
expected by the laboratory.  As a result,
for those runs where MCB was not de-
tected, the maximum calculable DRE was
less than 99.99%.  There were 11 of 17
runs when MCB was not detected.

Method 23

    Similar detection limit problems were
encountered for MCB with the M23 proce-
dure.  The detection limit was only low
Destruction Removal Efficiencies

    DREs were calculated from measured
fuel feed rates and compositions, flue
gas flow rates, flue gas molecular
weights, POHC flue gas concentrations,
and POHC molecular weights.

    Tables 4 and 5 show the calculated
DREs obtained from the VOST and MM5 sam-
pling, respectively.  The values are the
mass weighted averages for each, boiler
test condition.  For toluene, carbon te-
trachloride, and trichloroethylene the
                                            73

-------
DREs  ranged from a  low of 99.9976 to a
high  of 99.9999.  The MCB results were
lower, primarily because of the high
analytical limit of detectability men-
tioned previously.  The reported MCB
DREs  were calculated by inserting the
minimum detectable  quantity for the 11
runs  where no MCB was detected.  There-
fore, the reported  MCB DREs are lower
than  the actual DREs.

          TABLE 4.  DREs BY VOST
Condition
no.
1
3
5
2
4
6
Boiler
load/EA (%)
half/38
half/31
half/50
full/38
full/41
f ull/1 7
CC14
DRE (%)
99.9979
99.9990
99.9993
99.9989
99.9998
99.9992
TCE
DRE (%)
99.9999
99.9995
99.9976
99.9998
99.9999
99.9996
          TABLE 5.  DRES BY VOST
Condition
no.
1
3
5
2
4
6
Boiler
load/EA (%)
half/38
half/31
half/50
full/38
full/41
f ull/1 7
Toluene
DRE (%)
99.9997
99.9992
99.9991
99.9991
99.9994
99.9991
MCB
DRE (%)
99.952
99.978
99.972
99.946
99.948
99.973
    For the M23 results, there were only
four test runs in which POHCs were found
in the flue gases.  In the cases where no
POHCs were detected, the analytical detec-
tion limit was used to calculate the DRE.
Therefore, the actual DREs occurring in
the boiler were at least as great or
greater than the calculated DREs.  With
the exception of MCB, the M23 analytical
procedures were sensitive enough to sub-
stantiate at least 99.99% DRE.  Because
most of the DREs were based on detection
limits, the individual values by test
condition are not given.  Instead, Table
6 shows the number of runs for which DREs
greater than and less than 99.99% were
verified.

          TABLE 6.  DREs BY M23


POHC
Toluene
CC14
TCE
MCB
No. of
cases DRE
>99.99%
17
17
12
0
No. of
cases DRE
>99.99%
l(a)
0
3(b)
18
                                                   measured DRE v 99.98
                                                   average measured DRE = 99.46
    Table 6 illustrates the average DREs
calculated for each of the four POHCs by
the M23 procedure.  The DREs reported in
Table 6 illustrate that the EPA Method
23, coupled with GC/FID-ECD analysis, was
sufficient to subtantiate DREs of at
least "four nines" occurring in the test
boiler for toluene, CC14, and TCE.  Be-
cause of the low instrument sensitivity
to MCB, however, this method could not
provide adequate sensitivity for DRE mea-
surement of this POHC under the given
test conditions.  The best DRE calculable
for MCB under these conditions averaged
95.64.

Methods Comparison

    A direct comparison of the POHC con-
centrations measured by the various me-
thods is not possible because only four
of the 72 determinations by M23 yielded
detectable POHC quantities.  It is pos-
sible, however, to compare the DREs ob-
tained with the three methods as shown
in Table 7.  The test results for all six
test conditions was combined to give a
mass weighted average DRE for each POHC.
The VOST and MM5 test procedures yielded
higher DRE values for each POHC than did
the M23 procedure.  The reason that the
DRE values are higher for VOST and MM5
than for M23 is because VOST and MM5 had
lower limits of detectability.  With the
exception of MCB, VOST and MM5 were able
to measure actual flue gas concentrations
and thus actual DREs could be calculated.
                                            74

-------
Method 23 could only provide information
that the ORE was at least as great as the
calculated value, based on the higher li-
mit of detectability inherent with M23.

        TABLE 7.  DREs COMPARISON
Compound
Toluene
CC14
TCE
MCB
DRE by
VOST or MM5
99.999
99.999
99.999
>99.961
DRE by
Method 23
>99.998
>99.954
>99.995
>95.642
It is important to emphasize that the M23
DREs reported in Table 7 are more a re-
flection of limitations in analytical
methods sensitivity than of the ability
of the test boiler to destroy or remove
the test POHCs.  The test boiler likely
destroyed or removed a larger percentage
of the input POHC load than is reported.
REFERENCES
1 .  Protocol for the Collection and An-
    alysis of Volatile POHCs Using VOST,
    Envirodyne Engineers, Inc.,
    St. Louis, Missouri, August 1983.

2.  Sampling and Analysis Methods for
    Hazardous Waste Combustion (First
    Edition, Arthur D. Little, Inc.
    Cambridge, Massachusetts, February
    1983, pg 36.

3.  Federal Register, 45 FR 39766,
    October 3, 1980.
ACKNOWLEDGEMENTS

    This project was conducted under Con-
tract No. 68-03-3040, Work Assignment No.
SBE04 for the U.S. Environmental Protec-
tion Agency, Industrial Research Labora-
tory, Incineration Research Branch,
Cincinnati, Ohio.

    The authors wish to thank Mr. Jon N.
Bolstad of Engineering-Science and Mr.
Ned J. Kleinhenz of SYSTECH Corporation
for their diligent efforts as field team
leaders for the two test crews.
                                           75

-------
                     PARAMETRIC EXPERIMENTATION WITH A  PILOT-SCALE
                           BOILER BURNING  HAZARDOUS COMPOUNDS
                                     C. D. Wolbach
                                   Acurex Corporation
                            Mountain View, California  94042


                                       ABSTRACT

     Thermal destruction of wastes by direct incineration or by cofiring with
conventional fuels in boilers, furnaces, or kilns is one of the most effective methods
currently available for disposal of hazardous organic material.  While direct
incineration of hazardous wastes is regulated by Part 264 of the Resource Conservation
and Recovery Act  (RCRA) as adopted in January 1981, boiler cofiring is currently exempt
from RCRA provisions.  However, the potential for boiler cofiring regulations has been
evaluated by the Environmental Protection Agency (EPA) and they are now in the process
of preparing regulatory positions in drafting regulations for promulgation.  To support
this effort, EPA's Incineration Research Branch (IRB) in conjunction with the Office of
Solid Waste is conducting research and development programs on incineration
effectiveness and regulatory impacts.

     The global purpose of this study was to gather data to aid the EPA in selecting a
strategy or set of strategies for regulating the combustion of hazardous wastes in
boilers.  The specific objectives were twofold:  to identify which of several boiler
operational parameters have a major impact on boiler destruction and removal efficiency;
and, to evaluate and if practical, establish a mathematical model  for predicting an
upper limit on the amount of cofired waste that could be emitted.   In particular, those
parameters that could be easily changed by an operator or might represent major
differences between boiler types were studied.  A secondary objective of the study was
to gain sufficient information to allow judgements of what particular parameters to
monitor closely during full-scale testing.
GENERAL APPROACH AND TEST MATRIX

     The experimental goals of the
program, in sequence, were to:

        Determine the furnace operating
         parameters which affect the
         furnace three-dimensional
         time-temperature environment

        Quantify these effects

        Identify the relationships
         between operating parameters,
         time-temperature history, and
         destruction of hazardous waste
         components in the fuel
     The test plan divided the tests
into three subsets:  a baseline study
burning only distillate oil, a single
compound study of one hazardous waste
compound cofired with distillate oil,
and a multiple compound study burning
more than one compound simultaneously.
The baseline study was designed to
define the thermal environment within
the facility under the various operating
conditions of choice, to define the
magnitude of the effects of changes in
variables on the thermal environment,
and to demonstrate the ability of a
simulation to predict the thermal
environment resulting from a specific
set of variable settings.  The single
                                            76

-------
compound studies had two goals:  to
define the magnitude of variable effects
on destruction efficiency; and to define
the effects of cofiring wastes on the
expected thermal environment.  The final,
multiple compound studies were to
determine the destruction efficiencies of
several compounds with known thermal
environments.  Chlorinated wastes were
selected for sensitivity and ease of
analysis.  Chlorobenzene was selected for
the single compound tests because of its
high heat content, compatibility with
distillate oil, and simple structure.
For the multiple compound tests
chlorobenzene was fired with carbon
tetrachloride, chloroform, methylene
chloride, and dichloroethane.

     Within each task a set of
independent variables were identified
that are relevant to the operation of
incinerators and boilers.

     Table 1 shows the test matrix of
nominal operating conditions used in the
program.  Firing rate and stoichiometry
parameter values are listed at nominal
conditions based on analysis of the pure
distillate oil fuel.  As shown, two
nominal firing rates (0.2 and 0.38
megawatts) and three nominal
stoichiometries  (10, 25, and 50 percent
excess air) were used.

     The highest concentration of waste
to fuel was achieved with chlorobenzene
(10 percent v/v).  Other materials were
        injected  at  levels  of  approximately  1.5
        to  3.5  percent  (v/v).

            A total  of  40 tests  were  run
        generating  99 individual  data points.
        Fourteen  firing conditions  were studied
        and the destruction and  removal
        efficiency  relative to four compounds  was
        examined.  A computer  simulation  was
        prepared  that predicted  the thermal
        profile within  25C and  the destruction
        and removal  efficiency within two orders
        of'magnitude.

        FACILITY  DESCRIPTION

             All  tests  were performed in  the
        Acurex  Pi lot-Scale  Furnace.  The  furnace
        has a single burner, front  wall fired,
        into a  horizontally oriented  firebox.
        The premixed fuel oil/waste mixture  is
        pumped  out  of drums through a
        pressure-atomizing  nozzle and stabilized
        at  the  front wall.   The  fuel  flow is
        monitored by rotometer.   Combustion  air
        is  preheated and injected in  the  annular
        region  around the fuel delivery tube.
        The IFRF/Acurex burner design allows
        swirl adjustment by rotating  swirl
        blocks.  Air flow  is monitored by hot
        wire anemometer.

             The  walls  of the  radiant section  are
        refractory  lined with  optionally  mounted
        waterwalls.  Ports  along the  top  and
        sides of  the furnace accommodate  ceramic
        type R  thermocouple probes  for gas
        temperature'profiles'.  Type K probes are
                 TABLE 1.  TEST MATRIX OF VARIABLES AND NOMINAL VALUES
           Variable
                   Values
     Firing  rate  (load)
0.23, 0.35 _+ 0.03 megawatts (0.8, 1.2 +_ 0.10
million Btu/hr)
     Flame shape  (swirl)               7.5, 5.0, 2.0

     Excess air rate  (stoichiometry)   10, 25, 50 percent, +2 percent

     Waste to fuel  ratio

     Waterwall area

     Compounds
10, 3, 1.5 percent

0, 8 percent

4
                                            77

-------
 embedded in  the refractory  for wall
 temperature  determinations.
 Thermocouples  are also placed at the
 manifold inlet to and the outlet of each
 cooling panel, and to and from the burner
 quarl.  Cooling fluid flow  measurements
 are by rotometer.  Downstream of the
 convective a stainless steel  probe
 extracts gas samples for continuous
 emissions analysis.  Electro-optical
 analyses are used to define  levels of CO,
 C02, 02, NOX,  and total unburnt
 hydrocarbon  (THC).  Samples  for the
 Volatile Organics Sampling  Train (VOST)
 are extracted  with a stainless steel
 probe from the end of an 8-in. diameter
 15-ft long duct.  The gases  are then
 exhausted to a 45-ft high stack.

 RESULTS AND  DISCUSSION

      Table 2 provides a complete summary
 of the program results.  The  roman
 numeral in the first column  is an
       identifier for the  run condition.   The
       run  condition is  summarized in  the second
       column.   The first  item in parentheses
       is the nominal firing rate in million
       Btu/hr.   The second and third items are
       the  percent excess  air and the  swirl
       setting,  respectively.  The final item is
       the  percent of the  radiative section
       surface  area covered by water walls.   The
       number in the compound columns  represent
       the  average penetration value,  C/C0,
       (mass  of  waste out  per unit, time/mass
       rate of waste in),  the value of one
       standard  deviation,  and (in parentheses)
       the  number of data  points.  The
       relationship to ORE is:  ORE =  1-C/C0.
       The  effects of the  various firing
       conditions of DRE's  are discussed  in  the
       following sections.   A semiquantitative
       summary  on the variable effects  is given
       in Table  3.  These  values are rough
       estimates of the  difference between
       either temperature  or ORE for two  runs
       where  only the specified parameter was
    TABLE 2.  AVERAGE FRACTIONAL BREAKTHROUGH C/Cn FOR  EACH WASTE AND  FIRING CONDITION
      Condition
                        Chlorobenzene*,t
   Carbon
tetrachloride
                     Chloroform
                                     1,2 Dichloroethane
  I*  (0.8,  10, 7.5, 8)   1.4  0.1 x 10"* (4)  1.6  0^2 x 10'4  (2)  1.6  0.2 x  10"5 (3)    5  2 x

  Ia (0.8,  10, 7.5, 0)   4.2 x 10~6 (1)               i
 II   (0.8, 25, 7.5, 8)   1.4  0.3 x 10-5 (4)   5  3 x HT5 (5)

 Ha  (0.8, 25, 7.5, 0)   4  3 x 10-7 (6)

III   (0.8, 50, 7.5, 8)   8.9  6 x 10-5 (5)    2>1 x 10-4 (1)

Ilia  (0.8, 50, 7.5. 0)   2.5  2 x 10-7 (3)

 IV   (1.2, 10, 7.5, 8)   6  8 x 10-5 (5)

 IVa  (1.2, 10, 7.5, 0)   2.2  0.5 x 10~7 (3)

  V   (1.2, 25, 7.5, 8)   4.5  2 x 10-6 (5)

  Va  (1.2, 25, 7.5, 0)   0.5  0.3 x 10-7 (3)

 VI   (1.2, 50, 7.5, 8)   5.5  2 x 10-5 (4)

 Via  (1.2, 50, 7.5, 0)   3.4  0.7 x 10-7 (3)

VII   (0.8, 25, 5.0, 8)          -           6.3  3 x 10-5  (3)

 IX   (1.2, 25, 5.0, 8)   2.0  0.5 x 10-6 (3)   1>5  0>3 x 10-4 (3)
                                            ID"6 (3)



                                     1  0.5  x lO-6 (2)



                                     1  0.2  x HT6 (3)
                   2.6  2 x ID'5 (5)



                   2.0  0.8 x ID'5  (3)



1.2  0.3 x 10-* (2)   3.2  0.5 x 10'5  (3)



2.5  1 x ID'5 (2)    8.8  2 x 10'5 (3)



7.9  6 x lO-5 (3)    1.1  0.02 x 10'4 (2)   8  4 x 1Q-6 (2)
* t represents 1 standard deviation from average.
t ( ) Indicates number of valid data points.
+ Firing rate 1n million Btu/hr, percent excess air, swirl, percent water-wall.
 Run conditions suffixed with an "a" are without waterwalls. All other conditions are with
  water-walls 1n place.
I Not tested under that condition.
                                              78

-------
TABLE 3.  SEMIQUANTITATIVE ESTIMATION
          OF THE EFFECTS OF THE
          INDEPENDENT VARIABLES ON
          TEMPERATURE PROFILE AND DRE



Variable
Waterwalls
Excess air rate
Firing rate
Burner swirl
Compound
Range of
Temperature
Changes
C (F)
275 (500)
165 (300)
110 (200)
30 (50)
30 (50)
Range
of DRE
Changes
(Factor)*
100
10
3
1
20
*C/C0 is mass rate of waste out
 divided by mass rate of waste in.
varied.  For examplej a change in firing
rate between 0.8 million Btu/hr and
1.2 million Btu/hr changed the
temperature profile by about 110C
(200F).  The corresponding change in DRE
was about a factor of three.

     Thermal profiles were presented in
detail in a previous paper (1) and will
be discussed only incidentally to DRE
discussions.  The temperature profiles
clearly show a thermal boundary layer
within four inches of the furnace wall.
This boundary layer is significant, since
it represents greter than 40 percent of
the volume or thermal environment of the
furnace.  Because .temperature and
velocity are the dominant factors in
volume flowrate and destruction, this
zone will be significant when considering
destruction efficiencies in this furnace.

Effect of Waterwalls on Thermal
Environments and DRE's
     Addition  of waterwall panels
 replaced about '8 percent  of the
 refractory  surface area but extracted  50
 to  60 percent  of the heat  input.  With
 the waterwalls  in place the thermal
 environment of  the first  six  feet of the
 firebox was comparable to  a D-type
 packaged watertube boiler.  However, the
 pilot-scale waterwall surface temperature
 is  much lower  than for industrial boilers
 (600 to'800F).  The effect  will be to
lower DRE in the pilot unit with respect
to an industial boiler through wall
quenching..  This is offset by the larger
industrial boiler waterwall surface area
(>50 percent) compared to the pilot
unit.

     Waterwalls have a dominant effect on
DRE.  This is shown in Table 4.  For
chlorobenzene DRE varied over two orders
of magnitude with the introduction of
wall cooling.  Qualitatively, this was to
be expected' since the waterwalls
decreased the corresponding temperature
profiles 220 to 280C (400 to 500F).
Not surprisingly, the magnitude of the
effect is within the range predicted by
theory.  Using the Arrenius expression
for a rate constant, k = Aexp  (-Ea/RT),
and algebraically manipulating the
kinetic expressions to present the ratio
of breakthrough fractions for a given
waste under two different temperatures,
one obtains
= exp [E
                        - T2)/RT1T2]
Cj and ,C2 are the breakthrough quantities
at two different run temperatures
(assuming the same feedrate in), Ea is
the empirical pseudo first order energy
of activation for the disappearance of
the waste, R is the appropriate gas
constant, and Tj and T2 are the effective
temperatures.  Using effective
temperatures of 1,100 and 1,380K, an
activation energy of 50 kilocalories per
gram mole, and a gas constant of
1.982 calories per gram mole K, the
calculated ratio of breakthrough
quantities is 100.
 TABLE 4.  AVERAGE MEASURED PENETRATION
           C/C0 OF CHLOROBENZENE
              With Water    Without Water
Condition    Walls (x 105)  Walls  (xlO7)
(0.8,10,7.5)
(0.8,25,7.5)
(0.8,50,7.5)
(1.2,10,7.5)
(1.2,25,7.5)
(1.2,50,7.5)
14 + 1
1.4 + 3
8.9 + 6
6 +8
0.5 + 0.2
5.5 T 2
(4)
(4)
(5)
(5)
(5)
(4)
40
4 + 3
3 + 2
2 + 0.5
0.5 + 0.3
3 +_ 0.7
(1)
(6)
(3)
(3)
(3)
(3)
                                            79

-------
Effect of Excess Air on Temperature
Profiles

     The effects of excess ai r on
temperature profiles followed
expectation.  A comparison of Figures 1
and 2 shows effects of excess ai r on
centerline axial temperature profiles for
different firing rates.  An alternate
demonstration is given in Figure 3.
Here, radial temperature profiles at
two positions are shown for three excess
air levels at a nominal 0.8 million
Btu/hr firing rate.  All three conditions
are without waterwalls.  On the
centerline, gas temperatures vary as much
as 140C (250F) at cross section C-3
(approximately 80 in. from the burner
face) but only approximately 40C (70F)
at position E-2 (114 in. from the face).
The equivalent temperature difference at
the higher firing rate of 1.2 million
Btu/hr are on the order of 80C and 25C
(150 and 50)F), respectively.  The
lower difference at the higher firing
rate is caused by radiative effects from
the hotter flame zone radiating into the
cooler postflame zone.

     Qualitatively these excess air
results will have opposing effects on
destruction efficiencies.  At the higher
excess air level the increased oxygen
content will promote destruction while
the decreased residence time and
temperature will suppress destruction.
In fact, there appears to be a peak
destruction efficiency between excess air
rates of 10 and 50 percent.  This is
shown graphically in Figure 4.

Effect of Swirl on Temperature and ORE

     From the observed temperature
profiles, the effect of increased swirl
is to increase the temperature gradient.
However, the temperature changes only
about 15C (25F) at the centerline.
This is comparable to the measured
temperature variance between runs (see
Table 2), and is probably not
 significant.  There may  be a significant
 variation  in  DRE with  swirl, but the data
 is  difficult  to interpret.  A significant
 difference  is seen only  in the long, thin
 flame  (swirl  setting 2.0), but it is in
 the  opposite  direction for chlorobenzene
 and  carbon  tetrachloride.

 Products of Incomplete Combustion

     In order to maximize the number of
 variable data points, the analytical
 design for  this program  was simplified by
 not  performing GC/MS analyses.  This
 reduced the ability to identify and
 quantitate  PIC's.  Because the program
 was  not directed toward  PIC's, much raw
.data (such  as the numerous
 chromatographic strips)  that contains PIC
 information has not been reduced.
 However, even with limited examination of
 the  data some important  information is
 available.  For example, in every run
 methylene  chloride was-observed at levels
 equal to or greater than the POHC of
 interest,  even when it was not present as
 a POHC in the waste (Figure 5 displays
 typical chromatograms from the analysis).
 Although some of this methylene chloride
 can  be attributed to contamination, the
 levels were consistently higher than
 corresponding blank values.  Peaks in the
 retention time range for methylene
 chloride were routinely,  off scale.
 Several other chlorinated organics not
 fired in the fuel have been identified as
 being in the emissions of one or
 chlorinated organics not fired in the
 fuel have been identified as being in the
 emissions of one or more runs.  These
 include such  species as       
 trichloroethylene, the dichloroethylenes,
 chloroform, and possibly chloromethane.
 An example  is given in the summary
 Table 5.  The significance of this
 finding is  the following  although
 DRE's for both POHC's  (chlorobenzene and
 carbon tetrachloride) are greater than
 99.99 percent, the DRE of total
 chlorinated organics as fired is only
 99.985 percent.
                                           80

-------
 2,500
  2,400
  2,300
  2,200
  2,100
  2,000
           A-7- (1.2, 10, 7.5, 0)

           A-6 (V.2, 25, 7.5, 0)

          AB-4 (0.8, 10, 7.5, 0)

          TB-2 (0.8, 25, 7.5, 0)
                                                                            1.40T
                                                                            1,300
                                                                            1,200
                                                                            1,100
           1.68
                             2.90               3.70
                                  Axial Distance (m)
                                                                 4.47
 Figure  1.   Center-line Axial  Temperature Profiles  for  Four Firing
               Conditions Without  Waterwalls.
  2,000 -
   1,800
  1,400
   1,200
   1,000
         0.8 million  1.2 million Excess
          Btu/hr     Btu/hr   Air  -
Condition I  - O      IV -   (10%)
Condition II - D      V  -   (25%)
Condition III - A      VI - A  (50%)
                                                                             1,100
                                                                             i.ono
                                                                             900
                                                                             800
                       2.29      2.90 '          3.50
                                 Axial  Distance (meters)
                                                                  4.47
 Figure 2.   Center-line Axial  Temperature  Profiles for Six Firing
               Conditions with  Waterwalls.
                                        81

-------
 2,500
 2,400
 2,300
g 2,200
3
 2,100
 2,000
     H
     
 1,900J_
                Cross Section C-3
            rr^
                      o	o	Q
                                                  O Run B-5 Condition la 10% Excess air
                                                  D Run B-3 Condition lln 252 Excess air
                                                  A Run B-l Condition Ilia 50% Excess all
                                                      Cross Section E2
                                                                            % II
             10.16     20.32      30.48      40.64      30.48     20.32      10.16
                                Cross sectional distance (cm)
 Figure 3.   Effects  of  Excess Air on Temperature  for Three Firing
              Conditions  Without Waterwalls.   Normal  Firing  Rate  is
              0.8 million Btu/hr.
           ,-3
          10'
       I  in"1
       5  IS
          10'
           .-G
                    10
                            20
                                                         H1ll1on Btu/hr
                                                          (firing rate)
                                                           0.8   1.2
                                            Chlorobenzene      o    O
                                            Carbon Tetrachlorlde     
                                    30      40
                                  Excess air rate
                                                    50
                                                           60
                                                                   70
Figure  4.   Effect of  Excess Air on Fractional  Breakthrough  for
              Carbon Tetrachloride and Chlorobenzene.
                                      82

-------
                            TABLE 5.  SUMMARY OF RUN C-7

Mass flow
inn
Mass flow
out**
Out/in
Average
ORE
Chloro-
benzene
17.34

1.48 x 10~5

8.54 x lO-7
99.9999
+0.00016tt
Carbon
tetrachloride
9.72

8.32 x 10-4

8.56 x 10-5
99.9914
+0.0035
Chloroform*
0

1.22 x 10-3

N/A
N/A
.Subtotal1"
27.05

2.06 x 10-3

7.62 x 10-5
99.9924
jKJ.004
Other* Total 
0 27.05

2 x 10-3 4 x ID'3

N/A 1.48 x lO-4
N/A 99.985
+0.020
   Major identified  PIC.
 "" Includes  POHC's plus  chloroform.
 * Estimate  of unidentified  and/or unquantified PIC's.
  -f-stimate  of total  chlorinated  organic species.
 11  Estimate  of total  g/min.
   Average of three  samples  (g/min).
"ft Two standard deviation.    .  ,
,-Start

^__  10.32 CH2BrCl

r
f
f~
Lstop
. 13.13 CHCI3
' 13.94 1.2 OCA




      Standard 1,000 ng each on
      blank trap
                                              E
                                                Start
                                                             9.96 CH2 BrCI


                                                                  12.85 CHCI3
                                                Stop
Sample run C-7B carbon
tetrachloride
           Figure 5.  Typical  Chromatograms  Demonstrating  Presence of PIC's,
                                        83

-------
           PRODUCTS OF INCOMPLETE COMBUSTION FROM HAZARDOUS WASTE INCINERATORS
                                     Andrew Trenholm
                                     Roger Hathaway
                               Midwest Research Institute
                              Kansas City, Missouri  64110

                                      Don Oberacker
                          U.S. Environmental Protection Agency
                                 Cincinnati, Ohio  45268
                                        ABSTRACT

The Environmental Protection Agency (EPA) is preparing a Regulatory Impact Analysis
(RIA) of hazardous waste incineration.  Midwest Research Institute (MRI) provided input
data to the RIA from tests conducted at eight incineration facilities.  This paper pre-
sents information gathered during these tests related to products of incomplete combus-
tion (PICs).  PICs were defined as any hazardous organic constituent detected in the
stack gas but not present in the waste feed at a concentration of 100 |Jg/g or higher.
Benzene and chloroform were the compounds found most often among approximately 30 iden-
tified PICs.  PIC emissions were comparable to principal organic hazardous constituents
(POHC) emissions in concentration and total mass output.  Three mechanisms are proposed
to account for the presence of PICs:  (a) POHCs in the waste feed below detection lim-
its, subject to low destruction and removal efficiency (ORE), and therefore detected in
the stack gas; (b) introduction into the system from some source other than the waste
feed (e.g., auxiliary fuel, in-leak air, scrubber water); and (c) actual PICs (compounds
occurring as a result of combustion reactions).
INTRODUCTION

     Incineration is a popular option for
ultimate disposal of combustible hazard-
ous wastes.  While incineration provides
a high degree of destruction for com-
pounds in the waste feed streams, ques-
tions have been raised about possible by-
products of combustion reactions.  This
paper addresses some of these concerns
based on results of tests at full-scale
operating incinerators.

     Midwest Research Institute (MRI)
performed a series of tests at operating
hazardous waste incineration facilities
and analyzed the collected samples for
products of incomplete combustion (PICs).
These, tests were performed as part of
technical support to EPA's preparation of
a  regulatory  impact analysis (RIA) for
hazardous waste incinerators.

     Prior to  this  study by  MRI, the ac-
cumulated data base on products of incom-
plete  combustion  of Appendix VIII com-
pounds, principal organic hazardous con-
stituents (POHCs),  was limited primarily
to  laboratory  scale studies.  In those
studies, certain  specific compounds were
thermally destroyed in high concentra-
tions  to  test  for possible PICs.  Com-
pounds tested  include  kepone1  and  poly-
chlorinated biphenyls.2   These  studies
are illuminating  for specific cases, but
the data base for PICs from combustion of
the complex matrices of organic constitu-
ents which  currently are being  fed  to
                                           84

-------
hazardous waste  incinerators  was virtur
ally  nonexistent before  MRI' s  study.

     This paper  first describes briefly
the MRI  study  and identifies some  of  the
limitations of the  data base.   Then,  re-
sults  of the tests  conducted  are dis-
cussed in terms  of  three  possible  mecha-
nisms  to explain the presence  of PICs in
the stack gas, and the relative  contribu-
tion  of  each  mechanism.   The conclusion
section  contains some general  observa-
tions .   This paper  is intended  to be an
overview of the  PIC phenomenon.   It does
not provide  final definition  of mecha-
nisms  for PIC formation.
MRI STUDY OF HAZARDOUS WASTE INCINERATION
  FACILITIES

     PICs were  defined for this study as
any  Appendix VIII compound detected  in
the  stack  gas which was not present  in
the waste  feed in concentrations greater
than  100 (Jg/8-   Tne study included mea-
surement of  PICs at eight full-scale  op-
erating  incinerators with a variety  of
key  design  and operating parameters  and
waste  feed  characteristics.  Each incin-
erator had a  liquid  injection burner,  and
some  facilities also  included  a rotary
kiln  or  hearth.  Three incinerators  had
no  air pollution  control  devices.  The
remaining  five  had wet.scrubbers for  HC1
control, and  four  of these had  other  par-
ticulate control devices.

      Combustion chamber designs were  var-
ied.   Several  facilities had  multiple
combustion  chambers.  The  incinerators
ranged in  waste feed capacity  from 50 to
6,000  kg/hr  (110 to 13,000 Ib/hr), with
corresponding heat  inputs from 1  to  78
GJ/hr (1  to 74 million Btu/hr).  Operat-
ing  conditions  for  three key parameters
exhibited  wide  ranges:   combustion tem-
perature,  820  to  1450C (1400  to
2650F);  residence  times,  0.07 to 6.5
sec;  and percent excess air,  60 to 130%.

      The  facilities included some  exam-
ples  of  incineration of specific process
wastes generated onsite as well as  varied
mixtures of  wastes disposed offsite.   The
wastes fired at different facilities  var-
ied  also with respect  to  waste  type (liq-
uid,  solid,  and  gas), heating value,
chloride  concentration,  and  ash  and
moisture  content.   The organic liquids
and many  of the  solids  had a heating
value high  enough to  sustain combustion
without auxiliary fuel.   Percentages of
chloride  ranged  up to 25%, the highest
values  occurring  for  organic  liquids.
Ash values  ranged up  to 9% for organic
liquids and  29%  for solids.  Water con-
tent ranged  up to 50% for  organic  liquid
and solid wastes,  and to nearly 100% for
aqueous wastes.

     Samples were  collected of all inflow
and outflow  streams at the incineration
facilities.  All samples  were analyzed
for  Appendix VIII  constituents by gas
chromatography/mass spectrometry.  Lower
detection  limits for  compounds  in the
waste feed  ranged from 1 |Jg/g up to  100
pg/g, depending  on the variety and com-
plexity of  the  sample matrix.   The sam-
pling and  analysis procedures  used were
capable of  quantifying a wide variety of
compounds present in  the stack gas with
lower  detection  limits of 1-2 ng/L.

     Certain  compounds which  are sus-
pected of being  PICs were  not amenable to
measurement  by  the methods used in these
tests,  e.g., formaldehyde and phosgene.
Also, at  seven of the eight facilities,
carbon  tetrachloride  and trichloroethyl-
ene were  present in the waste  feed,  lim-
iting  their  identification as common
PICs.
PRESENCE  OF PICs  IN THE STACK GASAN
  OVERVIEW

     Twenty-nine  compounds were classi-
fied  as  PICs from the eight  incinerator
tests.   These  compounds are  shown  in
Table  1  along with  the  number of  sites
where  they  were detected as PICs  and the
range  of concentrations  measured in the
stack  effluents.   Figure 1 shows an al-
ternate  display of  the  data  categorized
as volatiles,  semivolatiles,  and trihalo-
methanes  (THMs).   The total  concentra-
tions  of  all PICs, and of  the three  cate-
gories of PICs are shown for each plant.
THMs  are not included in  the totals be-
cause  they  may  have originated in  the
liquid influent to the control systems at
some plants.

     Total  stack output  of Appendix VIII
compounds for the eight plants is shown
                                            85

-------
TABLE 1.  PICs FOUND IN STACK EFFLUENTS

PIC
Benzene
Chloroform
Bromodichlorome thane
Dibromochloromethane
Naphthalene
Bromoform
Chlo r ob enzene
Tetrachloroethylene
1,1, 1-Trichloroethane
Toluene
o-Nitrophenol
Methylene chloride
Phenol
2,4, 6-Trichlorophenol
Carbon disulfide
o-Chlorophenol
2,4 Dimethyl phenol
Methylene bromide
Bromo chlo r omethane
Trichlorobenzene
Hexachlorobenzene
Diethyl phthalate
Pentachlorophenol
Dichlorobenzene
Chloromethane
Methyl ethyl ketone
Bromomethane
Pyrene
Fluoranthene
Number
of sites
6
5
4
4
3
3
3
3
3
2
2
2
2
1
1
1
1
1
1
.1
1
1
1
1
1
1
1
1
1
Concentrations
(ng/L)
12-670
1-1,330
3-32
1-12
5-100
0.2-24
1-10
0.1-2.5
0.1-1.5
2-75
25-50
2-27
4-22
110
32
2-22
1-21
18
14
7
7
7
6
2-4
3
3
1
1
1
                   86

-------
11 * 	
Fll 1 g 	 ^ 	 r
11" v\ 	 	 >
\\
51
CN
O
to aU9|Dl|4l|dD|s] 	 . 	 .
(

O

of | [OUeildOJOIlpIJl-p'^Z




SI
*
o
1 	
* 
H-
X
2;

s
r
t/)
>
S
>
S
5
X
LO
S
to
S
3 4 5 678
V\ = Trihalomethanes - chloroform,
romethane (not included in other totals).
1 2
evice used.
, V = Volatile, SV = Semivolatile, TH/
n, bromodichloromethane, dibromochio
1-8 identify the plants.
   m
2 .S'S J
! 1 "1 
* U h- -D Z
* * *
* *
*
to
B
0
ri
4J
rt
4J
fi
0)
O
fi
0
O
O
M
t-)
0
M
CO
H
01
INCMO^tNCM                ^Sl
                                    87

-------
as a comparison between PICs and POHCs in
Figure 2.  In general, PIC concentrations
are slightly higher  than POHC  concentra-
tions, although  this ratio varies widely
from plant  to plant. The PIC  output  was
also compared to the total POHC  input in
the waste feed.   As  Table 2 shows, only
one of the eight plants exceeded 0.01% of
POHC input,  leading  to the  conclusion
that the PIC output rate  only  infre-
quently  exceeds  0.01% of  the POHC input
rate.

     Three possible  mechanisms to explain
the presence  of compounds in  the  stack
gas classified  as PICs were assessed.
These mechanisms and the relative contri-
bution of each  are discussed in  the  fol-
lowing sections.

Compounds Originally Present in the Waste
  Feed Below 100 |Jg/g with low DRE

     For this study, a compound present
in the waste  feed below 100  (Jg/g but de-
tected in the  stack would be  classified
as a PIC.   The  approximate  stack concen-
tration  for a compound present just below
100 |Jg/g, if  it were subject to  a  DRE of
99.9%, would be  10 ng/L.  Since this  con-
centration  is substantially  above the de-
tection  limit  of the  stack  sampling
method,  the compound is identified as a
PIC.  Data  gathered  on the program con-
firmed that this situation might be  rea-
sonably  expected.

     In  certain cases, the  waste feed
samples  were  relatively "clean" and  de-
tection  limits   were on the order of
1 M8/8-  These  waste feed samples  fre-
quently  contained  compounds  at concentra-
tions  below 100 |Jg/g.  When DREs were
calculated  for  these compounds, they  gen-
erally fell below 99.9%.   Although this
mechanism would not account for a large
percentage  of the total mass  output of
Appendix VIII compounds, it  could explain
the presence  of several of  the PIC com-
pounds detected in  the stack  effluents.

Compounds  Introduced  from  Some Source
  Other  Than Waste Feed

     PIC compounds may also  be introduced
to  the  incineration process  by  streams
other than  the  waste feed.   Three  sources
are:   (1) scrubbing  liquid  used in  con-
trol devices;  (2) inleak of ambient  air
through spaces in ductwork; and (3) aux-
iliary fuel  streams.   Instances of all
three possible sources of introduction to
the incineration process occurred during
the course of the testing program.

     Municipal water  is often  used  as
makeup to  scrubbers.  Appendix VIII com-
pounds commonly are found  in measurable
concentrations in  municipal water sup-
plies.  These compounds are primarily the
THMs mentioned earlier, and they are  in-
troduced into the water via chlorination.
The THMs are in low concentrations in the
water inflow streams,  but can be stripped
out of the water with high  efficiency by
the hot  stack gases.   In  a number of
cases the  scrubber  water  effluents con-
tained substantially  less  of these or-
ganic compounds than  did  the makeup wa-
ter.   This  phenomenon could  account
for the presence of THMs which  appear as
PICs in the stack effluent of plants with
wet  control  devices  using chlorinated
makeup water.  In  addition to THMs con-
tributed by treated water supplies, other
trace organics can  sometimes be found in
well water or lagoon water used as makeup
to scrubbers.

     Typically,  a  plant   exhibited   a
scrubber makeup flow  rate of 70 L/min  (18
gpm) and a stack flow rate of 100 Nm3/min
(3,500 dscfm).   THMs  were  measured  in
makeup water at concentrations  of 100
|jg/L  for  chloroform and 10 (JgA each  for
bromodichloromethane,  dibromochlorometh-
ane,  and  bromoform.   The  scrubber water
was stripped  clean of all THMs (the  ef-
fluent  contained  < 1 Mg/L  of  all com~
pounds).   This  stripping resulted in-a
contribution  to  the  stack  effluent  of
0.007 g/min  (70 ng/L)  for  chloroform  and
0.0007 g/min  (7 ng/L) for  each  of the
other  THMs.   These concentrations were
easily detectable by  the sampling methods
used.  In  one case, well water used for a
quench makeup was  found to contain four
Appendix VIII  compounds  in quantities
greater than 100 |Jg/L.

     A second possible  source of Appendix
VIII  compounds  in the stack effluent is
ambient air  that  may enter the negative
pressure  system  through spaces in duct-
work  or other openings downstream  from
the  combustion chamber.   Heat balance
calculations performed  upon the eight in-
cinerators indicated  that from 0 to  50%
                                            88

-------
                                                                      J
                                                                                    p.
                                                                                    H
                                                                                    PH
                                                                                    O
                                                                                    4J
                                                                                    O
                                                                                    O

                                                                                    4J
                                                                                    3
                                                                                    O,
                                                                                    4J
                                                                                    3
                                                                                    O
                                                                                    I

8     8
                     UOJ4DJ4U33UO3
                               89

-------
      TABLE 2.  CALCULATED VALUES FOR PIC EMISSIONS AS A
                  PERCENTAGE OF POHC INPUT
Calculated values for PIC emissions as a percentage of POHC
input:

          01 - PIC output (g/min)
          '"   POHC input (g/min) X
Plant     PIC output (g/min)
                POHC input (g/min)
  1

  2

  3

  4

  5

  6

  7

  8
0.090

0.0046

0.092

0.00062

0.011

0.00090

0.26

0.0076
9,800

  100

  780

   27

1,200

  310

8,300

1,200
0.00092

0.0044

0.012

0.00023

0.00091

0.00039

0.0031

0.00061
                               90

-------
of the  measured stack  flow rate might
have entered the  system downstream from
the  combustion  chamber.   It is not un-
usual for ambient  air to contain higher
concentrations of some Appendix VIII com-
pounds  than  the concentrations measured
in the  incinerator  stack.3   The ambient
air concentrations may be higher in areas
of a facility where hazardous  wastes are
handled.

     Ambient air samples collected at one
facility contained measurable  concentra-
tions  (between  2  and  30 ng/L) of  eight
Appendix VIII compounds,  as  shown in
Table 3.  The inleak  rate  at this plant
was  calculated  to be 25% of the total
stack flow.   This inleak rate was used to
calculate the figures  shown for contribu-
tion to stack gas concentrations.  All of
these compounds  were detected at low lev-
els in  the  stack,  so  this source is not
suspected of  contributing a large mass
flow rate of compounds; but  it still can
result  in  the  identification  of addi-
tional compounds as PICs.

     A third potential source of Appendix
VIII  compounds  is  the  auxiliary  fuel
stream, which is  not  usually classified
as a waste  feed.   Uncombusted fractions
of hazardous compounds may be detected in
the  stack gas,  and identified as  PICs.

     One of the facilities  tested fired
auxiliary fuel  oil along with  liquid and
solid waste.  A sample  of the fuel oil
was found to contain  eight Appendix VIII
compounds including five at concentra-
tions exceeding 100 Hg/g,  as  shown  in
Table 4.  All of these compounds had been
identified as POHCs in  other waste feeds
at this facility, but the auxiliary fuel
was the primary contributor of two of the
compounds.   While this particular example
does not influence  the  PIC  emissions of
the facility, had the same  fuel  oil been
used at a  facility where the  compounds
were not  listed as POHCs in the waste
feed, the undestroyed fraction would be
detected in the  stack  as a PIC.

Compounds Occurring as Actual  Combustion
  By-Products

     The third  mechanism to explain the
presence of compounds identified as PICs
is their formation  as actual combustion
by-products.  Due to the extremely
complex  nature  of waste feed and  stack
effluent, it  is  impossible to arrive at
specific  reactions  or possible kinetics
based  simply on  the results  of this
study.  However, a review of the informa-
tion collected is.useful.  The discussion
which  follows is  presented  in  three
parts, according  to  the  category of  com-
bustion by-product.

          Identifiable fragments of  feed
constituents  which  result from  partial
oxidation or  simple  substitution  reac-
tions .

     An example in this category would be
studies reporting the appearance of hexa-
chlorobenzene as a by-product of PCB com-
bustion.2   Several   of  the plants MRI
studied exhibited PICs which  were  unique
to  the  type  of waste being  burned and
were  identifiable products of reactions
of  waste  constituents.   Two  facilities
which burned  large quantities of aniline
both  exhibited  high concentrations  of
nitrophenols  in the stack.   Only one
plant showed measurable concentrations of
polychlorinated benzenes  in the waste,
and this  plant  exhibited di-, tri-,  and
hexachlorobenzene in the  stack.   Like-
wise, chlorophenols  were detected  in the
waste at only the one plant where ortho-,
tri-, and pentachlorophenol were detected
in  the stack.   These cases suggest that
the compounds identified  as  PICs  were
formed as a result of specific constitu-
ents in the waste feed.

          Reaction products which are the
re'sult of complex recombination or substi-
tution reactions.

     This  category  typically includes
compounds of high molecular weight.  High
molecular weight  compounds identified as'
PICs which might have resulted from  com-
plex  reactions   included  naphthalene,
fluoranthene, and pyrene.

          Simple  fragments  which cannot
be readily traced to any' single reactant,
but  rather  seem  to  be  universal  by-
products  of  the combustion  of  organic
compounds.

     This category,  which  includes many
low molecular weight compounds,  was  the
most broadly represented among the plants
studied.   Each  plant exhibited  one  or
                                            91

-------
TABLE 3.  APPENDIX VIII COMPOUNDS FROM AMBIENT AIR


Compound
Benzene
Chlorobenzene
Toluene
Chloroform
1,1, 1-Trichloroethane
Tetrachloroethylene
Trichloroethylene
Carbon tetrachloride



PIC
PIC
POHC
POHC
POHC
POHC
POHC
POHC
Concentration
in stack gas
(ng/L)
130
3.9
14
20
0.3
4.2
2.5
8.4
Potential contribution
to stack gas concentration
(%)
4
11
50
4
100
17
21
5
       TABLE 4.   APPENDIX VIII COMPOUNDS FOUND
                   IN AUXILIARY FUEL OIL
                   SAMPLE
             Compound
Concentration
    (M8/8)
      Toluene
      Methylene chloride
      Tetrachloroethylene
      1,1,1-Trichloroethane
      Methylene bromide
      Benzene
      Chloroform
      Trichloroethylene
    1,500
      440
      380
      220
      210
       91
       23
       20
                         92

-------
more fragments which might have  resulted
from the combustion of organic compounds.
These  included  chlorinated  methanes
(chloromethane,   methylene   chloride,
chloroform,  and  carbon  tetrachloride),
and  chlorinated  ethanes  and ethylenes
(1,1,1,-trichloroethane,  trichloroethyl-
ene  and  tetrachloroethylene).   In addi-
tion, bromoform was detected in high con-
centrations  at one facility which was
burning  brominated  wastes.   Other frag-
ments which were found as PICs could have
resulted from  the  combustion of aromat-
ics, and included benzene, phenol, tolu-
ene, and chlorobenzene.

Relative Contribution  of Different PIC
  Mechanisms

     Figure 3  summarizes, for each of the
three PIC formation mechanisms discussed,
the  relative potential  contribution  of
each to  the  total mass output of  PICs.
Low concentration POHCs in the waste feed
potentially accounted for 0  to 68% of the
total mass of  PICs  detected at each fa-
cility.  The higher percentages  were  for
the plants with  the lowest total mass  of
PIC  emissions.  The stripping  of THMs 
from the scrubber waters potentially  ac-
counted  for  up to 14% of the total mass
of PICs  detected.   The rest of the com-
pounds were  most likely actual products
of  combustion  reactions.  Figure 3 also
includes a bar which  represents  the per-
centage  of the PICs which occurred most
frequently and in the  highest concentra-
tions.   The  six  compounds in this cate-
gory  accounted for a  substantial per-
centage  of the total  mass of PIC output
at  most  of the  facilities tested.  Ta-
ble 5  lists  these  compounds, the concen-
tration  range  measured,  and how often
they were detected.
CONCLUSIONS

     This program did not initially focus
on  identification  of  PIC  mechanisms;
thus, definite limits were  imposed on in-
terpretation  of  the PIC data.  The dis-
cussion  regarding  appearance of PICs in
the  stack  is  fairly well documented for
the  case of the compounds  stripped from
the  scrubber  water, but  otherwise the
discussion  of PIC  formation mechanisms
is somewhat conjectural.  Possible
     TABLE 5.  SIX MOST COMMON PICs
Compound
Concentration
range (|Jg/L)
No. of
plants '
as a PIC
Chloroform
Benzene
Naphthalene
o-Nitrophenol
Bromoform
Phenol
1.5
13
4
25
0.21
5
- 1,330
670
120
35
23
19
mechanisms for formation of reaction by-
products, in particular, would require
more information before being confirmed
as explanations.  More thorough investi-
gation is also needed for the non-POHC
fraction of the waste feed which, for the
most part, remained unidentified in this
study.  Since the non-POHC fraction typi-
cally constitutes 30 to 90% of the waste
feed, by-products from non-POHC combus-
tion might prove very important in ex-
plaining the PIC phenomenon!.

     In summary, general observations
from the study are:

          Stack gas concentrations of
          PICs were typically as high as
          or higher than those for POHC
          compounds.

          PIC output rate infrequently
          exceeded 0.01% of POHC input
          rate.

          The three likely mechanisms
          that explain the presence of
          most PICs are:

               Low DREs for Appendix VIII
               compounds present at low
               concentrations (< 100
               M8/g) in the waste feed;

               Input of Appendix VIII
               compounds to the system
               from sources other than
               waste feed; and
                                            93

-------
                                          U
       O
       oo
00
Os
                         CM
                         col
           00
                                         U
                                         to
                                         U
                                         to
     I	I
                         I    I   I   I   I
                O
                >o
.000000

si

[
=[=
N.
CO

col
o-l

CN
ON
>[ 	
R

ol

|

"1
"I
I 1 1 1 1 1 1 1 1 1 1

U
to

Q.
^
U
oo
VI
*-
n
a. ^
D-
(U
"c
0)
CO
52
_Q
U E
to 
V)
a. 1
D
O
Q.
^ 
^ u
to
u- II
U
to
05
U Ou
to 
o.
o. C
^ II
T3
1 1 (y

u
u_
Figure 3. Relative Contribution of Different PIC Mechanisms
o ooo ooo oooo
o ON co t'N. o > *} co CM 
                 4U3DJ3J
                                                 94

-------
               Actual products of incom-
               plete combustion, or prod-
               ucts of complex side reac-
               tions .

          Benzene is the most likely com-
          pound to appear as a PIC, based
          on both frequency of occurrence
          and presence at higher concen-
          trations .

          Measurement  of Appendix VIII
          compounds at low concentrations
          in the  waste feed, auxiliary
          fuel,  and inflow  streams  to
          control systems is often neces-
          sary to explain the presence of
          PICs.
REFERENCES
     Duvall, D.  S.  and Rubey,  W. A.,
     Laboratory Evaluation  of  High Tem-
     perature Destruction of Kepone and
     Related Pesticides.  Cincinnati, OH;
     U.S.   Environmental    Protection
     Agency;  1976  December,  EPA-600/
     2-76-299.

     Ahling, B. and Lindskog,  A., Thermal
     Destruction  of PCB and Hexachloro-
     benzene,  The Science of the Total  .
     Environment, Vol.  10 (1978),  51-59.

     Singh, H. B. ,  et  al.,  "Distribution
     of Selected Gaseous Organic Mutagens
     and  Suspect  Carcinogens  in Ambient
     Air,"   Environmental Science and
     Technology, Vol. 16 (1982), pp. 872-
     880.
                                            95

-------
                      PRELIMINARY ASSESSMENT OF COSTS AND CREDITS

                  FOR HAZARDOUS WASTE CO-FIRING IN INDUSTRIAL BOILERS
                             R. McCormick and L. Weitzman
                                  Acurex Corporation
                                    Cincinnati, OH
                                       ABSTRACT
     This paper outlines the results of an IRB-sponsored study to develop preliminary
cost vs. credit information for hazardous waste co-firing in industrial  boilers.  An
overview of the cost/credit estimating methodology is presented, along with pertinent
background information on hazardous waste co-firing practices.  Details  of all  cost
estimates are provided in the project report, Reference 1.
BACKGROUND

     The practice of burning hazardous
wastes in industrial boilers has become
increasingly popular over the past decade.
The reasons for this trend are twofold:

(1)  Increased prices for conventional
     fuels, favoring waste substitution or
     co-firing as a means of reducing fuel
     expenditures; and

(2)  Stricter environmental regula-
     tions and higher costs for hazardous
     waste disposal by conventional
     methods, promoting waste co-firing as
     a means to eliminate or reduce waste
     disposal costs.

     Two regulatory actions under the
Resource Conservation and Recovery Act
(RCRA) in particular have favored boiler
co-firing as a hazardous waste disposal
alternative. The first such action was the
virtual ban on liquid waste landfill ing
which promoted thermal destruction (in
Incinerators, boilers, cement kilns, etc.)
as the favored disposal method for hazard-
ous organic liquids.  The second regula-
tory action favoring boiler co-firing was
the promulgation of emissions limitations
for incinerators burning hazardous wastes.
Since  1980, hazardous waste incinerators
have been subject to Destruction and
Removal Efficiency (ORE) requirements for
hazardous organic constituents of the
waste, particulate emission limitations,
and HC1 removal requirements.  No such
regulations have been imposed on boilers
burning hazardous wastes.

     This regulatory dichotomy has come
under  increasing attack over the last four
years.  From the standpoint of environ-
mental safety and health, there is no
conceptual difference between hazardous
waste  combustion in incinerators and
boilers (although actual emission char-
acteristics may vary).  Boilers were
originally exempted from'regulation
because the? implied objective of waste
.co-firing was energy recovery, rather than
waste  disposal.  However, this distinction
has been clouded by the recent trend
toward waste heat recovery in incineration
facilities.

     In order to assess the need for and
impacts of regulations governing hazardous
waste  co-firing in boilers, EPA is cur-
rently conducting a Regulatory Impact
Analysis (RIA) study.  One of the major
elements of this program is an economic
assessment of the impacts of boiler
                                            96

-------
performance standards and associated air
pollution control  requirements.  The study
addressed herein is intended to provide
preliminary cost/ credit information for
current boiler co-firing practices.  As
such, it serves as a precursor to the eco-
nomic impact assessment phase of the RIA.
OBJECTIVES AND SCOPE

     The underlying objective of the study
is to provide a preliminary evaluation of
the costs and credits associated with
hazardous waste co-firing in industrial
boilers.  Specific objectives are as
follows:

(1)  Identify the major waste/boiler
     co-firing scenarios that need to
     be addressed for the purposes of
     the RIA.

(2)  Identify the major costs and credits
     associated with boiler conversion to
     waste co-firing.

(3)  Develop preliminary cost data
     for the major and most probable
     boiler retrofit activities.

(4)  Develop a parametric approach to
     estimating retrofit costs, incre-
     mental operation and maintenance
     (O&M) costs, and credits so that
     cost/credit tradeoffs can be pro-
     jected as a function of waste type,
     waste:fuel co-firing ratio, boiler
     design and capacity, and potential
     air pollution control requirements.

     Due to the preliminary nature of the
study, none of the objectives are ad-
dressed in a completely thorough or
rigorous manner.  The goal is simply to
provide a basis for  initial decision-mak-
ing and more detailed future study.
 CO-FIRING  SCENARIOS  EVALUATED

      The range of hazardous waste co-fir-
 ing  scenarios addressed  in the study is
 necessarily  limited  in terms of waste
 type, boiler design  characteristics,
 capacity,  and waste:fuel co-firing ratios.
 The  evaluation is limited to those scenar-
 ios  which  best represent current practice
 and  probable future  practice.
Waste Characteristics

     As a starting point, the evaluation
is limited to hazardous organic liquid
waste co-firing.  This limitation is
imposed because organic liquids are the
prime candidates for boiler co-firing in
terms of volume, handling characteristics,
primary fuel firing compatibility, and
compatibility with standard industrial
boiler designs.

     Second, the economic evaluation is
based on in-house generation of the
organic liquid wastes being co-fired.  The
assumption here is that industrial boiler
operators do not routinely accept wastes
from outside sources, thus functioning as
commercial waste disposers.

     The third major assumption is that
the majority of wastes co-fired in indus-
trial boilers possess desirable fuel
properties.  Heating values are assumed to
be greater than or equal to 8,000-10,000
Btu/lb, such that the waste will support
combustion.  Water, ash, and halogen
contents are also assumed to be reasonably
low, so as to preclude flame quenching,
excessive ash buildup on the boiler tubes,
excessive bottom ash handling problems,
and/or cold-end corrosion problems.

     Finally, reasonable mid-range waste:
primary fuel co-firing ratios are con-
sidered.  The range is 10-50% waste with
primary fuel, with the percentage based on
gross heat input.

Boiler Designs and Capacities

     The basic boiler designs considered
in the study are those characteristic of
the  industrial size range (up to  250,000
Ib/hr steam) originally designed to burn
natural gas, distillate oil, residual oil,
or a combination of these fuels.  The
three most common design types, addressed
in the study, are as follows:

1.   Scotch firetube, N-pass design,
     packaged boilers for natural  gas,
     distillate oil, or combination
     gas/distillate oil firing.  Single
     burner design, with no economizer or
     air heater.  Typical capacities are
     10,000 to 30,000 Ib/hr steam, up to
     50,000 Ib/hr steam.  Saturated steam
     up to 150 psig.
                                            97

-------
 2.  Watertube design for natural  gas,
     distillate oil,  gas/distillate,
     or gas/distillate/residual  oil
     firing  (with  oil  preheat  equip-
     ment).   Single or multiple  burner
     design.   Steam capacities of  20,000 to
     100,000 Ib/hr saturated at  125 to  250
     psig  design rating.  Economizers  not
     atypical  at more than  50,000  Ib/hr
     steam,  but air heaters rare.

 3.  Watertube design with multiple
     burners  for gas/oil  firing.   Steam
     capacities of 100,000 to  250,000 Ib/hr
     and up,  with  turbogenerator steam
     pressures and superheat in the larger
     size  ranges.  Economizers typical, and
     possibly  air  heaters.

      These  boiler designs are most common
 (and compatible)  for liquid waste co-fir-
 ing, although a limited  number of coal-
 fired boilers are also being  used to
 co-fire hazardous liquids.  Both stoker
 and  pulverized coal-fired boiler conver-
 sion for  waste co-firing are  addressed to
 a minor extent.
CAPITAL REQUIREMENTS FOR BOILER SYSTEM
RETROFIT

     As indicated in the preceding sec-
tion, one of the baseline assumptions of
this study is that boilers co-firing
hazardous wastes were originally designed
to burn one or more conventional fossil
fuels  natural gas, oil, or coal.
Therefore,  an initial capital investment
is required to retrofit a boiler for waste
co-firing.  This capital investment can be
divided into three components:

(1)  Addition of waste storage and feeding
     equipment,

(2)  Boiler modification to accommodate
     the waste fuel, and

(3)  Air pollution control device (APCD)
     addition for parti oil ate/HC1  removal,
     if required by regulation.

The total  investment for retrofit is
usually modest, unless air pollution
 control  device  addition is mandated.
 Costs  are  summarized below; details may be
 found  in Reference  1.

 Waste  Handling  and  Feeding Equipment

     Liquid wastes  are usually transported
 from the process area to the boiler
 location by tank truck and pumped into a
 settling tank,  or holding tank.  Typical
 installed  costs for individual 5,000,
 10,000,  and 20,000  gal tanks designed for
 hazardous  liquid waste storage are approx-
 imately  $34,000, $46,000 and $61,000
 (December  1982), respectively.

     From  the storage tank, wastes are
 usually  pumped  directly to the boiler
 through  self-cleaning filters or dual
 strainers  to remove any suspended solids.
 Dual feed  pumps, costing $1,000 to $1,500
 each,  are  desirable to provide redundancy
 in the event of equipment failure.  Piping
 will cost  $2,000 to $5,000.  Costs for
 in-line  filters can range from several
 hundred  to several  thousand dollars,
 depending  on the flowrate, pressure,
 solids loading, and solids size distribu-
 tion.

 Boiler Retrofit

     In  any co-firing situations, retrofit
 at the boiler is a  relatively minor
 consideration.  High costs for boiler
 retrofit are usually incurred only when a
 completely new burner assembly, or burner
 assembly plus blower, damper, and control
 system is  required.  This may be the case
 when stoker or pulverized coal-boilers are
 retrofitted to burn liquid waste, or even
 when gas-oil fired boilers are converted
 over for waste co-firing.

     Equipment cost vs. capacity data for
 three potential  retrofit requirements are
 presented  in Figure 1.  The lower curve
 applies to one of the simplest retrofit
 cases:   liquid waste nozzle addition/re-
 placement.  The costs shown in Figure 1
 are the nozzles only; total  installed
 costs including pressure and flow control
 valves, plus labor,  can easily run  two to
three times the costs for the nozzles
 alone.

-------
  Figure  1.
 60 r
Burner Component and
    System Costs
                        Liquid waste nozzle
                       A Complete burner assembly
                        Burner assembly plus
                        blower and controls
        100
              200   300   400
               GPH CAPACITY
                   500
                         600
     The middle curve is for complete
liquid fuel burner assemblies with natural
gas firing capability built-in.  The costs
include pilot, spark ignition assembly,
and damper assembly.  The upper curve is
for complete burner, systems, including
forced-draft blowers and master controls
for flame safeguard, fuel modulation, air
pressure and flow regulation.

      Installation costs for burners and
complete burner systems are typically 50%
of the equipment cost.  If structural
modifications are required to install the
burner, this installation cost can easily
double.

Air Pollution Control Device Addition

      Industrial boilers designed for
natural gas or fuel oil firing are gener-
ally  no.t equipped with any APCD's, includ-
ing boilers co-firing hazardous liquid
wastes.  If Federal standards similar to
the RCRA performance standards for incin-
erators were to be promulgated for boilers
co-firing  hazardous waste, it is likely
that  at least some boiler facilities would
need  to provide one or more APCD's for.
particulate and/or HC1 removal.  Obviously,
this  would depend on the ash content of
the waste, chlorine content of the waste,
and the regulatory requirements imposed.

      It is impossible to predict industry
trends  relative to APCD addition at this
point, should,performance standards for
particulate and/or HC1  be promulgated.
However, six scenarios can be envisioned:

(1)  The waste being co-fired at the site
     in question contains negligible ash
     and chlorine, so APCD addition is
     unnecessary to comply with regulatory
     requirements.

(2)  Chlorinated wastes are not being
     burned at the site, but particulate
     emissions exceed the standard as a
     result of waste co-firing.  The
     mandated particulate removal effi-
     ciency is in the 95 to 99% range.   In
     this case, a "cold-side," dry elec-
     trostatic precipitator (ESP) is the
     likely choice for particulate control.

(3)  This scenario is identical to the
     previous one, except that the man-
     dated particulate removal efficiency
     is more than 99.5%.  In this case,  a
     baghouse would probably be more
     economical than an ESP for particu-
     late control due to the excessive
     plate area required for an ESP to
     achieve this efficiency.  This is a
     generalization, however, which could
     be reversed by site-specific factors.

(4)  The co-fired wastes do not contain
     significant ash, but chlorine is
     present in quantities such that the
     waste/fuel mix contains more than
      0.5% organically-bound chlorine.   In
     this case, a packed bed absorption
     system may be  required for HC1
      removal.  This is not considered
     a  typical waste co-firing  scenario,
     due to the adverse boiler  corrosion
      impacts associated with burning
     chlorinated materials.

(5)  The co-fired wastes contain signifi-
      cant ash  and chlorine, such that
      particulate  and  HC1 emission  stand-
      ards would both  be  violated.   In  this
      case,  a venturi  scrubber/packed  bed
      absorption system would  be the
      likely choice  for  air pollution
      control.   Again,  this  is  not
      considered a typical waste co-firing
      scenario  for the  reason  stated
      above.

 (6)   Either particulate  or HC1  emission
      standards are  exceeded,  but the
                                            99

-------
     boiler facility owner/operator
     chooses to eliminate waste co-firing
     rather than install the necessary air
     pollution control equipment.

     These scenarios apply specifically to
natural gas/fuel oil-fired boilers, but
also to coal stoker-fired boilers in prin-
ciple because stokers are usually equipped
with no more than cyclonic separators for
particulate control.

Electrostatic Precipitators

     Electrostatic precipitators are one
of the most common particulate control
devices for large-scale utility boilers,
and are well suited for industrial boiler
retrofit to control particulate emissions
unless (a) the gas is particularly wet, or
(b) the gas contains significant HC1 such
that corrosion is a major potential  prob-
lem.  ESP advantages are high particul ate
collection efficiency at neglibile pres-
sure drop.

     Basic ESP costs are determined by two
primary factors: insulation requirements
and collection plate area requirements.
Insulation requirements are a function of
the cold-end temperature and gas dew
point, i.e., condensation is undesirable
due to the associated corrosion problem.
For standard, uninsulated ESP's, cost is
related to plate area by

     P * 129,400 + 4.42ft              (1)

where P = purchased cost in June 1983
          dollars
      A = collection plate area, ft2

If insulation is required,

     P * 201,000 + 6.62A              (2)

     The basic equation for plate area
is
     A * Q In (1- n)/w
(3)
where Q * gas flow rate, cfs
      n * fractional collection efficiency
      w = drift velocity ft/s

For the purpose of this study, a drift
velocity range of 0.1-0.3 ft/s is assumed.
              The total  installed cost for a
         basic ESP system is roughly double the
         purchased equipment cost.

         Baghouses

              Baghouses, or fabric filters, are
         also viable candidates for retrofit
         control  of particulate emissions  from
         boilers  co-firing hazardous wastes,
         especially if extremely efficient par-
         ticul ate removal  (^99.5%) is needed.
         However, baghouses are subject to the same
         condensation/corrosion problems that can
         plague ESP operation;  thus, the same
         limitations on  gas moisture content and
         acidic constituents apply.

              The purchased cost for a fabric
         filter system is a complex function of gas
         temperature, composition, particulate  .
         loading  and size distribution, and system
         design features.  Some of the major
         factors  impacting costs are as follows:

         (1)   Type of system used:

              - intermittent pressure, mechanical
                shaker cleaning

              - continuous pressure, pulse jet
                cleaning

              - continuous pressure, mechanical
                shaker cleaning

              - continuous pressure, reverse air
                cleaning

         (2)   Type of bags (filter medium)  used.
              The nominal  operating temperatures of
              various fabrics are:
              Cotton
              Ifylon
              Acrylic,  polyester
              Nomex
              Tefl on
              Fiberglass
180F
200 F
275F
425F
400F
550F
         (3)   The  need  for  insulation to  prevent
              condensation

         (4)   The  cloth area  required (function of
              capacity  and  air-to-cloth ratio)

         (5)   Whether the system  is custom built or
              standard.
                                           100

-------
     The approach used to estimate a
purchased equipment cost for a baghouse
system, incorporating the design consid-
erations listed above, does not lend
itself to the simplification needed for
this paper.  Therefore, the reader should
refer to Reference 1 for a procedure to
estimate purchased costs for fabric filter
systems.  Installation costs may then be
estimated as 95-100% of the purchased
equipment cost.

Venturi Scrubber/Acid Gas Absorber Systems

     If the gas is extremely wet or
contains significant HC1 as a result of
co-firing chlorinated wastes, a venturi
scrubber/acid gas absorber system would be
the probable choice for particulate and/or
HC1 removal.  Such a system would likely
include a small nickel alloy quench to
reduce the temperature of the boiler exit
gas to saturation, a high nickel alloy
venturi scrubber, a cyclonic separator and
integral packed tower absorber of fiber-
glass construction, polypropylene tower
packing, a caustic recycle system, a
corrosion-resistant fan, and an acid-
resistant stack.  This type of system
would be preferred in cases where wastes
containing significant quantities of ash
and/or chlorine are to be co-fired.
     Total equipment cost, C, for this
generic system is approximately given by
the relationship derived in Reference 2:

     log C = 0.81 log Q + 1.7

where Q is the inlet gas flowrate, ranging
from 1,000 to 50,000 acfm.  These costs 
are based on the assumption of a 30-in WC
venturi pressure drop, which would be
typical for hazardous liquid waste co-fir-
ing.  If no venturi scrubber is needed,
(i.e., little or no particulate control
problem), the total equipment cost for the
system is reduced by approximately 15%.
Installation costs for this type of
scrubbing system are typically about 50%
of the equipment cost.

Indirect Costs
     Indirect engineering and construction
costs will be associated with any major
retrofit activity such as burner system
replacement/addition or APCD addition.
These indirect costs are usually estimated
as percentages of the total  direct cost
for equipment purchase and installation,
as shown in Table 1.
                          TABLE 1.  INDIRECT COST MULTIPLIERS
                 Indirect Cost
       Percentage of Direct Cost
        Engineering and supervision

        Construction and field expenses

        Construction Fee

        Startup

        Contingency
              10 to 20a

              10 to 20b

              7.5 to 10

              1 to 3

              10
        TOTAL
                                                             38.5 to  63
        a 20% applies to ESP addition
        b 20% applies to ESP or baghouse addition
                                           101

-------
     These percentages are somewhat higher
than would be expected for new facility
construction because they apply to retro-
fit situations.
INCREMENTAL O&M COSTS

     Incremental O&M costs for boiler
operation are likely whenever waste,
rather than fuel, is burned.  These incre-
mental costs can include increased con-
sumption of electric power and water,
operating labor, higher costs for residue
disposal and maintenance, add-on costs
for scrubbing chemicals, liquid nitrogen,
and waste analysis, capital recovery
charges, and taxes and insurance.  Residue
disposal and maintenance may be particu-
larly significant and are discussed below.*
The other O&M costs are discussed in
Reference 1.

Ash Disposal

     Co-firing hazardous waste with
natural gas or fuel oil will in all
likelihood increase the quantities of ash
deposited on the furnace floor and the
hopper fallout.  Addition of baghouses or
ESP's will also increase the quantities of
ash that must be disposed.  However, the
major potential cost impact of hazardous
waste co-firing is in ash characteristics,
rather than the quantities of ash gen-
erated .

     Under current RCRA regulations for
incinerators and other treatment pro-
cesses, any residue (e.g., ash) from the
processing of a hazardous waste is, by
definition, a hazardous waste as well.   If
this criterion was expanded to include ash
from boilers co-firing hazardous waste,
the cost impact could be significant
(depending, of course, on the quantity of
ash handled).

     Conventional  landfill ing costs in
major  industrial areas throughout the U.S.
range  from about 0.1 to  l.Ojd/lb for dry
*This  assumes  that there are  no air
pollution  controls added.   If controls are
added, then  capital  recovery  costs are
likely to  be significant,  and if the
control  is a scrubber,  then the cost of
caustics is  important to consider.
solid waste.  Secure landfilling  costs  in
the same areas range from 1.2 to  4.8/lb
for the same type of waste.  Typically,
the secure landfilling costs are  an  order
of magnitude higher.  Total annual cost
for ash disposal  can be estimated using
the general relationship:

   Annual cost, $ =

   (Ash content of waste, lb/1b)x

   (Annual waste throughput, lb)x

   (Landfilling cost, $/lb)         (4)

Maintenance

     Maintenance is potentially the  most
significant annual cost associated
with co-firing hazardous wastes in boilers
and also the most difficult to predict.
Within the boiler itself, severe convec-
tive pass fouling and acid-gas corrosion
are the two major concerns.  The major
conclusion pertinent to this study is that
incremental boiler maintenance due to
waste co-firing can range from virtually
nil to the other extreme of catastrophic
tube bundle failure.  This type of corro-
sion-related problem is most likely  when
chlorinated, other halogenated, or phos-
phorous-bearing materials are burned and
economizers or air heaters are incorpor-
ated in the boiler design.  Likewise,
fouling problems are most likely to  occur
when high ash content and/or alkali  metal
contaminated waste are burned in boilers
with narrow, staggered tube spacings in
the convective passes.

     A considerably larger data base is
available on APCD maintenance costs.
Assuming that the proper type of equipment
is specified for a given,application, and
that suitable materials  of construction
are used, the annual maintenance cost for
ESP's and wet scrubbing  systems is typic-
ally 5% of the .equipment cost.  A 5%
multiplier can also be used to project
hardware maintenance for baghouse systems,
but the  cost for bag replacement must
also be  considered.


CREDITS

     The  credits associated with co-firing
hazardous  wastes include the obvious fuel
                                            102

-------
savings, plus elimination of certain
on-site and/or contractor costs for
dispoal of the waste.

Fuel Savings

     Unless the full capacity of the
boiler is not needed to satisfy plant
steam requirements, one of the basic goals
of waste co-firing is to maintain the same
stream flow and quality.  Therefore, the
net fuel savings as a result of waste
co-firing can be estimated by

     Of = Qw - Qs (1/H2 - 1/ni)      ,(5)

where Qf = Net fuel savings, Btu/hr

      Qw = Heat input with waste, Btu/hr

      Qs = Desired (constant) heat output
           with steam, Btu/hr

      ni = Boiler efficiency during
           baseline operation as measured
           by the ASME heat loss method

      \\2_ = Efficiency during co-firing

From equation (5) it is obvious that the ,
fractional heat input with the waste must
be  greater than the efficiency loss
associated with co-firing in order to
realize a net fuel savings.  Assuming that
this is the case, net fuel savings is
given by:
     Ff = Qf/HVf
(6)
where Ff = mass or volumetric fuel con-
           servation rate

      HVf = fuel heating value

     Based on these estimates, the boi.ler
facility operating schedule, and  local
fuel prices, the annual fuel conservation
credit can be calculated directly.

Waste Disposal Elimination

     The savings associated with  elimina-
tion of the waste disposal problem are
        highly variable, and not always easy to
        estimate.  Because the vast majority of
        waste candidates for boiler co-firing are
        combustible liquids, it is likely that the
        alternative form of waste disposal would
        be incineration, either on-site or at a
        commercial waste disposal facility.

              If the waste was burned in an on-site
        incinerator prior to boiler co-firing, the
        variable  incineration costs such as
        auxiliary fuel, power, water, chemicals,
        residue disposal, labor, and maintenance
        would be  eliminated by conversion to
        boiler co-firing.  Fixed costs for the
        incineration.facility would not be elim-
        inated, however.  Thus, the analysis must
        be done on a case-specific basis.

              If the waste was shipped to an
        off-site  incineration facility prior to
        boiler co-firing, the cost analysis
        is more straightforward.  However,
        estimating an  across-the-board cost
        savings is still difficult.  Commercial
        waste disposal  firms typically charge
        whatever  the local market will bear,
        rather than a  fixed percentage above cost.
        Moreover, most firms charge a premium for
        chlorine  and ash content, and may offer a
        discount  for volume and/or fuel value.
        Typical costs  are 4 to  5/lb,
        although  special wastes  such as PCB oils
        may  demand a price of 30/lb or even
        higher.
REFERENCES

1.  McCormick, R.J. and L. Weitzman.
    Preliminary Assessment of Costs and
    Credits for Hazardous Waste Co-Firing
    in Industrial  Boilers (Draft), EPA
    Contract No. 68-02-3176.  U.S. Envi-
    ronmental  Protection Agency, Cincin-
    nati, Ohio.  October 1983.

2.  McCormick, R.J. and R.J. DeRosier.
    Capital and O&M Cost Relationships for
    Hazardous Waste Incineration Contract
    No. 68-02-3176.-  U.S. Environmental
    Protection Agency, Cincinnati, Ohio.
    July 1983.  p 199.
                                            103

-------
                            HAZARDOUS WASTE  PRETREATMENT AS  AN
                             AIR  POLLUTION  CONTROL TECHNIQUE

                  James J.  Spivey, C. Clark Allen, Robert L. Stallings
                                Research Triangle  Institute
                            Research Triangle Park, NC  27709

                                   Benjamin L. Blaney
                          U.S.  Environmental Protection Agency
                                  Cincinnati, OH  45268
                                        ABSTRACT

     There are many identified hazardous waste streams that contain volatile compounds
that can be emitted to the atmosphere from landfills, surface impoundments, storage
tanks, or other disposal facilities.  One way to minimize or eliminate these emissions
is to pretreat hazardous waste streams to remove volatile compounds.

     This project examines specific hazardous and potentially hazardous v/aste streams
(e.g., phenolic sludge from plastics production) and the technical applicability and
economic feasibility of specific pretreatment techniques (e.g., steam stripping) for
removing volatile compounds from them.  In the absence of competitive removal mechanisms
(such as biological decay), all the volatile constituents will ultimately be released to
the atmosphere.  The results show that air stripping, steam stripping, or evaporation
(coupled with carbon adsorption of the offgases); steam stripping; and batch distilla-
tion are the most widely applicable pretreatment techniques.  An example of air strip-
ping coupled with carbon adsorption for removal of volatile constituents from hazardous
waste streams containing aromatic solvents is given.  The cost-effectiveness of pre-
treatment varies widely with waste stream characteristics and type of pretreatment, with
typical values being between $0.099 and $2.00 per kilogram (kg) of volatile material
removed.   This paper is a summary of a more comprehensive report prepared by the Re-
search Triangle Institute, which is unpublished as of May 1984.   The reader is referred
to the final report for Task 12-5 of EPA Contract 68-02-3149 for further information.
INTRODUCTION

     The purpose of this study is to
conduct a preliminary assessment of the
technical and economic feasibility of
various pretreatment techniques for the
removal of volatile constituents from
hazardous waste streams.  This study was
conducted in response to increasing
concern over the potential adverse health
and environmental consequences associated
with emissions of volatile substances.
Such emissions from hazardous waste
treatment, storage, and disposal facili-
ties (TSDFs) are of particular interest.
According to preliminary unpublished
results from an Environmental Protection
Agency (EPA) national survey of TSDFs
conducted in 1981 (U.S. EPA 1983a), there
are about 4,820 TSDFs in this country
managing a total of about 151 billion
liters, or approximately 100 million
megagrams (Mg), of hazardous waste annu-
al ly.  Although no comprehensive estimate
of the total volatile emissions from all
domestic TSDFs has been reported, Breton
et al.  (1983) have estimated that nation-
wide emissions of 54 selected volatile
compounds regulated under the Resource
Conservation and Recovery Act (RCRA) may
be as high as 1.6 million Mg per year.
These emissions from TSDFs may represent
as much as 19.5 percent of the total
nationwide emissions of volatile organic
compounds.   These results, although based
on limited data, suggest that emissions
                                           104

-------
of volatile compounds from TSDFs are
probably significant.

     There are a number of sources within
TSDFs from which volatile emissions can
be emitted.  These sources include

          Aerated impoundments
          Landfills
          Surface impoundments
          Cooling towers
          Storage tanks
          Waste handling and transfer
          operations.

     Reasonably detailed data on stream
composition and physical properties are
required to evaluate pretreatment tech-
nology for removing volatile constit-
uents.  Available compilations of hazard-
ous and potentially hazardous waste
stream composition and generation rate
information were evaluated to find the
one most suitable for this study.  The
Waste Environmental Treatment (WET) Model
(U.S. EPA 1983b) was judged to have the
most pertinent information for an engi-
neering assessment of this type.  It is
also the most comprehensive and, in spite
of some limitations (e.g., volatile
constituents are not identified for some
streams and waste streams can vary in
composition), the WET Model data were
useful for assessing potential pretreat-
ment techniques.

     A thorough screening of current
technology was conducted to determine
which pretreatment techniques could be
used for volatile constituent removal/
recovery.  Nine broad engineering tech-
niques were selected:

          Steam stripping
          Chemical oxidation
          Adsorption (liquid or gas
          phase)
          Biological treatment
          Ozonation/radiolysis
          Distillation
          Wet oxidation
          Solvent extraction
          Physical, separation.

Techniques that are  used primarily for
ultimate destruction, such as incinera-
tion and pyrolysis, were excluded.  The
individual compatibility of these tech-
niques and associated economics with all
WET model waste streams was evaluated.
PRETREATMENT TECHNIQUES

     For each pretreatment technique,
there is a set of hazardous waste stream
characteristics (or criteria) that deter-
mine if the technique is applicable to
that stream.  For example, one criterion
for using liquid phase carbon adsorption
pretreatment to remove volatiles is that
the waste stream in contact with the
adsorbent must not contain excessive
concentrations of metallic ions or sol-
ids.  Using this and other criteria, an
appropriate WET Model stream has been
selected, and a detailed example has been
prepared to show how carbon adsorption
pretreatment might be used.  A similar
analysis and an example have been pre-
pared for each pretreatment technique
listed above and are presented in the
final report for this work, which will be
available in the near future.

     A distinction can be made between
pretreatment processes that are applic-
able at the site of generation and proc-
esses that could be used at a TSDF that
accepts waste materials from a variety of
sources.  At a typical TSDF, the waste
streams are generally not segregated by
source.  Thus, pretreatment using carbon
adsorption, steam stripping, and batch
distillation, which has the capability of
handling a variety of waste types, is
likely to be most applicable at TSDFs.

EXAMPLE CASECARBON ADSORPTION

     The adsorption of organic compounds
from both liquid and gaseous phases onto
activated carbon is a mature process
technology with widespread use as an
integral unit operation in such indus-
trial manufacturing processes as corn
syrup and Pharmaceuticals production and
sugar refining, in industrial and munici-
'pal wastewater treatment, in drinking
water purification, in the separation and
recovery of organic compounds from vapor
streams, and in pollution control of
atmospheric emissions.  Although acti-
vated carbon has been and continues to be
the dominant adsorbent used, other adsorb-
ents such as resin or polymeric materials
and zeolite molecular sieves have found
increasing use for a number of special
applications in recent years.

     Since the selection of a pretreat-
ment process depends on the nature of the
                                            105

-------
 hazardous waste stream, a number of
 different process configurations incor-
 porating adsorption methods may be de-
 vised to accommodate a broader spectrum
 of hazardous waste streams.  For this
 discussion, adsorption processes may be
 broadly classified into two groups:
 liquid phase and gas phase adsorption
 processes.   Liquid phase adsorption
 usually involves the addition of rela-
 tively inexpensive powdered activated
 carbon (PAC) directly to the waste
 stream.   The volatile material is ad-
 sorbed and the waste is disposed of
 without removal of the PAC; or, if de-
 sired, the PAC can be removed and regen-
 erated.   Gas phase adsorption systems
 (for pretreatment of hazardous wastes)
 include processes where volatile constit-
 uents are separated from the hazardous
 waste stream into a gaseous stream for
 subsequent removal  by adsorption.   This
 allows adsorption to be used on slurries
 and  sludges that cannot be treated easily
 by liquid phase adsorption.   The phase
 separation  processes and adsorption
 process  that were considered in this
 study are air and steam stripping and
 evaporation.   For these gas phase sys-
 tems,  granular activated carbon (GAC) is
 the  predominant adsorbent.

 Process  Description

     The liquid phase activated carbon
 adsorption  process  involves  two basic
 steps  as shown  in Figure 1.   In, Step  1
 (adsorption), the waste stream contacts
 the  carbon, which selectively adsorbs the
 hazardous material(s) and  allows the
 purified stream to pass through.  Step 2
 (disposition  of contaminated or spent
 carbon)  represents a number  of process
 options.  When  the carbon  reaches its
maximum  capacity or when the effluent is
 unacceptable  for discharge,  the carbon is
 removed  from  the adsorber  for disposal,
destruction,  or regeneration as estab-
 lished by the option selected under Step
2.   In some cases, the  carbon can be
regenerated in  such a way that the ad-
sorbate  is recovered.   This  may be impor-
tant in  pretreatment of hazardous wastes
since the recovered volatile material  may
have some economic value (e.g., as a
solvent).

Process Operation

     The technical suitability  of a waste
              Hazardous
                                 Purified
              Waste Stream
                                 Waste Stream
                 Regeneration with
                  Destruction of
                 Hazardous Waste,
                   e.g.. Thermal
                   Reactivation
                  Regenerated with
                  Reclamation of
                 Hazardous Material
        Figure 1. Steps in carbon adsorption.

stream for  carbon  adsorption  pretreatment
depends mainly on  its physical  form and
the type and relative concentration of
constituents.  However,  other factors
that affect the treatment economics often
dictate which  streams are actually feas-
ible for carbon treatment;  such factors
include the required degree of  solute
removal, waste throughput rate, and
carbon utilization.  The discussion below
is intended to provide guidance in pre-
screening waste streams  for carbon treat-
ment.  Selection should  be  based on
laboratory  adsorption and regeneration
studies with the actual  waste stream.

     In general, carbon  adsorption is
applicable  only to single-phase fluid
streams; specifically, liquid solutions
and gas mixtures.  Although both aqueous
and nonaqueous streams may  be treated
with carbon, liquid phase waste treatment
applications to date have been  confined
to aqueous  streams.

      The suspended solids  level  in the
 waste influent to the carbon contactors
 (adsorbers) generally should be  less  than
 50 ppm for optimum process operation.
 Although suspended solids  levels  up  to
                                            106

-------
2,000 ppm may be handled in special
equipment, stream prefiltering is usually
employed with these streams.

     The amount and type of nonvolatile
organics and dissolved inorganics must
also be considered.  Oil and grease in
the waste stream should be less than
10 ppm to avoid carbon' fouling.  High
levels of dissolved inorganics may cause
such problems as scaling (corrosion) and
serious loss of carbon activity with
thermal carbon reactivation.   In some
cases, however, these problems may be
mitigated through pH control  and soften-
ing of the waste influent or by acid
washing the carbon before reactivation.

     Technically, a high concentration of
solute(s) in the waste  influent does not
limit carbon adsorption treatment.
However,  in actual practice the most
concentrated influent to be treated on a
.continuous basis contained approximately
10,000 ppm total organic carbon (TOC).
Nevertheless,  higher solute concentra-
tions up  to 15 percent  may possibly be
carbon treated.  This limit of 15 percent
solute concentration is strictly a prac-
tical and economic one.  Technically,
there is  no real limit  to solute concen-
trations  that  are  treatable by carbon
adsorption.  In  addition, when the carbon
is  regenerated,  the higher solute influ-
ent concentrations result fn  higher
loading  of the carbon (mass solute
adsorbed/mass  of carbon) and  thus less
frequent regeneration may be  required.

      The flow  rate of the waste  stream
does not affect  the technical  aspects  of
carbon  treatment,  but it does have  a
major impact on  treatment economics.
Carbon  treatment benefits significantly
from economies of  scale, especially where
thermal1 reactivation  is used.   If carbon
usage is less  than 500  kg/day,  then the
economics generally favor  disposal  of  the
spent carbon.   At  carbon usage over
4,000 kg/day,  onsite  thermal  reactivation
is  usually feasible.  Where carbon  rates
are between  500  and 4,000  kg/day,  offsite
reactivation  should be  considered.

      The identification of waste streams
that are suitable  for carbon  treatment is
more difficult when multicomponent  sol-
 utes are present.  Some "secondary"
constituent  may interfere with the  ad-
 sorption of  the solute(s)  of  interest.
For example, long-chain organic soaps
were found to cause poor carbon adsorp-
tion performance on a wastewater stream
from a polyvinyl chloride production
plant (U.S. EPA 1971).  Thus, preliminary
laboratory studies on the actual waste
stream are essential in prescreening
candidate waste streams for carbon treat-
ment.

     In summary, the following charac-
teristics may be used as guidelines to
identify wastes streams that are likely
candidates for carbon treament:

          Aqueous waste streams with
          organic solute concentrations
          that are less than 15 percent.

          Waste streams where the aggre-
          gate concentration of high
          molecular weight nonvolatile
          organics is substantially lower
          than the concentration of the
          volatile organics.

          Waste streams where suspended
           solids are  less than 50 ppm  if
          the stream  is not  prefiltered
           and less than 2.5  percent if
          prefiltering is used.

          Waste streams where oil and
           grease concentrations are less
           than  10 ppm.

           Waste streams where the concen-
           tration of  dissolved  inorganics
           is  low (less than  100 ppm)
           unless waste stream precondi-
           tioning and spent  carbon wash-
           ing before  reactivation opera-
           tions are  included.

     The  removal efficiency  of carbon
treatment can be controlled  to prac-
tically any  level through the design of
the carbon  contactor.  Typical carbon
treatment efficiencies are better than
99 percent with influent concentrations
below  1,000  ppm.  At  higher  influent
concentrations, removal efficiencies can
exceed 99.99 percent  removal to yield
effluent  concentrations at several ppm.
Of course,  as with  most alternative
treatment processes,  carbon  treatment
removal efficiencies  must be compared  to
capital and operating costs, which  in-
crease dramatically as efficiencies
approach  100 percent.
                                            107

-------
 Process Economics

      Several  variables and/or alterna-
 tives in the  design and operation of a
 carbon treatment system can have a major
 impact on the economics of the process.
 These factors include:

           Type of carbon (GAC or PAC)
           Flow rate
           Contact time
           Process configuration (series,
           parallel,  or moving bed)
           Number of stages
           Flow direction (packed or
           expanded;  upflow or downflow).

 WET Model  Example

      Waste stream 02.02.14 from the WET
 Model was  selected as  an example to show
 a typical  carbon  adsorption  system de-
 sign, associated  material  balances,  and
 treatment  economics.

      This  stream,  with  a nominal  rate of
 426,000  kg/day  (17,750  kg/h,  based on 365
 days  per year operation)  contains  ben-
 zene, toluene,  and phenol  at  concentra-
 tions ranging from 3,000  to 5,000  ppm.
Although these  concentrations are  on the
 upper range of  the concentrations  cur-
 rently being treated in  commercial  prac-
tice, GAC  has been shown to be effective
for these  constituents at these  levels.
The composition of WET stream 02.02.14 is
given in Table  1.
                                  The process flowsheet  for carbon
                             adsorption pretreatment of  the above
                             waste stream  is shown  in Figure 2.  The
                             process design and configuration  shown
                             includes waste stream  preconditioning,
                             moving-bed adsorption, thermal reactiva-
                             tion, and furnace offgas cleaning units.
                             The preconditioning system  is designed to
                             remove suspended solids to  below 50 ppm
                             and to allow pH adjustment  of the waste
                             influent for maximum adsorption effi-
                             ciency of the carbon.

                                  The material balance for the pre-
                             treatment example is shown  in the tabular
                             portion of Figure 2.   Although the mate-
                             rial balance indicates essentially 100
                             percent removal of the three organics, in
                             actual  practice the waste effluent
                             (stream No.  8) may contain these constit-
                             uents at concentrations of several ppm.

                                  The major process uncertainties in
                             the design of this carbon adsorption are:
                             (1) the equilibrium capacity of the
                             carbon  for the three organics in a multi-
                             component aqueous solution of this par-
                             ticular composition,  (2)  the adsorber
                             residence (retention) time,  and (3)  the
                             carbon  recirculation  rate.   For this
                             case, a carbon loading of 0.3 kg adsor-
                             bate/kg carbon and a  minimum adsorber
                             residence time of 30  minutes was  assumed.
                             Also, to  size  the reactivation furnace,  a
                             furnace  residence time of 30 minutes was
                             assumed.   These assumptions  are  consist-
                             ent with  the  ranges  used  in  current
                             practice.
  Table 1.  Composition  of WET  Stream
      02.02.14, Quench  Blowdown  from
      Ethylene Production by Thermal
        Cracking of Heavy  Liquids
Component
Benzene
Phenol
Toluene
Solids
Water
Mass
fraction
0.005
0.003
0.004
0.01
0.98
Flow rate,
kg/h
88.8
53.3
71.0
177.5
17,359.5
       Total
1.002
17,750.1
Economics

     The capital and operating costs for
the above example case were based on
24-h/day, 330-day/yr operation, an ad-
sorber design capacity of 200,000 gal/
day (126 percent of waste stream rate
including recycle streams to the ad-
sorber), and reactivation furnace through-
put rate of 25,500 kg carbon/day.   The
capital costs of the major components of
the carbon treatment system including
support equipment, installation, engi-
neering, legal, financing, and adminis-
trative costs are presented in Table 2.

     The annual operating costs for the
system are also included in Table  2.   The
major operating costs include:   labor,
electricity, fuel  (natural gas), mainte-
nance materials, and carbon makeup.
                                           108

-------
                               ~l
                                     c
                                     OJ
                                     S JS
                                     k. en
                                     "tt <0

                                     s.*

                                     II
                                     <- S:
                                     D. to
                                     t N
                                     O ea
                                     8-=
                                     ro 
                                     si
                                     "
                                     o 92
                                     s 
                                     a- E
                                     O)

                                     iZ
109

-------





























CM
S
1
o

&















1
J
S
&
g
1
S
&












^
r
1=

S
<
ei
c -
|f

i~
c
i
o
c
>
g
2
c e
^
 C
15
13
w

id
3
M
S
S
g
01
N
i
1
1
o
c
S
3

to -5
S E
I
f^cor^mr^mcOf'^coLQ* o oo r*-
rC rC irT T-  ^f" CN~ ^r* co* CN co* co - ^-"
* CN CN CN
r*
^

i i i i i i i i i i i i i i



i i i i i i i i i i i i i ">


o
1 1 1 1 1 1 I I I i i i i CO
1 1 1 1 1 1 1 1 1 1 1 1 1 in
CM
	 8


o ooo
^ t^ r- r~
cocormu5ri*cor**'COco* r*o)co
tP*S'~'~CN'"SCD'~<0<>' *~
CO CO t- r- ^.
C: i t ' C: i v i " i i i i

g i g i i g i i i  i g g i

 i  i i  i i iSiSSi

* *~ t ^-
r. " - ' SP '

T CMco^incor^cooo^'CMco *^-

(ocNcnr-- r-nc:
ci - r~-


"~

1 1 1 1 S CM co |
t- r- CO


co co in
1 g 5 1 1 1 1 "*
CM

S 2 !2
, CO CO i . i i 00
1 in . 1 1 1 1 
CM" o>" TJ-"
CO


r.
CD
O T-
1 1 1  I 1 1 


11111111

1 1- 1 1 1 1 1 1

11111111


' 	

ncois*coojor-cN





i ,_ I 1 1 1 1

1 1 1 1 1 1 1



in
10 1 1 1 1 1 1


in
00 ...
*
, , ,', , ,


00

ST}" O O O
| O) O O | O


1 	

1 1 1 1 1 1 1

1111111


1111111

CO ^t" U3 CO r*^ CO C3>
CN CN C>J CN CN CM C*J
no

-------
             TABLE 2.   CAPITAL AND OPERATING COSTS FOR CARBON ADSORPTION
               PRETREATMENT WITH THERMAL REACTIVATION OF THE CARBON FOR
                               WET MODEL STREAM 02.02.14

      Capital Costs
        Influent pump station
        Carbon adsorption system (2 pulsed-bed contactors)
        Carbon regeneration system (fluidized-bed furnace)
        Carbon inventory (150,000 Ib @ $0.85/1b)
        Construction costs
                Total Installed Cost

      Annualized Operating Cost

        Operating labor (36,000 man-hours @ $15/m-h)
        Maintenance (5% of capital cost)
        Electricity (824,000 kWh @ $0.05/kWh)
        Steam (13,680,000 Ib @ $4/1,000 Ib)
        Fuel (1,430,000 therms @ $0.59/therm)
        Water (20,800,000 gal @ $0.40/1,000 gal)
        Carbon makeup (829,000 Ib @ $0.85/lb)
        Taxes, insurance, administration (4% of capital cost)
        Capital recovery (16.3%; 10% over 10 years)

                Total Operating Costs

        Product Recovery Credit
                Net.Operating Costs

        Waste treated (kg/yr)
        Total volatiles removed (kg/yr)

             Unit treatment, cost ($/kg waste treated)
                                 ($/kg volatiles removed)
                            21,000
                           181,000
                         1,925,000
                           128,000
                           925,000
                        $3,180,000
                           540,000
                           159,000
                            41,200
                            54,700
                           840,000
                             8,300
                           704,700
                           127,200
                           518,300

                        $2,993,400

                                 0

                        $2,993,400

                       155,490,000
                         1,865,880

                       0.019 $/kg
                       1.60 $/kg
       ^Construction fee (10%), contingency (15%), engineering (15%), startup (1%).
Advantages and Disadvantages

     The major advantages of carbon
pretreatment are:

          It is a mature technology in
          commercial use for waste treat-
          ment applications.

          Carbon adsorption can handle a
          broad range of organic constit-
          uents and concentrations.

The disadvantages of carbon pretreatment
include:

          Carbon adsorption treatment,
          especially with thermal reacti-
          vation, is a complex and labor-
          intensive operation.
          Carbon adsorption has substan-
          tial operating costs.

     Table 3 shows the results of the
analysis of other types of pretreatment
processes.  Applicable pretreatment
processes are shown for some typical
waste types that may contain volatile
constituents.  This table shows that air
stripping, steam stripping, or evapora-
tion (coupled with carbon adsorption of
the offgases); steam stripping; and
distillation are the most widely applic-
able techniques for volatile removal.

     Insofar as the hazardous waste
streams are typical of the application,
the pretreatment technique, the volatile
removal efficiency (kg volatile removed/
kg volatile in the stream), unit cost
                                          111

-------
     TABLE  3.  APPROPRIATE  PRETREATMENT
             PROCESS BY WASTE  TYPE

                          Applicable
   Waste type      pretreatment process(es)


 Organic liquids    Distillation
 Aqueous, up to
   20% organic

 Aqueous, less
   than 2%
   organic
 Sludge with
   organics
 Some sludge in
   organic or
   aqueous stream
Steam stripping
Solvent extraction

Steam stripping
Steam stripping with
  carbon adsorption
Resin adsorption
Air stripping with
 carbon adsorption
Ozonati on/radi olys i s
Wet oxidation
Biological treatment

Air stripping with
 carbon adsorption
Evaporation with carbon
  adsorption
Ozonation/radiolysis
Wet oxidation
Chemical oxidation
Evaporation with carbon
 adsorption

Physical separation
($/kg stream treated), and cost-effec-
tiveness ($/kg of volatile removed) are
typical of what may be expected if the
pretreatment technique were used on other
streams of a similar nature.

     One further important result of this
study is an engineering judgment regard-
ing the most applicable pretreatment
technique(s) for all WET Model streams.
This is presented (in the final report
for this project) in a matrix that
matches the 12 pretreatment techniques
considered in this report with all WET
Model streams that contain volatile con-
stituents.   Based on this matrix, the
following pretreatment techniques are
considered most applicable for removing
volatile constituents from the WET Model
waste streams.

          Air stripping, steam stripping,
          or evaporation/carbon adsorp-
          tion
          Steam stripping
          Batch distillation.

SUMMARY AND CONCLUSIONS

     In summary, this paper is a prelim-
inary examination of various pretreatment
techniques for the removal of volatile
compounds from hazardous waste streams
identified in one data base.  Although
the conclusions and analyses herein may
not be regarded as final, this paper does
provide insight into the potential  applic-
ability of pretreatment to reduce emis-
sions of volatile compounds from TSDFs.
                                                    The  conclusions of this  investiga-
                                               tion are:
          An analysis of potential  air
          emissions "from treatment,
          storage, and disposal  of  72
          hazardous waste streams indi-
          cated that a substantial  frac-
          tion of volatile hazardous
          components could be emitted to
          the atmosphere from landfills,
          land treatment facilities,
          surface impoundments,  and
          cooling towers.   It is reason-
          able to assume that in many
          cases,  removal of organics  by
          pretreatment would cause  a
          comparable reduction in air
          emissions.
                                     Pretreatment of these hazardous
                                     waste streams could remove most
                                     (90 to 99 percent) of the
                                     volatile materials.  A number
                                     of alternative processes are
                                     available for pretreatment for
                                     most of the waste streams.
                                     The cost-effectiveness varied
                                     from $70 to over $2000/Mg of
                                     volatiles removed.   The cost-
                                     effectiveness of pretreating
                                     specific waste streams is not
                                     necessarily in the  same range
                                     as the example cases.   The
                                     actual cost-effectiveness would
                                     depend on the chemical and
                                     physical characteristics of the
                                     waste stream(s),  the design
                                     capacity of the pretreatment
                                     system,  and the degree of
                                     volatile removal  required.
                                           112

-------
           Pretreatment techniques using
           carbon adsorption, steam strip-
           ping, or batch distillation are
           the most applicable ones for
           the WET model streams evalu-
           ated.  This judgment considers
           cost, the range of applicabil-
           ity, and the extent to which
           the technology has been demon-
           strated.
REFERENCES

Breton, M., T. Nunno, P. Spawn, W. Farino,
     R. Mclnnes.  1983.  Assessment of
     Air Emissions From Hazardous Waste
     Treatment Storage and Disposal
     Facilities (TSDFs), Preliminary
     National Emissions Estimates, Draft
     Final Report, EPA Contract 68-02-
     3168, August 1983.
U.S. Environmental Protection Agency.
     1971.  Wastewater Treatment Facil-
     ities for a Polyvinyl Chloride Pro-
     duction Plant.  U.S. EPA Project
     No. 12020 DJI.

U.S. Environmental Protection Agency.
     1983a.  National Survey of Hazardous
     Waste Generation and Treatment,
     Storage, and Disposal Facilities
     Regulated under RCRA in 1981.
     Preliminary Highlights of Findings.
     August 30, 1983.

U.S. Environmental Protection Agency.
     1983b.  W-E-T Model Hazardous Waste
     Data Base, 2nd draft.  Prepared by
     SCS Engineers for Office of Solid
     Waste.  July 21, 1983.  This data
     base was incorporated in slightly
     revised form in the "RCRA Risk Cost
     Analysis Model Phase III Report,"
     Office of Solid Waste.  March 1984.
                                           113

-------
                   PROGRAMS OF THE INDUSTRIAL WASTE COMBUSTION GROUP
                          AT THE U.S. EPA CENTER HILL FACILITY
                                   George L.  Huffman
                        Chief, Industrial Waste Combustion Group
                           Energy Pollution Control  Division
                          U.  S. Environmental Protection Agency
                                Cincinnati, Ohio  45268
                                        ABSTRACT
    The primary focus of EPA's in-
house Thermal Destruction research
being conducted at their Center Hill
Facility (part of their Cincinnati-
based Industrial Environmental
Research Laboratory) is to:

     *    Establish how combustion
          variables affect failure of
          a simulated hazardous waste
          incinerator or a simulated
          industrial boiler that
          co-fires hazardous waste
          with conventional fuel to
          achieve 99.99% ORE (Destruc-
          tion and Removal Efficiency);

     *    Determine how, when and why
          PIC's (Products of Incomplete
          Combustion) are formed
          (e.g., too low an oxygen con-
          centration or too low an
          operating temperature) and
          to determine how or whether
          they can be subsequently
          destroyed or removed (e.g.,
          by secondary combustion or
          scrubbing/adsorption tech-
          niques); and

         Determine which organic
          chemicals are the hardest
          to burn [this will assist
          in making better selections
          of POHCs (Principal Organic
          Hazardous Constituents)].
 Key research being  done at Center
    Hill  features:

 Conducting parametric testing, in
 bench-scale and mini-pilot combus-
 tors  on simulated hazardous waste
 mixtures  to determine the range
 of operating conditions under which
 efficient hazardous waste destruc-
 tion  can  be expected and where
 PICs  can  be minimized.  The combus-
 tion  equipment located there
 consists  of a bench-scale non-flame
 "TDU-GC"  (Thermal Destruction Unit-
 Gas Chromatograph), a "Microspray
 Reactor"  (a laminar-flow inter-
 mediate-size unit having a quiescent
 flame), a "Turbulent Flame Reactor"
 (a larger, mini-pilot, water-jacketed
.unit  having the ability to induce
 turbulence into a flame), and a
 "Controlled Temperature Tower" (a
 refractory-lined, 4-foot outside
 diameter  by 7-foot  high reactor
 which provides for  turbulence,
 secondary combustion and simulated
 cold  walls).  The operation of each
 of these  units will be supported
 by an onsite GC/MS  to provide
 comprehensive analytical capability.

 Examining key combustion variables
 to determine their  impact on POHC
 DRE's and on PIC formation and des-
 truction.  Key variables will
 include residence time, operating
                                            1T4

-------
    temperature, degree of turbulence,
    degree of atomization, pressure drop
    across the burner nozzle, oxygen con-
    centration, air-to-fuel  ratio, waste-
    to- fuel ratio, waste mixture constitu-
    ent impacts, secondary combustion
    impacts, and the impacts resulting
    from quenched destruction react-
    ions due to cold walls in combustors.

*   Carrying-out parametric testing on key
    process variables attendant to the
    operation of a two-stage air pollution
    control system (consisting initially
    of a packed-bed scrubber for HC1 con-
    trol and an activated carbon bed for
    unburned hydrocarbon removal) to define
    the amounts of pollutant removal that
    is possible (i.e., to provide data
    on the "R" part of the ORE).

    Specific hypotheses that will be
addressed in this In-House Thermal Destruc-
tion Research Program include:

    (1)  The ORE performance of a hazardous
         waste combustor can be predicted
         by CO/C02 or THC (total hydrocar-
         bon) in the exhaust gas; conse-
         quently, continuous monitoring
         based on one or more of these
         parameters may be suitable as a
         compliance monitoring technique.

    (2)  Flame cooling (due to flame
         proximity to cooler heat trans-
     fur surfaces) and post-flame heat
     heat loss can have a marked effect
     on overal 1 ORE performance and on
     PIC formation.

 (3) The standardized POHC test mixture
     used in EPA's ongoing full-scale
     test burns is a representative mix-
     ture for characterizing the ORE
     and PIC formation/destruction per-
     formance of a given combustor
     across a range of operating con-
     ditions.

(4)  High levels of PIC's can be fed
     into the post-flame zone during
     the combustion of hazardous waste,
     and they can be significantly
     lessened there due to the large
     gas residence times that exist
     in that zone, or the opportunity
     for secondary combustion there.

(5)  A two-stage scrubber/carbon bed
     system can effectively remove HC1
     and organics formed in or passed
     through thermal destruction
     devices burning hazardous wastes.

     The first three hypotheses will be
tested in the flame-mode Turbulent Flame
Reactor/Controlled Temperature Tower,
the fourth hypothesis in the non-flame
TDU-GC system and the fifth, obviously,
in the scrubber system discussed above.
                                           115

-------
                       DETERMINATION OF THE THERMAL DECOMPOSITION
                  PROPERTIES OF 20 SELECTED HAZARDOUS ORGANIC COMPOUNDS
                                     Barry Del linger
                                     Juan L. Torres
                                     Wayne A. Rubey
                                     Douglas L. Hall
                                     John L. Graham
                                  University of Dayton
                                   Research Institute
                                   Dayton, Ohio 45469

                                    Richard A. Carnes
                           U.S. Environmental Protection Agency
                             Combustion Research Facility/NCTR
                                    Jefferson, AR 72079

                                        ABSTRACT

     Laboratory determined thermal decomposition profiles and kinetic data for a list of
20 selected hazardous organic compounds are reported.  All data were obtained in flowing
air at mean gas-phase, high-temperature zone residence times ranging from one to six
seconds.  The extrapolated temperatures required for 99.99% destruction of the parent
compound at two seconds mean residence time, Tgg.g9(2), ranged from 600C for 1,1,1-
trichloroethane to 950C for acetonitrile.  The processes and parameters potentially con-
trolling incineration efficiency are discussed, and four previously proposed methods of
ranking compound incinerability are reviewed.

     The possible chemical mechanisms for destruction of hazardous organic compounds are
examined and used to explain trends in the experimentally determined thermal d.ecomposition
data.  It is proposed, through proper application of the principles of organic chemistry,
kinetics, and physics, that laboratory, gas-phase thermal decomposition data generated
under controlled conditions can be incorporated into models of full-scale incineration,
serve as a viable ranking of waste incinerability, and used to predict the formation of
products of incomplete combustion.
INTRODUCTION

     The ultimate goal of hazardous waste
incineration is to destroy the waste mat-
erial with as high a destruction efficiency
(OE) as possible.  Under the Resource
Conservation and Recovery Act (RCRA) of
1976, an incinerator operator must show
that the facility can adequately destroy
those hazardous waste constituents which
are most difficult to incinerate.  In
theory, the permit writer will select com-
pounds within the mixture which are of
sufficient toxicity, concentration, and
thermal stability so as to be designated
as principal organic hazardous constitu-
ents (POHCs).  It must then be shown,
possibly by trial burn, that the designated
POHCs can be destroyed or removed by the
particular incineration system to a
destruction and removal efficiency (DRE)
of 99.99 percent.

     The development of a ranking of the
incinerability for compounds which are
candidates for POHC selection is of obvious
utility.  The US-EPA currently is using a
ranking based on the heat of combustion per
gram of pure compounds.  This method has
received considerable criticism, and the
development of an alternative ranking
scheme is of a very high priority to the
EPA.

     Experimentally determined gas phase
thermal stability under controlled
                                           116

-------
laboratory conditions has been proposed as
an alternative ranking method.  This
report presents the results of the labora-
tory determination of the gas phase thermal
decomposition properties of twenty (20)
hazardous organic compounds.   The com*)
pounds were selected by the EPA based on
their frequency of occurrence in hazardous
waste streams, apparent prevalence in the
stack effluent, and representativeness of
the spectrum of hazardous organic waste
materials.

EXPERIMENTAL PROCEDURES

     All of the experimental data pre-
sented in this  study  were generated using
the thermal decomposition unit-gas
chromatographic (TDU-GC) system which was
designed and built with funding provided
by the US-EPA (Cooperative Agreement No.
807815-01-0)  [!].  Samples of the 20 com-
pounds were prepared and introduced to the
system by several procedures depending
upon their physical state and vapor
pressure.

     To initiate a test, the sample is
introduced into the system and gradually
vaporized in a flowing gas stream (i.e.,
nitrogen, air, or nitrogen/oxygen mix-
tures).  The vaporized sample passes
through a controlled, high-temperature
tubular reactor where it undergoes thermal
decomposition.  The products evolving from
the thermal decomposition of the compound
and the remaining parent compound are
swept into a Varian VISTA 4600 high-
resolution gas chromatograph for analysis.
The sample insertion chamber, the reactor,
and the entire transport system are fabri-
cated of fused quartz to minimize inter-
action with the sample.

     In the TDU-GC system, the sample is
deposited  in a sample  insertion chamber
which is packed with quartz wool.  The
chamber is slowly heated to 250C to 300C
by applying a linear temperature program
(10-20C/min).  The sample molecules are
thermally desorbed and gradually swept into
the thermal reactor.

     The sample emerging from the reactor
is trapped at the head of the chromato-
graphic column by maintaining the GC oven
at a cryogenic temperature ('Mninus 30C).
A 15 meter dimethyl silicone chemically
bonded stationary phase fused silica capil-
lary column in conjunction with a flame
ionization detector was used in the major-
ity of the investigation.

     For the ethane analyses, Tedlar bags
were used to capture the reactor effluent
at the splitter.  The captured gas was sub-
jected to analyses using a Varian Aerograph
Series 1800 gas chromatograph equipped with
one meter x 4.0 mm ID packed column (5 K
molecular sieve, 45/60 mesh) operated
isothermally at 150C.  This system was
utilized to ensure resolution of the three
C2 hydrocarbons.

RESULTS

     For each of the twenty (20) test com-
pounds, the fraction of the feed material
undestroyed at a given set of temperatures
and mean reactor residence times (tr) was
determined.  This resulted in the genera-
tion of what may be termed thermal decom-
position profiles, i.e., a semilogarithmic
plot of fraction remaining (fr)
vs. the reactor temperature (C) at con-
stant residence time.  Thermal decomposi-
tion profiles were generated in flowing
air at four mean residence times, ip = 1.0,
2.0, 4.0, and 6.0 seconds.  An example of
the determination of this family of pro-
files is given in Figure 1 for chloroform.
The thermal-decomposition profile for
chloroform is representative of the great
majority of the compounds tested and serves
to illustrate several features.

     The data for the test compounds have
been summarized in Table 1 with entries
for the temperature for the onset of decom-
position, T0nset (2) (C), the inter-
polated temperature for 99% destruction,
Tgg (2) (C), and the extrapolated tempera-
ture for 99.99% destruction, Tgg gg (2)
('C).  All these values are for tr = 2.0
seconds in flowing air.  Using only the
data presented in this table, the thermal
decomposition profile for the compounds
may be approximately reconstructed.  The
table lists the compounds in order of
decreasing temperature required for 99%
destruction efficiency.  A slight reorder-
ing occurs if Tgg^g(2) is used for rank-
ing.  However, the numerical differences
for the reordered compounds are small.

     For the conditions possibly
encountered during gas-phase thermal decom-
position in an incinerator (600C to
1,400C and oxygen levels of 0.1 to 21%),
two possible global decomposition pathways
                                           117

-------












cC
O


o
H- 
^4
t/>
2
IE
O
0
LU
a

t
S*
S
LU
U.
O
^_
cc

n
l/>
r

Ul

CO
&
1


















"o
E

LO r
re
u

^-^
i

 
"*

^^
CM
*"cn^~
encj
ov 
en
I-

^ i
CM ^"^
^-^CJ
cno
ov  


CM
| \ 
'('"* '
cucj
LOO

H?~

i
re re
O 
r- 3
J- S
r- i.
0. 0
E Lu
LU






"2

o
o.


O






O




f*^
O
X
r-
<3-




o
in
en





o
o
en




o
to
t-.





2
CO
IE
CM







0)

r-
S.
4J
r
C
O
+J
a)
u
>

QJ OJ
H


P co ^~ CD *d~ "^i* ^o cn en en *=i* co f~~ " co "d* en en CM
CO-=t--*CO CMVDCM LO CO COCMCOI~-.CMlOCM-^-CO


CO CM LO LO CM
locncoio mi  in i  co oo LO co i =) i  r--i  co
OOOO OOO O CD OOOOOOOOO
XXXX XXX X X XXXXXXXXX
coLOLoen i  oco co *cncncri
i  COCMi  i  COCM IO CM COi  CMCncO"  i  CMi 




OOCDO OOO CD CD OLOOOOOOOO
usr^ooin ^J-COCM co cn coco^DLooocDd-CMCD
co co co co cot^co t-~ r~- r-~r--i  p-.t~.t-~u3%j=re >-> -r- jr xi i  too) c i roNi  oi-reo .c
i-4-> XCMCU S- O i--C X-r-CMCUCM^: C-^ C+J Xi  i 	 P
o cu cu "I >,!- re o cu "a .a 4->cuco-i-cu^: cu
eCSICf" Q.QCJ 1C i    ; LUCCl=CSs:inCJi 
118

-------
         100
          10
       8-
       iu
         O.I
         0.01
CHLOROFORM
 O tr= 1.0
 n fr = a.o
 A fr =4.0
 o tr=e.o
                        100
                               400
                                            5OO
                                    EXPOSURE TEMPERATURE, C -
                                                         600
                                                                       700
           Figure 1.
       Thermal Decomposition Profiles for Chloroform in  Flowing
       Air at Mean Residence Times of 1.0, 2.0, 4.0, and 6.0
       seconds.
predominate, pyrolysis and oxidation.[2]

     The global expression  for these two
reaction schemes are
      dt
          = k][A]a+ k2[A]a[02]b.
                        (1)
                                        large excess of molecular oxygen relative
                                        to the concentration of the waste material,
                                        the decomposition equation could be simpli-
                                        fied to an expression that is first order
                                        in the concentration of the sample.  This
                                        resulted in the integrated pseudo-first
                                        order rate expression:
where:  k-i and k2 are the global rate con-
        stants for pyrolysis and oxidation,
        repsectively, and a, a, and b, are
        the reaction order for the decom-
        position of species A with respect
        to A and 02.

The time dependence is included in this
expression.  The temperature dependence is
included'in the rate constants for the two
processes which  may be expressed by the
Arrhenius equation:
                                          fr = exp (-k2' tr)
                                                                                (3)
        k = A exp (-Ea/RT)
                         (2)
                                 where:  fr is the fraction of the parent
                                            species remaining, and
                                         k2' = k2[02l is the pseudo-first
                                            order rate constant

                                 This expression may be combined with the
                                 Arrhenius Equation to yield an expression
                                 for the required temperature for a given
                                 level of destruction in an atmosphere of
                                 flowing air:
                                                                   -1
where:
Ea is the activation energy for the
   process, cal mole"'
A  is the Arrhenius coefficient, s""1
R  is the universal gas constant,
   1.99 cal mole"1 K-1.
     For each of the test compounds, it was
found that when .the thermal decomposition
reaction occurred in an atmosphere with a
                                                          = BOS E
                                                                 a
                                                      In
                                                           In
                                                                                (4)
                                                where:
                                             is the temperature required for
                                             a given DE,  K,
                                             has units of kcal mole"  and
                                             the other variables are as pre-
                                             viously defined.;
                                 A plot of In fr vs. tr for the four
                                            119

-------
residence times will yield the rate con-
stant for the reaction at a given tempera-
ture.  A plot of In k vs. 1/T for the four
experimental temperatures should then
yield a straight line with the slope equal
to -Ea/R and an intercept of In A.

     Regression analyses of this type have
been performed on each of the twenty (20)
test compounds.  In all cases, the first
order kinetic plots yielded far better
fits than zeroth or second order plots.
The measured kinetic parameters, A and Ea,
are included in Table 1.

DISCUSSION

     Several methods have been proposed
for ranking the relative incinerability of
hazardous organic compounds.  These
include heat of combustion per gram mole-
cular weight (He/gram of the pure compound)
autoignition temperatures (AIT) of pure
compounds, theoretical approach based on
the kinetics of flame mode thermal decom-
position, and gas-phase thermal decomposi-
tion of pure organic compounds in flowing
air [3-7].  Other parameters can be envi-
sioned which might be appropriate as a
basis for such a scale.  An attempt was
made to correlate laboratory generated gas-
phase thermal decomposition data reported
in the previous section with the previously
proposed rankings scales.  This comparison
was complicated by lack of overlap of the
compounds investigated.

     For the heat of combustion scale, the
only agreement that could readily be dis-
cerned was the general increase in thermal
stability with decreasing heat of combus-
tion for the chlorinated benzenes.

     There appeared to be a general posi-
tive correlation for both Tgg(2) and
Tgg.ggU) with AIT for those compounds
with an AIT below 550C.  Above this tem-
perature the gas phase thermal stability
appeared to vary little with AIT.

     Direct comparison of thermal decom-
position data with the theoretical ranking
was hampered by the fact that only four of
the 20 compounds were ranked by the group
at NBS who developed the approach.  How-
ever, the thermal stability of the chloro-
benzenes ranked by NBS fell  in the same
order as suggested by the Tggsgg(2i) based
ranking.  Chloroform was predicted to be
rather unstable due to the relatively low
energy carbon chlorine bond.  This thermal
instability is evident from the gas-phase
thermal decomposition results, although it
appears to be even more fragile than pre-
dicted by the NBS ranking.

     Only five of the test compounds were
studied at Union Carbide.  The most signi-
ficant disagreement was for acrylonitrile
which was the least stable of the five
compounds based on the Union Carbide cal-
culated Tgg.gg(2), while the data from
this study indicates its thermal stability
rivals that of methane.  Ethane is slightly
less stable than monochlorobenzene, as
measured by Union Carbide, while this trend
is reversed in the present data, although
the difference is small.  With the excep-
tion of acrylonitrile, the two rankings are
similar although the data from this study
predicts the compounds to be typically more
stable by 40C than that predicted by
Union Carbide.

     No single proposed ranking scheme or
molecular parameter was identified which
correlated with all our thermal decomposi-
tion data, although trends were observable
in homologous subgroups.  By careful consi-
deration and utilization of the principles
of chemical reactions, it is possible to
explain the behavior of each of the twenty
compounds on a relative basis and identify
possible mechanisms that can be used to
extrapolate this limited data to other com-
pounds not studied in the laboratory.  The
20 test compounds may be divided into five
subclasses; these classes are discussed in
the following paragraphs.

Methane, Dichloromethane, Chloroform, and
  Carbon Tetrachloride

     The observed trend in this group is
decreasing thermal stability with increas-
ing chlorine substitution, except for car-
bon tetrachloride which is intermediate in
thermal stability between methane and
dichloromethane

     The data are in agreement with a mech-
anism based on abstraction of a hydrogen
probably by an electrophile.  Since the
carbon-hydrogen bond dissociation energy
(BDE) decreases with increasing chlorine
substitution up to chloroform, one would
predict decreasing thermal stability,
which is observed.  However, carbon
tetrachloride contains no hydrogens and
thus would not be susceptible to this mode
                                           120

-------
of attack.  It would instead be expected to
decompose by bond rupture.

Benzene, Monochlorobenzene, 1,2-
  Dichlopobenzene, 1,2,4-Trichlorobenzene,
  1,2,3,4-Tetrachlorobenzene, Hexachloro-
  benzene, Pyridine, Aniline, Nitrobenzene

   1  The observed trend is increasing ther-
mal stability with increasing chlorine
substitution.  Pyridine is more stable,
nitrobenzene is less stable, and aniline
has about the same thermal stability as
benzene.  All of the bonds in benzene, the
chlorinated benzenes, and pyridine are
probably in excess of 90 kcal/mole and one
would expect electrophilic addition to be
the predominant reaction path.  The chlor-
ines and the nitrogen in pyridine are more
electronegative than hydrogen or carbon
which leads to a destabilization of the
electron deficient intermediate resulting
from electrophilic addition, thus reducing
the rate of decomposition and greater ther-
mal stability than benzene.

     In cases of aniline and nitrobenzene,
their stability relative to benzene can
also be explained by electrophilic attack.
A resonance stabilized intermediate may be
envisioned for nitrobenzene, the nitrogen
carbon bond dissociation energy (BDE) is
70 kcal/mole which may be easily broken and
therefore, represents an alternative mode
of decomposition.

Ethane 1,1,1-Trichloroethane,
  Hexachloroethane

     ,0ne might expect ethane to be
destroyed by unimolecular decomposition
through bond rupture at the weakest bond.
Due to the 88 kcal/mole carbon-carbon bond
one would predict significantly less sta-
bility than methane but still moderate
thermal stability for ethane.

     Although similar in structure, one
would expect the  pathway for decomposition
for 1,1,1-trichloroethane to be concerted
elimination of HC1 which is a very low
energy process [10].  Based on this, one
would expect this compound to be one of the
least stable compounds studied, which is
in fact,  the case.  Hexachloroethane would
have to eliminate Cl2 to proceed by a
concerted pathway.  This process is more
endothermic and the decomposition of
hexachloroethane would instead be expected
to proceed through  carbon-chlorine or
carbon-hydrogen bond rupture, both BDE's
being approximately 73 kcal/mole.  Based on
these considerations, one would expect
hexachloroethane to be intermediate in
stability between ethane and 1,1,1-
trichloroethane, which is the observed
trend.

Tetrachloroethylene and Hexachlorobutadiene

     Both of these compounds are quite
stable.  This is probably due to the large
BDEs caused by sp2 hybridization of all
carbon atoms and the lack of hydrogen atoms
available for abstraction.  Electrophilic
addition however, may be the predominant
mode of attack under incineration condi-
tions.  One would then expect hexachloro-
butadiene to be thermally less stable
since it can form an ally! radical inter-
mediate which is known to be quite stable.

Acetorn'trile and Acrylonitrile

     These compounds are very stable and
all BDEs are 93 kcal/mole or greater.
Carbon nitrogen triple bonds are expected
to be much less reactive towards electro-
philic addition than double bonds.
Apparently, the only mode of decomposition
for acetonitrile is loss of a hydrogen
through bond rupture or abstraction.
These are both high energy processes which
account for its stability.  Acrylonitrile
may, on the other hand, be susceptible  to
addition at the carbon carbon double bond.

CONCLUDING REMARKS

      Preliminary calculations indicate
that  "fault" or "failure" modes of incin-
erator operation, or equivalently the
"extremes" of operational parameter dis-
tribution functions, may well control
measured incineration efficiency for full-
scale units [11,12].  Following this line
-of reasoning, one would conclude that  the
incinerator effluent would only contain
undecomposed feed material and products
of incomplete combustion  (PICs) which  were
formed under these conditions of failure.
Thus, "fault" modes are essentially worst
case  conditions and appear to have a
dominating effect on the composition of
the  incinerator effluent.

      A point to be made along this vein is
that  the temperature at which this study
was  conducted is probably representative
of fault modes.  The experimental
                                            121

-------
laboratory temperature range covered 0 to
99.935 destruction of the feed material which
is typically several hundred degrees below
mean temperatures quoted for hazardous
waste incineration.  If a given incinerator
does not meet the 99.99% destruction effic-
iency requirement, yet has a high mean
operating temperature, then a likely possi-
bility for its failure is that a fraction
of the waste feed experiences temperatures
somewhat lower than the mean (where destruc-
tion efficiency is low) i.e., the destruc-
tion efficiency and temperature range
measured in laboratory studies.

     Thus, one might expect the actual PICs
emitted from the incinerator to be the same
as those formed under the conditions
studied in the laboratory.  Furthermore,
this reasoning suggests that the relative
thermal stability of hazardous wastes
should be compared at fault mode tempera-
tures since only this fraction of the waste
is escaping incineration.  Since selecting
a suitable temperature for comparison of
every compound is difficult and still some-
what arbitrary, a ranking based on the tem-
perature required for 99% destruction at
2 seconds mean residence time wouldvseem to
be a reasonable measure of POHC destructa-
bility.  One could just as well select 90%
or 99.9%, but examination of the data shows
that the rankings over this range are
essentially identical.

     Products of incomplete combustion have
not been determined in this study.  PIC
determination would remove much of the
speculation in the discussion of reaction
mechanisms.  In previous research, the
formation of numerous PICs from a wide
variety of organic compounds has been
observed [13].  These PICs have, in some
cases, been produced in as much as 30%
yields and been as hazardous or more haz-
ardous than the parent compound.  The det-
ermination of PICs should be an integral
part of future research programs and consi-
dered for inclusion in an ultimate ranking
of waste incinerability.

     The reported research has addressed
non-flame, high-temperature, gas-phase
reaction chemistry.  Extension of this work*
to include so-called flame mode studies
represents a special challenge due to the
difficulties in scaling results.  Of all
incineration processes which must be
modeled to full-scale, gas-phase chemical
kinetics is the easiest and most success-
fully performed.  The temporal  and spatial
distributions present in small  laboratory
or bench-scale flames are not easily scaled
to the turbulent poorly-defined flames pre-
sent in full-scale systems.   Consequently,
an elementary chemical  kinetic  approach to
determining flame mode  destruction
efficiencies might prove most effective.

     The importance of hydroxyl radicals in
flames is well documented [14].  Thus, an
experimental program to determine the rate
of attack of hydroxyl radicals  on hazardous
wastes would produce kinetic results which
is easily scaled.  Kinetic data of this
type in combination with measurement or
estimation of hydroxyl  radical  concentra-
tions in full-scale systems  would allow
simple scaling of laboratory results to
full-scale.  These data over different
temperature ranges would be  applicable
to both flame and non-flame  modes of
destruction.

REFERENCES

1.  Rubey, W. A., Fiscus, I. B., Torres,
    J. L., Description and Operation of a
    Thermal Decomposition Unit-Gas
    Chromatographic System,  prepared for
    US-EPA under Cooperative Agreement
    CR-807815, 1984.

2.  K. J. Laidler, Reaction  Kinetics,
    Vol. 1., Homogeneous Gas Reactions
    eds., W. C. Agosta and R. S. Nyholm,
    Appleton-Century-Crofts, New York, NY,
    1971.

3.  E. P. Grumpier, E.  J. Martin, and
    G. Vogel, Best Engineering Judgement
    For Permitting Hazardous Waste
    Incinerators, Presented  at ASME/EPA
    Hazardous Waste Incineration
    Conference, Williamsburg, Virginia,
    May 27, 1981 .

4.  J. J. Cudahy, L. Sroka,  and W. TroxTer,
    Incineration Characteristics of RCRA
    Listed Hazardous Hastes, IT
    Environscience Draft Report Submitted
    under EPA Contract 68-03-2568, July,
    1981.

5.  W. Tsang and W. Shaub, Chemical
    Processes in the Incineration of
    Hazardous Materials, presented at
    American Chemical Society Symposium
    on Detoxification of Hazardous Wastes,
    New York, NY, August, 1981.
                                           122

-------
 6.  B. Dellinger,  J.  L.  Torres,  and R.  A.       11.
     Carnes,  The High-Temperature Gas-Phase
     Oxidation of Selected Chlorinated
     Aliphatic and  Aromatic Hydrocarbons,
     presented before  the Division of
     Environmental  Chemistry.of the 185th
     American Chemical  Society Meeting,
     Seattle, WA, March,  1983.

 7.  K. C. Lee, N.  Morgan, J. L.  Hansen,        12.
     and G.  M. Whipple, Revised Model  for
     the Prediction of the Time-Temperature
     Requirements for  Thermal Destruction
     of Dilute Organic Vapors and its  Usage
     for Predicting Compound Destructabil-
     ity, Presented at 75th Annual Meeting
     of the Air Pollution Control
     Association, New  Orleans, June, 1982.       13.

 8.  J. A. Kerr and A. F. Trotman Dickinson,
     Bond Strength  of  Polyatomic  Molecules,
     IN:  CRC Handbook of Chemistry and
     Physics, Chemical  Rubber, Co.,
     Cleveland, OH  59th Ed., 1979.

 9.  S. W. Benson,  Bond Energies*, Journal        14.
     of Chemical Education, Vol.  42, No. 9,
     pp. 502-518, September, 1965.

10.  S. W. Benson,  and H. E. O'Neal,
     Kinetics on Gas Phase Unimolecular
     Reactions, NSRDS  Report NBS-21, U.S.
     Government Printing  Office (1970).
B. Dellinger, J. L. Torres, W. A.
Rufoey, D. L. Hall, J, L. Graham,
Determination of the Thermal
Decomposition Properties of 20
Selected Hazardous Organic Compounds,
Final Draft Report prepared for
US-EPA under Cooperative Agreement
CR-807815,, 1984.

J. C. Kramlich and W. R. Seeker,
Laboratory-Scale Flame Mode Study of
Hazardous Waste Incineration,
Presented at Ninth Annual Research
Symposium on-Solid and Hazardous
Waste Disposal, Ft. Mitchell, KY,
May, 1983.

D. L. Hall, B. Dellinger, and W. A.
Rubey, Considerations for the Thermal
Degradation of Hazardous Haste,
Prepared for Presentation at 1983
(4th) International Symposium on
Environmental Pollution, Miami Beach,
Florida, October, 1983.

The Mechanisms of Pyrolysis, Oxidation,
and Burning of Organic Materials,
L. A. Wall, ed., Proceedings of the
Fourth Materials Research Symposium
held by NBS, Gaitherburg, MD, NBS
Special Publication 357 CODEN:XNBSAV
(1972).
                                            123

-------
                  HAZARDOUS WASTE DESTRUCTION USING PLASMA ARC TECHNOLOGY

                                     Thomas  G.  Barton
                              Pyrolysis  Systems Incorporated
                                 Well and,  Ontario  L3B 5P1

                                     Nicholas P.  Kolak
                  New York State Department  of  Environmental Conservation
                           Division of  Solid and  Hazardous Waste
                                  Albany,  New York  12233

                                      Chun Cheng  Lee
                            U.S. Environmental  Protection Agency
                       Industrial Environmental  Research Laboratory
                                  Cincinnati, Ohio  45268
                                          ABSTRACT
       A  thermal  plasma  properly applied  to toxic  and  hazardous waste  destruction pro-
vides  a  pyrolytic  environment with  distinct  advantages over  other competing  technolo-
gies.   Complete  atomization  of  organic fluids  have been  demonstrated to occur  in less
than one  third  of a millisecond.  Kinetic  recombinations of atomic  entities  are predict-
able  and  generally form acid  gas,  fuel  gas and  graphitic soot.   Since the 'process  is
pyrolytic,  the  scale of the  equipment  is  small  for  high throughput   rates.    Energy
requirements  for  the destruction of organic  fluids are typically  less  than  one  kilowatt-
hour  per  kilogram  of  waste.  A  mobile prototype  is being  constructed for  a  throughput
rate  of four  kilograms per  minute  of liquid organic waste.  The  mobility and  efficiency
of such systems may have significant  advantages  over fixed facilities.
         An  assessment  of  approximately
  forty emerging waste destruction techno-
  logies  was  conducted  by the Environmental
  Protection   Agency  (EPA).     Five  were
  selected for  demonstration  since  they
  showed   a high potential  for successful
  application to toxic and hazardous waste
  destruction.   Under  a co-operative agree-
  ment between the  New York  State Depart-
  ment   of    Environmental'    Conservation
  (NYSDEC) and the EPA Industrial  Environ-
  mental   Research  Laboratory  (IERL)  1n
  Cincinnati  a demonstration   program  was
  established and jointly funded to further
  Investigate mobile plasma pyrolysis.   A
  contract   was   awarded   by   NYSDEC   to
  Pyrolysis    Systems    Incorporated    of
  Hell and, Ontario  in  1983 for  the fabrica-
  tion  and  demonstration  of  the  mobile
  prototype.
        The  plasma  arc device  forms  the
 heart of  this technology.   A co-linear
 electrode arrangement  creates  an electric
 arc  which   is  stabilized  by  field coil
 magnets.  Low oressure air  is  used as a
 medium   through   which    an   electric
 current  1s  passed.    In  passing  through
 the   air,   the  electrical   energy   is
 converted to thermal energy  by  absorption
 by the  air  molecules  which  are activated
 into ionized atomic states, losing elec-
 trons in the process.   Ultraviolet  radia-
 tion  is emitted when the  molecules  or
, atoms relax  from  their highly activated
 states to lower  energy levels.
        Waste fluids can be  injected into
 'this plasma zone to interact with  decay-
 ing  plasma  species   to   become  rapidly
 atomized prior to exiting this electrode
 space.  An  integral part  of this process
                                           124

-------
                             ATOMIZATION  m  EQUILIBRATION   P  QUENCHING
                            Figure  1.   Process Schematic.
technology  is  the  pyrolytic  simulation
computer model  .   New compounds which  are
created from  the  recombination of  atomic
species are  predictable based on kinetic
equilibrium.   The minimization of  Gibb's
free  energy  is  used  to  determine  the
equilibrium   concentration   of   product
species  over  a  wide  range  of   selected
temperatures  and  pressures.  Thus,  appro-
priate  operating   conditions  can  be  pre-
dicted, for  any organic waste  fluid prior
to  its  destruction.   Similarly,  undesir-
able products  can be minimized or  elimi-
nated through  altering  the character of
the  fluid  (blending)  or  the  operating
conditions.   Changes in  enthalpy between
feedstocks  and  predicted  output  products
are  used  to  determine  the  plasma  energy
required  adiabatically  for  the  pyrolysis
of the waste being treated.
       A simplified  process  schematic for
the  mobile prototype  1s  shown  in Figure
1.  A five hundred  kilowatt  plasma device
is fitted'  to  one end of a stainless steel
reaction  chamber and  mated  to  a hollow
graphite   core   to   form  an   atomization
zone.  Residence time  in this  atomization
zone is  approximately  five hundred micro-
seconds.   The reaction chamber serves as
the  equilibration  zone where the  atomized
species  recombine to  form  new products.
This zone  is  equilibrated at  a  temperature
range  of  120G-1800  Kelvin  and  the resi-
dence  time in this  zone is  approximately
one second.
       A three stage spray ring  is  located
at the producer  gas outlet from  the plasma
reaction   chamber.    The  spray  ring   is
injected   with  water  and  appropriate
amounts  of liquid  caustic soda  to  quench
                                       125

-------
 the   product   gas,   neutralize   acidic
 products   and  trap  partlculates.    Salt
 water  and  partlculates  are  pumped  from
 the scrubber  sump and  sampled prior  to
 approved   discharge.    An  induction  fan
 continuously draws on the  reaction  cham-
 ber and  scrubbing system  to maintain  a
 slightly   negative  pressure  in  the  sys-
 tem.   Producer gas from the induction fan
 passes through a  three-way valve.   During
 normal  operations, this  gas  flows  to the
 flare  stack  where  it  is  electrically
 ignited.   Since this  gas is mainly hydro-
 gen and  carbon monoxide, it  burns  with  a
 clean  flame at a  temperature of 2000-2300
 Kelvin.    This flare  stack  serves   as  an
 air pollution  control  device to  prevent
 the release of fuel  gas to  the  environ-
 ment.   In  the evant  of  a  power  failure,
 the three-way  valve  directs the  product
tgas_ through an activated carbon filter to
'block' the  potential   release  of  un-
 destroyed toxic material to  the  environ-
 ment.  Provision  is made for  sampling the
 product gas prior to  the flare.
       During  normal  operation,  the plasma
 pyrolysis   unit   requires   approximately
 three  minutes  of  warmup time  prior  to com-
 mencing   injection  of  toxic  or  hazardous
 wastes.     During  warmup,  the  system  is
 Injected   with a  flushing fluid  such  as
 ethanol to maintain a  reducing  atmosphere
 inside the  atomizatlon  zone.   This  unit
 also requires  approximately twenty  minutes
 to  cool key components  in order to  conduct
 post  run  maintenance.   These short  cycle
 times  make  this  unit basically  an  on/off
 system  which  can  respond  rapidly   to
 adverse  conditions.   The  system  hardware
 is  arranged to provide  failsafe  operation
 even  in  the event  of total  power  failure.
 Normal  scheduled  maintenance of  pre-engi-
 neered components  such  as  fans,  motors,
 valves.,  meters  and  sensors  should  safe-
 guard   against   unforeseen   malfunctions.
 Thus,   consideration   for  'failsafe   is
 limited to monitoring essential  conditions
 and providing protection  in  the event  of
 an  adverse situation.
       A  HACSYM  350  process control  com-
 puter  is  used to  control  the total  opera-
 tion of the system.  The computer  provides
 operator  warnings  as well  as taking posi-
 tive  action in the occurrence of  adverse
 conditions.   Under total  loss of  power,
 the arrangement  of  hardware  takes  over
 this latter function.   Waste  feed  shutdown
 signals  are triggered  by the computer  if
 it  senses  any  loss  of  the plasma  arc,
 induction  system  and  cooling systems  or
 notes the presence  of unacceptable  pro-
 ducts in  the  stack  product  gas.    If  the
 arc is lost during  toxic  waste  injection,
 the response  time of  the  shut  off  valve
 may allow  as  little  as five  micrograms of
 undestroyed  waste  to  pass   through  the
 pyrolysis  system   to  be  trapped  on  the
 activated  carbon   filter.     Unacceptable
 products  directed  to the carbon  filter  may
 include  trace levels of principle  organic
 hazardous   constituents   or    excessive
 release of hydrogen  chloride.
       In addition to providing  waste feed
 shutdown   signals  and  taking appropriate
 action, the computer  system,  through  total
 system monitoring,  will  also  provide  oper-
 ator  warning  signals  which,  if  not  reme-
 died  in  a  preset  time  period, will  auto-
 matically  shut  the  system  down.    Such
 warning  signals  stem  from  monitoring  all
 temperatures,   pressures,   flows,    fluid
 reserves   and   standard  operating   para-
 meters.   Failsafe  criteria  are instituted
 to  safeguard the environment.  Obviously,
 the degree  to which the  environment  is
 protected  will   bear  directly  on  the
 accuracy  of monitoring  activities.
       All  hardware  is  designed  to  be
 located   within  a  forty-five foot  long
 moving van type trailer.   The hitch end
 houses    the   six-pulse,   water    cooled
 thyristor unit which  is fed  with primary
 480 volt,  three  phase power.   The  rear
 section  of  the  trailer houses  the  skid
 mounted  hardware  and  access  is  made for
 the connection  of   domestic  water  and
 sanitary   sewer  services.     The  central
 area of the trailer contains  a laboratory
 1n  which  the  control  and  monitoring
 equipment  are located.    In   the   mobile
 unit, trace  analysis  will be  performed  by
 a Hewlett-Packard  5792A gas chromotograph
 coupled to a  Hewlett-Packard 5970A mass
 selective  detector  (MSD).   The MSD  has a
detection  level  of one nanogram while  in
the spectral  scan  mode  and  10 picograms
while  in  the  selected  ion mode.   It  is
proposed  at this time to monitor the pro-
duct gas  before  it is flared.  Detection
of  trace  chemicals 1s  assisted  by  their
being  at  their highest  molar concentra-
tion at that  point,  where  they  have yet
to  be  altered through  combustion  in the
f1 are.
       Bulk gas  analysis  of the  scrubbed
product gas  will  also be  performed at  a
                                           126

-------
sampling  point  downstream of  the induc-
tion  fan.    A  Hewlett-Packard  5880A gas
chromatograph  complete  with  both  flame
ionization   and   thermal   conductivity
detectors   will    provide   analysis   of
hydrogen,   water,   nitrogen,   methane,
carbon    monoxide,    carbon    dioxide,
ethylene,   ethane,  'acetylene,  propane,
propylene,    1-butene    and    hydrogen
chloride.   A detection of  200 parts per
million  (ppm) of hydrogen chloride in the
producer  gas will  signal  a  ninety-nine
percent  removal   efficiency of hydrogen
chloride  in  the   scrubber  during  the
destruction    of    wastes     containing
thirty-five percent chlorine by weight.
       It  is  proposed that  this  unit be
demonstrated in  Canada during   the summer
of  1984.   A test  program  is presently
under review by  the  Canadian Federal and
Provincial  agencies.   Subsequent testing
within New  York  State  is contingent  upon
Federal  and State Agency  permits.

 PROBLEMS ENCOUNTERED

        During this  project, problem areas
 have  been  encountered  involving  design
 changes  which  stemmed  from  a continuous
 indepth-  review 'process.    The  as-built
 schematic  for the mobile unit is signifi-
 cantly altered from the proposal  original-
 ly  put  forward  in the  contract.   Many of
 the changes resulted from the  selection of
 alternative  hardware.   - Others  were  re-
 quired  to  improve  the  failsafe  perform-
 ance.    Examples  of  additional  equipment
 added  or equipment modified   are  the  de-
 ionized  water  system,  activated  carbon
 filter,  air dryer,  scrubber  water tank,
 scrubber water  pump,  plasma field current
 supply,  redundancy  in instrumentation and
 sensors  as well  as  a  large  selection of
 valves,  piping   and  wiring to accomodate
 the requested modifications.
        Modifications  to the   Westinghouse
 plasma  arc  system were  necessary to  ensure
 that  it would  provide  a  similar  process
 performance  at  an  equivalent  plasma effi-
 ciency.   Several  weeks  of design  discus-
 sion  and  testing  on  a  Marc   II prototype
 device  and its, modified version were con-
 ducted  to  verify the  final design.  This
design was finalized in February, 1984 and
the  unit  is  scheduled  for  delivery  in
May.  The prototype unit  achieved destruc-
tion  efficiencies  of  99.9999  percent for
askarel  solution  fed   at  a  rate  of 3.6
kilograms  per minute   on  the  laboratory
reactor  at the  Royal  Military  College in
Kingston,  Ontario.  Similarly,  parametric
data  for the modified  device show a pre-
dictable performance  envelop  for  a  rated
performance   between   200-500   kilowatts.
Plasma enthalpies .in  excess  of  ten  kilo-
watt-hours per  normalized cubic  meter were
also  achieved.
        Present   guidelines  call   for  moni-
toring air emissions  from  the unit.   How-
ever, it  has  been shown  that  the  parti-
tioning  of toxic compounds passing through
the  scrubber  leaves   better  than  ninety
percent   of  them  in  the  scrubber  water.
Calibration  of  the laboratory scale scrub-
bing   unit  shows that  better than ninety-
six  percent  of  low  volatile  non-soluble
 (LVNS) compounds are trapped by the  scrub-
ber fluid.   During typical  operation  the
producer gas and  scrubber fluid flowrates
 are  fifteen  cubic meters  per  minute  and
forty   liters   per  minute  respectively.
Calculation will  show that the  LVNS oro-
qucts  are more than  69,000 times  more
 concentrated in the  scrubber fluid  than
 in the   producer  gas  prior to  flaring.
 These  calculations  have  been  verified
 with   parafin oil  which is characteristic
 of   most  LVNS  products   of   interest
 (dioxin, etc.).  Thus the  question of how
 to measure  the  performance  of  this new
 technology  has  yet to be resolved.   It
 may be more practical to exercise control
 over  the performance of  the  system  by
 monitoring  for  toxic  effluents  in the
 scrubber  water  since  it  functions  on   a
 single pass basis.

 RESULTS

        The fabrication of  the mobile unit
 should  be completed  in May 1984.  Demon-
 stration testing  of  the unit is forecast
 to start in June  1984 and will continue
 until all regulatory  and monitoring con-
 cerns have been addressed-.  Guidelines to
 assess  this  technology and the  protocols
                                           127

-------
for  this assessment  are  currently  under
development  and  should  be  completed  by
June.
ACKNOWLEDGEMENTS

       Funding for  this  project has  been
made  possible from the  New  York  State
Department of  Environmental  Conservation
and  the  Environmental  Protection Agency
Industrial Environmental Research  Labora-
tory.   Special  recognition  is  given to
Mr.  John  Grimaud  '(NASA)  for  valuable
advice in the course of this effort.
REFERENCES

(1)  Barton, Thomas  G.,  Unpublished test
     data  from the  Department  of Civil
     Engineering, Royal  Military  College
     of Canada, February 1984.

(2)  Gordon,  Sandford,  Computer  Program
     for Calculation of  Complex Chemical
     Equilibrium  Compositions,  NASA N78-
     17724, 1971.

(3)  Lee,  Chun  Cheng,  A  Comparison  of
     Innovative Technology for  the Des-
     truction   of   Hazardous    Wastes.
     USEPA/IERL  Cincinnati,   Ohio  45268,
     1984.
                                           128

-------
                     METAL VALUE RECOVERY FROM METAL HYDROXIDE SLUDGES

                                 L.  G.  Twidwell,  Professor
                              A. K.  Mehta, Adjunct Professor
                               G.  Hughes, Research Associate
                 Department of Metallurgy and Mineral  Processing Engineering
                      Montana College of Mineral  Science and Technology
                                       Butte, MT  59701


                                         ABSTRACT

     Metal bearing hydroxide sludge material  is generated by the metal  finishing industry.
This product has traditionally been disposed of in hazardous landfill  sites.  Long-term
maintenance of such sites is required and metal values are unnecessarily lost.   The pre-
sent research effort was initiated to investigate the application of well  established hy-
drometallurgical unit operations to mixed metal bearing hydroxide sludge materials.  Test
equipment capable of treating 75-100 Ibs. per day of mixed metal sludge has been assembled.
The test system has been used on a variety of hydroxide sludge materials to demonstrate
successful dissolution and selective and effective metal value separation.
INTRODUCTION

     In recent years increased emphasis has
been placed on preventing the introduction
of heavy metal containing industrial waste-
waters in publicly owned treatment works
and the environment.  Legislation has
established regulatory authority for con-
trolling the discharge of heavy metals into
the environment.  It also has mandated
resource recovery whenever economically
feasible.  Many treatment and control tech-
nologies have come into existence to remove
metals from these wastewaters, i.e., a
sludge, concentrate, or regenerate  form is
created and is, in most cases, disposed,of
in a landfill.  Metals are recoverable, but
are not recovered significantly because of
a lack of proven technologies.

     Process wastewaters from the metal fin-
ishing and electroplating industry  contain
cyanides and heavy metals.  These waste-
waters have a detrimental effect on the en-
vironment if discharged directly.   Such dis-
charges are regulated by Federal, State,
County, or City ordinances, and require in-
stallation of treatment .technology.  One of
the treatment technologies presently in use
is oxidation, neutralization and precipita-
tion, which destroys cyanide and removes
heavy metals  as a hydroxide sludge.  This
product has traditionally been disposed of
in hazardous landfill sites.

     Disposal of sludges in landfills has
certain inherent disadvantages:

1.  Perpetual maintenance of the disposal
    site is required,

2.  Dilution in metal content by mixing
    with other types of waste materials,
    and,

3.  Permanent loss of non-renewable metals.

If heavy metals are recovered from metal
finishing sludges, it will alleviate or re-
duce the disposal.problem and provide for
conservation of energy and metal resources.
The present study outlines a technical
methodology to treat metal bearing sludges
by hydrometal1urgical techniques.

     The treatment of hydroxide sludges for
metal value recovery will produce several
beneficial results, i.e.., economic benefits
from the metal values recovered will help
offset the cost of recovery/treatment; non-
renewable resource metals will be recycled
for use by society, and there will be
                                            129

-------
 significantly less hazardous material to be
 disposed of in landfills.

 PURPOSE

      The purpose of the present study was
 to investigate at an advanced laboratory
 scale the potential for application of well
 established hydrometallurgical  techniques
 to a mixed metal sludge.  The design, de-
 velopment, fabrication, acquisition, and
 assembly of a treatment system has been
 conducted at the Montana Tech Foundation
 Mineral  Research Center in Butte,  MT.

      The objectives of the study included:

      Develop a flowsheet to separate and
       recover metal values from metal fin-
       ishing hydroxide sludge materials,

      Develop a test assembly of unit oper-
       ations to accomplish the  separation
       of metal  values on a scale of 75-100
       Ibs.  of sludge per day, and

      Verify that the large scale  unit
       operations accomplish the  separation
       of metal  recovery efficiency for each
       unit operation; delineate  process and
       materials  handling problems  when
       treating  complex mixed metal  sludge
       materials.

APPROACH

     A laboratory test program has  been
conducted to  support the  development of an
appropriate flowsheet made  up of unit oper-
ations designed  to accomplish the  stated
objective of separation'and metal  value re-
covery from mixed metal  hydroxide  sludge
materials.  The  laboratory  test  program was
based  on  a  comprehensive  review  of current
literature; discussions with  consultative
experts  in  the field  of extractive  metal-
lurgy; and  previous experimental research
conducted at  Montana  Tech Foundation.   A
flowsheet was designed based  on  the  inputs
gained from the above sources and modified
as dictated by the  laboratory test  program.

     The  flowsheet  that resulted from the
Phase  I study was  based on  the treatment of
a mixed metal sludge material containing
significant concentrations  of metal  values:
copper, nickel, zinc, chromium, and  vari-
able concentrations of other elements that
were not  considered recoverable as mar-
ketable products  but which  required  removal
 in  order  not  to contaminate the metal value
 products, e.g., iron, aluminum, and calcium.
 The  developed flowsheet is not an unalter-
 able sequence of operations; alternatives
 do  exist  and  will be discussed in the body
 of  this report.  As is often the case, unit
 operations may be accomplished by several
 different technical approaches, e.g., the
 unit operation of chromium oxidation may be
 accomplished  by use of chemical oxidation
 reagents or by use of electrochemical cells.
 The choice is  usually based on efficiency
 and economical considerations.

     Another  important point concerning the
 development of the flowsheet is that only
 commercial process unit operations and com-
 monly used reagents were considered.  That
 is, new developments in the separation of
 metals from complex solutions have been re-
 ported but are not yet adopted commercially,
 e.g., iron removal  from a mixed metal  solu-
 tion by solvent extraction techniques.
 These new developments are reviewed in the
 final project report and the consequences
 of th.e adoption into the present flowsheet
 are discussed.

     The flowsheet developed for mixed
metal sludges is presented in Figure 1.
The treatment sequence consists of the
 following unit operations:

     Sulfuric acid dissolution of the
      metal  hydroxides.

     Selective precipitation  of iron  from
      the  solution  as  a  potassium or
      sodium jarosite  ((KFe3(OH)g(S04)2).

     Solid-liquid  separation  of the leach
      residue and precipitated jarosite
      solid.

     Selective extraction of  copper from
      the  leach, solution  (with  subsequent
      copper recovery  as electrodeposited
      copper or as  crystallized copper
      sulfate by solvent extraction.

     Extraction of zinc  (and  residual
      iron)  in preference  to chromium and
      nickel  from the  leach solution  (with
      subsequent zinc  recovery  by  crystal-
      lizing  as zinc sulfate)  by solvent
     extraction.

     Selective oxidation  of chromium ions
     to form dichromate.anions.
                                            130

-------
    Selective precipitation of chromium
     from solution (using lead sulfate)  as
     lead chromate.   The lead chromate can
     subsequently be redissolved to pro-
     duce a concentrated chromic acid sol-
     ution with the  regeneration of lead
     sulfate.

    Solid-liquid separation of the pre-
     cipitated lead  chromate.

    Selective precipitation of nickel
     ions from solution as nickel  sulfide
     or crystallization as Mi SO..

    Solid-liquid separation of the nickel
     sulfide or NiSO..

    Recycle of the  resulting purified
     leach solution  to  the original leach
     as make-up water.   The unrecycled
     final leach solution may be pre-
     treated by ion  exchange before
     discharge.

    Laboratory testwork supports the con-
cept that metal values  can be separated
and recovered effectively and efficiently
from complex mixed metal sludge materials.

RESULTS

    The experimental approach and philos-
ophy for the laboratory verification
studies include preliminary test of con-
cept by screening experiments; development
of a two-level factorial design matrix for
the experimental bench-scale studies; ex-
ecution of the studies  in the design matrix
to establish which variables are most im-
portant and what the relative effect of
each particular variable is on the measur-
ed result; and subsequent use of the de-
sign matrix  effects (by using the Box-
Wilson "steepest ascent" approach) to
optimize the selection  of experimental
variables for further larger scale
testwork.

Large Scale Testwork

    The objectives of the large scale test
program were:  to size  the unit operation
so that 45.4 kg (100 Ibs.) of sludge could
be treated per day;  to  test the unit oper-
ations to ascertain  if effective and effi-
cient metal value extraction and recovery
could be'achieved; and'to determine what
chemical and mechanical problems might be
associated with treating approximately 200
liters of leach solution per day.

    The test system consists of leach
vessels, settlers, a filter press, solvent
extraction mixer-settlers, chlorine or
electrochemical oxidizer, pH monitors and
controllers, precipitating vessels,
crystal!izers, and an ion exchange column.

    A summary of the experimental  program
and conclusions drawn from the testwork
are presented below.  The discussion
follows the sequence of operations pre-
sented in the flowsheet in Figure  1.

    The sulfuric acid leach operation is
effective in redissolving the metal values
(example starting sludge compositions are
presented in Table 1).  The dissolution is
rapid and without control problems.  The
leach is carried out in a single two hun-
dred seventy liter vessel.  The conditions
required are well characterized, and
rather mild, i.e., one-half hour,  40-50C,
sludge/liquid ratio of 0.8, acid content
to control pH in the range 0.5-1.5, and
agitation sufficient to suspend the par-
ti cul ate in the solution phase.  Example
data are presented in Table 2.

    The sludge dissolution is essentially
complete in less than one-half hour.
Therefore, the leach operation is  not the
controlling step in the overall treatment
sequence.  The leach unit operation is
capable of treating over a ton of sludge
per eight hour day.  The filterability of
the Teach residue product is difficult,  
but may be overcome with the use of filter
a'ids.  The filterability of a mixed leach
residue-jarosite product is rapid  and
effective.  Therefore, in most testwork
the jarosite precipitation process was
performed in-situ with the leach residue
solids.

    The iron removal unit operation is via
the precipitation of potassium jarosite.
The precipitation process requires ele-
vated temperatures and relatively  long
reaction times.  Two hundred liters of
leach solution slurry can be treated in
six-eight hours.  The jarosite process
allows iron to be removed from an  acid
solution.  The product is a crystalline
compound that has excellent settling and
filtering properties (see Table 3).  The
iron removal process has been demonstrated
on .high iron sludge materials, i.e.,
                                           131

-------
 * 15-20% iron  in  the  starting sludge solids.
 This  means  that for these particular
 sludges  a significant quantity of leach
 residue-jarosite  solids  is  formed, e.g.,
 11.6  kg  of  solids or  17.8 kg of wet mater-
 ial.  Therefore,  the  disposal of 17.8 kg
 would be required instead of 45.4 kg or
 approximately  forty percent of the original
 sludge weight.  A significant quantity of
 sludge material exists that has iron con-
 tents much  lower than the above values.
 The jarosite process  is  also effective for
 treating the low iron containing sludges,
 e.g., two-four percent iron.  The quantity
 of leach residue-jarosite solids produced
 from  such sludge material would be rather
 small, i.e., a two percent  iron bearing
 sludge would yield 5.6 kg of leach residue-
 jarosite solid.  This quantity of solids
 translates, at 65% solids,  to 8.6 kg of
 disposable material instead of 45.4 kg or
 approximately  one-fifth  the original sludge
 weight.   Jarosites are widely produced in
 the zinc industry.  They are deposited in
 lined storage  ponds.  It is difficult to
 state whether  their heavy metal content
 means that the jarosite  should be consider-
 ed a  hazardous material, but even if that
 is the case at least a significantly
 smaller  weight of material must be
 considered for disposal.

      High iron sludges do (low iron sludges
 do not)  present a problem for chromium re-
 covery.   Significant amounts of chromium
 are lost when the jarosite precipitation is
 performed.  It is believed that the loss
 can be minimized by maintaining conditions
 such  that chromium is not oxidized and the ,
 pH is maintained below 2.5.

     Mechanical control  of the system is no
 problem.  Chemical control must be exer-
 cised to ensure that the pH is maintained
 in the range 1.8-2.5 and that the iron is
 in the ferric form.   Solid-liquid separa-
 tion  is effectively accomplished by allow-
 ing the  leach residue-jarosite to settle;
 decanting the solution from the solids;
 and pumping the small  volume of remaining
 slurry to a filter press.

     The removal  of copper is accomplished
 by solvent extraction (SX) using LIX 622.
 The extraction of copper from zinc,  chrom-
 ium, nickel, and aluminum is selective and
effective (>96% extraction per contact
stage).   Copper contents of a few mg/liter
are achievable in  two stages of contact and
one stage of strip.   The pH  of the solution
 exiting  the jarosite precipitation unit
 operation need not be adjusted prior to
 copper extraction.

     The SX testrack is designed to treat
 up  to 200 liters of solution per day.  The
 design throughput is 500 cc/min. for each
 phase.   Most tests to date, have been per-
 formed at 250 cc/min.  Ten contact mixer-
 settler  units are available for copper SX.
 Therefore, this unit operation is not the
 slow step in treatment sequence, i.e.,
 three concurrent streams could be treated
 (each in three cells) at one time; at 250
 cc/min., 360 liters could be treated per
 day.

     Large scale testwork has been conduc-
 ted for  up to six hours.  Control of flow-
 rate and interface levels is easily achiev-
 ed and requires constant attention only
 during initial loading of the system.  Once
 the system interfaces have been established
 little operator attention is required.  A
 summary of experimental  testwork for large
 scale continuous testing is presented in
 Table 4.

     The removal of zinc is accomplished by
 solvent extraction using DEHPA.  The ex-
 traction of zinc from chromium and nickel
 is selective.  Ferric iron, aluminum, and
 calcium are partially coextracted with the
 zinc.

     The extraction of iron (only between
 0.2-0.6 gpl  present)  with zinc is desirable
 because it provides a way of removing re-
 sidual  iron from the  solution.   The iron
 once extracted into the  organic phase is
 not stripped by ^$04 but is stripped by
 HC1 acid (4-6N).  Zinc is stripped by H2S04
 (200 gpl).   Therefore, a means  of bleeding
 iron from the process stream is to load
 iron and zinc into the organic  phase, strip
 the zinc by contacting with H2S04 (200 gpl)
 followed by stripping the iron  from the
 organic by contacting with HC1  (4N).   Both
 strip solutions can be recycled until  the
metal  content is appropriate for recovery
of zinc as  zinc sulfate  monohydrate and for
 disposal  of iron as ferric chloride
solution.

     Calcium is coextracted with zinc but
poses  no problem because it precipitates as
 gypsum in the H2S04 strip circuit.   It can
be effectively filtered  continuously  from
the solution  during solution recovery.
                                           132

-------
     Aluminum is co-extracted with zinc and
partially stripped,in the t^SO^. strip cir-
cuit.  Its presence must be considered dur-
ing the crystallization of zinc sulfate.

     The zinc SX testrack is the same de-
sign as used for copper removal.  Ten SX
cells are available.  The removal of 5 gpl
zinc requires four stages of extraction,
three stages of zinc stripping, and one
stage of iron stripping.  Therefore, the
removal of zinc is the limiting step in
the present treatment process.  Two hundred
liters can be treated at a flow rate of 400.
cc/min. for each phase in an eight hour
period.  Some flexibility does, however,
exist by control of the extracting reagent
concentration in the organic and by chang-
ing the organic to aqueous ratio in the
system.

     Control of flowrate and interface
levels is easily achieved and does not re-
quire constant attention once the initial
loading and interface levels have been es-
tablished; i.e., operator attention is
minimal.  The system can be shut off and
restarted without difficulty.  Chemical
control of pH is required in zinc extrac-
tion to achieve effective zinc removal.
Solution pH control is exercised by adjust-
ing pH after the first two stages of con-
tact.  Temperatures in the range of 40-55C
are desirable for rapid phase disengagement.
A summary of experimental testwork for
large scale continuous testing is presented
in Table 5.

     Chromium removal is accomplished by
first oxidizing the chromium with chlorine
gas, then precipitating the dichromate ion
as lead chromate.  Oxidation may alterna-
tively be accomplished by use of electro-
chemical cells.  Both systems have been
shown to be effective in laboratory scale
test reactors.  Large scale oxidation test-
work using chlorine has been performed suc-
cessfully; Table 6.  Both chlorine oxida-
tion and electrochemical oxidation reactors
will be further investigated during Phase
II testwork.

     A recycle system for stripping the
oxidized chromium from the leach solution
has been operated successfully:  the solu-
tion is exposed to lead sulfate in an agi-
tated reactor; lead chromate precipitates;
the lead chromate product is crystalline
and dense and settles rapidly; the solution
essentially free of lead chromate solid is
pumped from the solids for further treat-
ment for nickel removal; the lead sulfate
solid is separated from the chromic acid
and is recycled to the lead chromate
precipitation reactor.

     Nickel can be removed by sulfide pre-
cipitation.  The reaction is rapid and near
quantitative.  The pH is maintained in the
range 4-5 so hydrogen sulfide is not re-
leased.  The solid product is readily fil-
terable.  Quantitative removal  of nickel
is not necessary because practically all
the final solution can be recycled to the
leach-jarosite precipitation unit operation.
Therefore, the addition of a deficiency of
sulfide (less than the stoichiometric re-
quirement for complete nickel removal) is
desirable so that all the added sulfide
ions are consumed.  Then when the solution
is recycled to the acid leach step,
hydrogen sulfide gas will*not be formed.
An alternative nickel recovery unit opera-
tion is nickel sulfate crystallization.

     The sequence of unit operations des-
cribed above results in appropriate metal
value recovery from mixed metal sludge
materials.  Not only are metal  values
recovered but there is a significant
decrease in the quantity of waste material
that must be treated for permanent disposal.

ACKNOWLEDGMENTS

     This research was supported by a
Cooperative Agreement (R 809 305 010) be-
tween EPA-IERL and Montana Tech Foundation.
The financial support is gratefully
acknowledged.
                                            133

-------
TABLE 1.  STARTING SLUDGE MATERIAL BLENDED SAMPLE REPRODUCIBILITY
Composition
Cu

Fe

Zn
Cr
Barrel 5 (used
2.41
2.41
2.48
2.4310.05

8.26
6.57
6.03
4.41
4.05
5.8612.36
11
11
11
.33
.88
.65
1 1 .6210.29

19
18
17
17
16

.05
.06
.37
.16
.91
17.7111.34
8
8
8
8.

6
8
9
9
6
.40
.45
.75
53+0.22
Barrel
.15
.61
.99
.83
.96
8.3112.16
1.36
1.35
1.35
1.35+0.01 4.
in Solid (%)
Ni
Al
Cd

Cu
Pb
in kettle test)
5.08
4.08
5.08
99+0.91
18 (used in large
8.52
7.10
6.24
4.46
4.13
1.91
2.23
2.28
2.45
2.47
6.0912.43 2.27+0.35
4.05
4.15
4.55
4.25+0.20 0.
scale test)
2.66
2.81
3..13
2.77
2.58
2.7910.34 0.
0.39
0.41
0.41
40+0.01

	
0.04
0.08
0.12
0.11
09+0.03
1
1
1
.08
.00
.10
1.06+0.04

0
0
0
0
0

.31
.45
.44
.64
.64
0.5010.19
0.09
0.10
0.09
0.09+0.01

0.08
0.08
0.07
0.05
0.11
0.08+0.03
                                  134

-------
         TABLE 2.   TYPICAL SULFURIC  ACID  LEACH OF MIXED METAL HYDROXIDE SLUDGE:
                   STANDARD CONDITIONS




Condition*
100
650
1,000
50,600
gm
gm
gm
gm
sludge
sludge
sludge
sludge


Fe
92
95
55
92
.0
.4
.8
.0

Cu
93.7
94.9
94.3
93.7

Metal
Zn
95
90
94
95
.9
.5
.2
.1
Extracted
Ni
95.9
97.8
85.0
95.9
(
Cr
96.5
96.7
96.7
96.5

Cd
93.0
100.0
97.0
93.0

Al
89.9
95.7
96.0
96.9
*Standard Conditions:   one-half hour  leach; ambient temperature; sludge/liquid ratio =
                       0.8;  acid content equivalent to weight of solids in sludge.
                                            135

-------
     TABLE 3.  JAROSITE FILTRATION RATES
Feed Solids
Solids Loading
 Rate (kg/m2/Hr.)
Cake Solids (%)
Present
 Study
 40-50
 25-55

 66-71
Japan*   Canada*
50-55     20-30
80-100    40-50
78-80
75-78
*Data supplied by
 Ingersoll-Rand
                        136

-------
 TABLE 4.   SUMMARY OF LARGE SCALE TESTS ON
           SOLVENT EXTRACTION OF COPPER WITH
           LIX 622
    Condition
    Copper extraction from
        leach solution
                  Per-
                  cent

Sequential Series
 One	

Raffinate from    98.9
 contact
 (40 lit.)

Sequential Series
 Two	.

Raffinate from
 contact
 (60 lit.)

Sequential Series
 Three	

Raffinate from    98.0
 contact
 (20 lit.)

Sequential Series
 Four	

Raffinate from    96.9
 contact
 (90 lit.)

Sequential Series
 Five	

Raffinate from    99.0
 contact
 (160 lit.)
                          Copper content
                            in solution
       Initial
        1.37
94.4    0.39
        2.32
        3.89
        3.05
Fi nal
(gpi)
0.017
0.022
0.047
0.120
0.030
                        137

-------
  TABLE 5.   SUMMARY OF LARGE SCALE TESTS ON
            SOLVENT EXTRACTION OF ZINC WITH
            D2EHPA
     Condition
 Sequential  Series
  Three	

 Raffinate from
  contact
  (20 lit.)

-Sequential  Series
  Four	

 Raffinate from
  contact
  (50 lit.)

 Raffinate from
  contact
  (90 lit.)

 Sequential  Series
  Five	

 Raffinate from
  contact
  (160 lit.)
    Zinc extraction from
	leach solution

       Zinc content
       in solution

Per-   Initial   Final
cent    (gpl)    (gpl)
97.4    5.70
97.8    5.89
98.8'   4.94
98.9    6.20
0.15
0.13
0.060
0.070
                        138

-------
 TABLE 6.   SUMMARY OF LARGE SCALE TESTS ON
           CHROMIUM PRECIPITATION
   Condition
      Chromium removed
	from solution

      Percent chromium
         content in
          solution

Per-  Initial    Final
Cent   (gpl)     (gpl)
Sequential Series
 Four	

Starting solution
 (10 liters)

Fi nal fi1trate
 after 30 min.
 exposure

Sequential Series
 Five	

Starting solution
 (42 liters)

Final filtrate
 after 30 min.
 exposure
99.5
99.7
       1.65
0.008
       2.34
0'.007
                       139

-------
 Recycle
Solution
Acid
SIudgo
            Slurry
                       Decant
               Settler
                    Wash .
                      11
Storage
                          FIItrate
                                                 COPPER SX
                                                Loaded Organic
           Aq
                                                Strip Organic
                            Jaroslte
                              Cake
            IRON REMOVAL
                                           Organic
                                                                     Pregnant Aqueous
                              Organic
                                                                                Cu
                                                                                EW
                                        Strip Acid
                                                Raffinate
                                          To Storage then to Zn SX
                                                            COPPER REMOVAL
         Figure 1.   Flowsheet  for Treatment of Mixed Metal  Hydroxide  Sludges.
                                                 140

-------
 Copper
Raffinate
       Storage
                                                        ZINC  SX
                                                      Loaded Organic
                                   Organic
        Strip
       Organic
      Organic
                              Aq
                           Organic
  Aq
Organic
Organic
   Strip Acid
                                                   Raffinate
                                  pH Adjust
                  (Bleed Stream
               to Crystallization)
                                                   To Storage
                                               then to Oxidation
                                         ZINC REMOVAL.
                                     Figure 1.   (Continued).
                                                  141

-------

Zinc






Na2S
C12 PbSO.,
4 Solution
Raff I note



1



Storaga





Decant Decant |
( lit Recycle
1 1 Makeup
1 1 * Water
PreclpI- Preclpl-
tat Ion tat ion
Oxidation "\^ ^/^ \^ ^S~








Wash | Wash

Restdual
Solids
rn
i i
1 	 Filter Filter
Press | Press "~|
to
Initial
Loach












PbCrO, Cake NIS Cake
1
IX
CHROMIUM REMOVAL
                                                              Smal I
                                                             VoIume
                                                            Discharge
                                               NICKEL. REMOVAL.
                   Figure  1.   (Continued).
                                142

-------
                  OPERATIONS AT THE USEPA COMBUSTION RESEARCH FACILITY
                                    Richard A. Carnes
                          U.S. EPA Combustion Research Facility
                                      Jefferson, AR
INTRODUCTION

     The safe disposal of this Nation's
industrial waste is the major goal estab-
lished by Congress in the Resource Conser-
vation and Recovery Act (RCRA) of 1976.
More specifically the USEPA has determined
that incineration of those wastes compat-
ible with thermal destruction is the
preferred disposal option.  Although
incineration is the method of choice,  its
proper implementation and continued safe
operation are required before it will be
accepted by local citizens or heavily
invested  in by  industrial generators or
waste disposal firms.  To these ends the
USEPA Combustion Research Facility (CRF)
has established  its mission and goals.
PURPOSE

     The CRF has a multi purpose mission
in that  it will assess the performance
capabilities of the rotary kiln/air pol-
lution control device presently  installed
and will follow these capability analyses
to the planned liquid injection unit and
its appurtenances when they are fabricated.
It will be through performance testing
that the CRF will assist in forming the
framework of the USEPA policies and
regulations regarding thermal destruction
of  industrial wastes as the disposal route
of choice.
     The CRF will provide a physical plant
for the Agency to test the selection of
appropriate Principal Organic Hazardous
Constituents (POHCs) for specification in
permits and it will directly support the
identification of those critical and
perhaps lesser known operating variables
that contribute to the production of
Products of Incomplete Combustion (PICs).
The CRF has as an operating purpose the
integration of data generated at both the
laboratory and bench scale so as to
provide the Agency with a potential
"incinerability" matrix and  identify those
waste components associated with thermal
environments which should be avoided to
prevent the formation of toxic products.

     The daily operations at the CRF will
provide the Agency and the technical com-
munity as a whole a better understanding
of the complex thermal processes occurring
in an  incinerator and the capability to
describe the thermal process so as to be
able to characterize and assess performance
in full-scale systems and to extrapolate
performance data from one waste to another
or from one technology design scale to
another, or even another design itself.

     Finally, the CRF will serve as the
focal point for the USEPA for definition
of compliance monitoring and will provide
valuable insights on what must be moni-
tored so proper assurances are given that
                                            143

-------
 the  incinerator is performing within pre-
 selected ranges.   The CRF will provide
 data on sub-system evaluation leading to a
 total  system Destruction and Removal
 Efficiency (ORE).   In this way the CRF
 will  make an invaluable contribution to
 the  national  permitting process,  the
         enforcement  process,  and  to citizen
         acceptance of incineration  as  an  accept-
         able  neighbor.

              As  can  be  seen  in  Table 1, the
         objectives of the  research  at  the CRF  are
         varied.
             TABLE 1.   RESEARCH OBJECTIVES OF THE COMBUSTION RESEARCH FACILITY
                                                 Specific Objective
                              1         2
                           Process    Seal ing
                           Operating   and
                           Variable   Design
      Reliability
      & Control-
      abil ity
             4
        Destruction
        of Various
          Wastes
   5
Spec ial
 Test
 Burns
Components
   and
Subsystem
A - Determine  Destruction
    Mechanisms &  Kinetics,
    Identify PICs
    (Unregulated
    Substances)

B - Engineering Character-
    ization of System
    Operat ion

C - Effects of Waste Feed
    Constituents &
    Physical Properties &
    of Feed System
    Configuration

D - System Response to
    Abnormal Operational
    Modes

E - Permit Compliance
    Testing

F - Determine Destruction
    Efficiencies for
    Regulated Waste
    Materials &
    Substances
X


X
X


X
                                           144

-------
.PROBLEMS  ENCOUNTERED

      The  conduct of research  at  the  pilot
scale is  a most difficult  task.   Being
 involved  in  hazardous  waste disposal  at
this  scale compounds the difficulty.   In
the course of  events that  has transpired
at the  CRF  in  getting  operations ready
to conduct hazardous waste research  there
are few problems that  have not been
encountered.

      Some problems that caused the  largest
cost  and  time  increments are

      (1)  Identification of  a properly
          tested and approved waste
          container.

      (2)  Design and procurement of  air
          pollution  control  equipment
          standard to  the  field.

      (3)   Installation and operation of
          the  on-line  combustion gas
           analysis system.

      (4)  Preparation  of documents  for
          obtaining  the necessary EPA
          permits  under RCRA.
 FACILITY. STATUS

      At the present time, the technology
 that is available at the CRF is a pilot
 scale rotary kiln with afterburner and an
 air pollution control system consisting of
 a variable throat venturi wetted elbow and
 packed tower scrubber.

      The present configuration of the kiln
 is such that the feed waste is countercur-
 rent to the primary burner.  The burner is
 fitted with an in-line Validyne Engineer-
 ing Model 11-1-1000 propane meter.  The
 primary combustion air is monitored by an
 in-line pitot. .The kiln temperature  is
 controlled using a shielded thermocouple
 at the feed face.  Additional monitoring
 equipment on the kiln includes the combus
tion gases for both composition and flow
rate.  Combustion gases are monitored for
CO, C02 and 02 using an on-line Bendix
Combustion Gas Analyzer.  The organic
components can be determined by extractive
sampling using either liquid techniques
(EPA Method 5) or solid adsorber sampling
systems, such as the Volatile Organic
Sampling Train (VOST).

     The afterburner temperature is
controlled using a shielded thermocouple
located at the center point of the chamber.
Additional monitoring facilities consist
of surface temperature, exit gas tempera-
ture, 'and provisions for determining
combustion gas composition and flow rate
(essentially the same as are available for
the kiln gases).

     An Anderson 2000 air pollution
control system is fitted to the exit of
the afterburner chamber by a refractory
line elbow.  Here the high temperature
gases are contacted with a variable throat
wetted venturi operating in the range of
30  inches of water gravity.  The gases are
cooled and large particulate (if any)  is
removed.  The gases are then passed
through a packed spray column and  a two
stage mist eliminator.  Thus far,  results
show the system performing very well
regarding halogen acid neutralization  and
particulate removal.

     The Bendix Combustion Gas Analysis
system  is utilized to continuously monitor
CO, C02, and 02  in the kiln transfer
duct, the afterburner and the stack.
Temperature  is monitored at more than
35  locations throughout the system.  These
range from routine maintenance tempera-
tures to those in the afterburner  and  in
the stack.   Velocity measurements  are
taken at the head of the kiln and  the  head
'of  the  afterburner.  Combustion gas flow
measurements ate taken  in the kiln exit,
afterburner  exit and 'stack as  is moisture
content and  the chlorine and HC1 content
 in  each of the three places.
                                            145

-------
      The CRF has in place an approved
 quality assurance (QA)  program  that  is in
 complete agreement with all  aspects  of the
 required program units.  The quality
 control  (QC) aspects of operation  also
 conform  to  EPA requirements  and  are  an
 integral part of all analytical  operations.
 A detailed  health and safety program has
 been  established at the CRF  and  is part of
 daily operations.  The  general  require-
 ments that  have been placed  on  this  pro-
 gram  are:   (a)  protection of all personnel
 who will be  responsible for  the  operation
 of the facility, (b) protection  of the
                                         surrounding  area,  including the community
                                         and  environment,  and  (c)  protection  of the
                                         facility and  its  equipment.
                                         RESULTS

                                              Thus  far  experiments  have been  con-
                                         ducted  and samples  collected  during  the
                                         combustion of  propane  (background  blank),
                                         toluene (feedstock  for chlorobenzenes),
                                         hexachlorobenzene (HCB),  and  trichloroben-
                                         zene  (TCB).  Table  2 presents the  experi-
                                         mental  conditions for  the  experiments.
                      TABLE  2.   HCB AND  TCB  EXPERIMENTAL  TEST SERIES
EXP.
NO.
1.
2.
3.
4.
5.
POHC(l)
IDENTITY
Propane
Propane
Toluene
Propane
Toluene
Propane
Toluene
Propane
Toluene
TAB
1093
1093
982
871
871
(C)
879
871
871
760
760
FEED RATE SAMPLING ,.
(kg/hr) POSITIONS^'
-59 (a),(b)
2,43
2.43 (a)!(b)
~ 2.43 (a) lib)
2.43
REMARKS
Background
Background
Background
Background
Failure Mode-AB
burner shut off
6.
7.
8.
9.
Propane
Toluene
HCB

Propane
Toluene
HCB

Propane
Toluene
HCB

Prdpane
Toluene
1,2,4-TCB
1093
 982
 810
1093
871
760
760
871
                                        2.43
                                        0.10
   2.43
   0.10

^-46
   2.43
   0.10
                                        2.43
                                        0.10
(c)



(c)



(c)



(c)
                               2 min. every  15 min.

                               HCB fed as solution
                               in Toluene
                                                                    HCB fed as solution
                                                                    in Toluene
                                                                    HCB fed as solution
                                                                    in Toluene
                               1,2,4-TCB fed as
                               solution  in toluene
                                                                            (continued)
                                            146

-------
TABLE 2 (continued)
EXP.
NO.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
(1)
(2)
POHC(1)
IDENTITY
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
Toluene
1,2,4-TCB
Propane
1,2,4-TCB
Propane
1,2,4-TCB
Propane
1,2,4-TCB
Chemical(s)
Position (a)
TAB
CO
1093
1093
1093
1093
982
871
871
1121
1121
1121
fed to
= kiln

-------
Selected  DRE's for the  experiments  using
the chlorobenzenes are  presented  in Table  3.
                  TABLE  3.   RESULTS  OF  SELECTED  CHLOROBENZENE  EXPERIMENTS
Exp.
NO:
8
7
6
9
10
11
17
18
19
TAB
CO
810
982
1093
1093
1093
1093
1121
1121
1121
rko
760
760
871
871
871
871
788-760
732-704
677-649
Feed Rate
(kg/hr)
0.1
0.1
0.1
0.1
0.25
0.5
1.0
*
*
*
Emission Rate
(qm/hr)
2. 23xlO-3
2.03x10-3
4.41xlO-3
3.47x10-3'
1.66xlO-3
0.50x10-3
0.31x10-3
0.25x10-3
0.67x10-3
0.86x10-3
1.90x10-3
2.74x10-3
TOTAL EMISSION
104.6 MQ
342.3 Mg
0.305 grams
ORE
%
99.998
99.998
99.996
99.997
99.998
99.999
99.9997
99.9999
99.9998
99.9998
99.9998
99.9997
DE%+
99.9991
99.992
97.2
+ DE = Destruction Efficiency as opposed to ORE.
* The 1,2,4-trichlorobenzene was fed at a rate of 66.7 gm per minute for an elapsed time
  of 30 seconds.  Thus a total of 33.35 gms was fed  in each of the experiments.  Assuming
  that all feed was volatilized and removed from the kiln during this time, and taking the
  appropriate (measured) flow rate for each of the conditions, we compute DE as follows:


                   DE = 1 - -
                            C0

                       33 35
          where C0 =	'-	and would be the concentration in the kiln flue
                     kiln flow
gas If DE=0.

Kfln flow for experiment 17=5.70 (DSCMM), 18=6.28 (DSCMM), 19=6.02 (DSCMM)
                                            148

-------
     In addition to the aforementioned
sequence of experiments, the ability to
accurately determine flow rates and spe-
cific concentrations of combustion pro-
ducts allows an estimate of the heat
release from the stoichiometric equations
for the oxidation of propane:
                       Researchers at the CRF haye'compared
                  the total heat release against four (4)
                  methods of measurement.  The results of
                  these measurements are presented in
                  Table 4.
                    TABLE 4.  COMPARISON OF HEAT RELEASE (MMBtu/hr.)
Heat Release From
Stack Gas Analysis
Heat Release Based
On Inline Gas Meters
Heat Release Based on Combustion
    Gases and Surfaces Losses
C02
1.94
H20
1.87
Propane Flow
1.97
1.97
     The excellent agreement between the
various estimates of the heat release
lends credence to the various measurement
techniques used.
FUTURE OF CRF

     Presently ,the CRF is anticipating
receipt of a final Part B RCRA incinerator
permit to be issued by the State of
Arkansas under Phase II authority granted
by the USEPA.  Until such a permit is
received operations at the CRF shall be
limited to commercially available chemi-
cals and synthetic mixes.  Once permission
is received the CRF is geared to secure
several candidate waste streams for active
investigation.

     Research shall be directed toward a
better understanding of residence time.
It has often been said that the operation
of an  incinerator  is governed by the
three T's - time, temperature and _turbu-
lence.  The Time spent (its residence
time)  in an afterburner  is normally deter-
mined by the prescription, Equation (1).
                                                       '   (1)
                  where V is the chamber volume and Q the
                  throughput of combustion gases measured at
                  the exit temperature of the chamber.
                  There are several problems with this.  The
                  gas temperature within the chamber can be
                  expected to be significantly above that at
                  the exit and therefore to move the same
                  mass of gas through the unit area requires
                  a higher velocity.  Also, the turbulence
                  within the chamber must result in not all
                  elements of the combustion gas remaining
                  within the chamber for the same time
                  period.  For these reasons the CRF is
                  presently undertaking a detailed investi-
                  gation to define tr as some sort of
                  average time within the high temperature
                  chamber.  It was recognized that helium
                  dflution could be used for an actual
                  measurement of residence time in addition
                  to the above mentioned utility for high
                  temperature flow measurements.  Instead of
                  a steady input flow as required by the
                  former measurement, a measure of the time
                                            149

-------
of  arrival  of  hel inn  introduced  as  a sharp
pulse  will  then  allow a measurement of the
residence time within the  chamber.   Pre-
liminary results of the residence time
distribution measurements  have been very
exciting and  it  is anticipated that these
experiments will be completed  in the near
future.  Figure  1  is  a schematic of where
helium will be introduced  in the system
and where analysis will occur.
                    FIGURE  1
                            PORT
GO  < WATER TRAP *
 WPUTFUEL
 INPUT FUEL
                                  OUTPUT GAS
                             WPUT FEED

                              He
                              PORT
     The concept of hot zone sampling has
been investigated by others but not to the
extent as called for at the CRF.   At the
present time there is a stationary probe
located in the transfer line from  the kiln
to the afterburner and  in the duct from
the afterburner to the scrubber.   These
probes are in zones that are non-laminar
in flow so that mixing might be induced
and thus concentration uniformity  of
chemical species would occur.  There is no
attempt to sample iso-kinetically.  At the
present time samples are being collected
using the solid absorbents XAD-2 and Tenax
in which further sample manipulation
follows  in  the  laboratory.   Results  of
these efforts will be forthcoming  in
separate  CRF documents,  some  of  which will
be  submitted to referred journals.

     Other  areas  in which the CRF  is
conducting  investigations are:

     o  Chlorine  balance  in the  scrubber
        water

     o  Fixed vs_  Travers ing sampling  in
        the stack

     o  Flame out failure mode experiments

     o  Sub-system destruction/removal
        efficiency calculations

     o  Product(s) of incomplete combus-
        tion analysis and prediction of
        species

     o  Correlation of pilot  scale data to
        lab and bench scale results

     o  Evaluation of modified solid
        absorbent clean-up procedures.

     In the future several reports and
journal publications will transmit the
results and significance of these and
other studies at the CRF.

     In an attempt to assure that there
exists a high degree of coordination
between the several IERL incineration
research programs it has been determined
that each should study several of the
synthetic waste streams referred to as EPA
"Soups."  This approach has been taken to
obviate the variations that can be ex-
pected in actual  industrial  wastes and is
expected to yield data that will allow
direct intercomparison among the several
programs. This procedure, in addition to
generating  information of direct value
itself, should also assist in understand-
ing the means by which laboratory data
should be extrapolated to pilot and full
scale technology.  The protocol which is
herein presented is designed to extract
                                            150

-------
the maximum of information from a series
of incineration tests at the CRF on the
first of these soups.

     The experiments described below are
anticipated to yield data directly applic-
able to the following:

     o  A determination of the effects of
        afterburner temperature, residence
        time and excess air on the thermal
        stability of the compounds that
        make up the EPA soup #1

     o  The effect of kiln conditions on
        the OE of the afterburner

     o  The effect of kiln conditions on
        the production of PICs from the
        soup feed,

     o  The effect of feed rate of the
        soup on both ORE for the compo-
        nents of the soup and on the
        production and stability of PICs

     o  The effect of variations of RTD on
        the observed ORE for the after-
        burner

     o  Development of thermal stability
        data that will allow comparison to
        other technologies

     o  Investigation of scrubber holdup  .
        effects

     o  Obtain operational experience on
        the use of'the VOST system for
        sampling  and analysis of POHCs
        (and PICs) which are suitable for
        purge and trap analytical methods.

     The philosophy that has underlain the
CRF  operations from the beginning of the
program has been that all experiments
should be conducted from a technically
conservative viewpoint.  That is to say,
every effort should be expended  to design
experiments that would yield useful data
while at the same time conducting the
tests so as to have minimal impact on the
facility and the immediate environment.
This attitude will  be considerably im-
pacted when the HEPA/carbon filtration
system is  in place since use of the system
will always ensure that emissions are
within permissible-1imits.  The experi-
mental protocol suggested for the soup
experiments is presently dictated by this
conservative attitude which requires
beginning with low feed concentrations and
maximum thermal conditions within the
incinerator.  As it  is shown that no more
than permissible emissions are found under
these severe conditions the conservative
restrictions are gradually relaxed.  At
each step the emission levels determine
the advisability of proceeding to the next
less severe operations set.  This process
continues until either the entire set of
experimental conditions has been explored
or until a set of operational parameters
has been attained for which non-permitted
conditions are found.

     The proposed set of test conditions
will be carried out using the EPA soup as
indicated below:

     Soup components:

          carbon tetrachloride (CC14);
          trichloroethylene (02^013);
          monochlorobenzene (C5Hr;Cl); and
          Freon 113  (C2Cl3F3) in a
          toluene matrix.

It  is proposed that  several solutions be
made up of this combination as follows:
                                            151

-------
          Feed Mix BI  (in the proportions as follows)
            Ce^Cl        93.5 ml corresponding to 31.5 gm
            C2H3C13       68- ml corresponding to 58.0 gm
            CC14          62.7 ml corresponding to 92.2 Cl2
            C2C13F3       66- ml corresponding to 37.4 gm Cl2 + 30.4 gm
            C7H8        2500.0 ml
            TOTALS         2.79 liters containing 219.1 gm Cl2 + 30.4 gm

          Feed Mix 62  (in the proportions as follows)
                         467.5 ml corresponding to 157.5 gm
                         340.0 ml corresponding to 290.0 gm
            CC14         313.5 ml corresponding to 461.0 gm
                  F3     330.0 ml corresponding to 187.0 gm Cl2 + 152 gm F2
                        1349.0 ml
C7H8
TOTALS
                           2.80 liters containing 1095.5 gm Cl2 + 152 gm F2
     It  is noted that Mix BI has a
heating value of 114 MBtu/gal whereas Mix
83 has a heating value of approximately
62 MBtu/gal.
                                        The feed rates  to  be  used  for  both
                                   Mixes BI and  62 will  be 2.84 1/hr
                                   (0.75 gal/hr).
                                   The actual  proposed  experimental matrix  is
                                   shown in Table 5.
                        TABLE 5.  EPA SOUP - EXPERIMENTAL MATRIX
Experiment
A!
A2
AS
A4
AS
AS
A?
AS
Ag
AID
TV
N
1600
1400
1200
1200
1600
1400
1200
1200
1400
1400
&
2100
1800
1600
1400
2100
1800
1600
1400
2000
2000
Feed
Ql
D*|
Bl
BI
B2
B2
B2
B2
B2
B2
Sampl ing*
K
4
4
4
4
4
4
4
4
2
4
A
2
2
2
2
2
2
2
2
2
2
Stack
2
2
2
2
2
2
2
2
2
2.
Slowdown
1
1
1
1
1
1
1
1
1
1
*Integer refers to the number of samples to be taken.
                                            152

-------
     Experiments Ag and AIQ will be fail-
ure mode experiments.  In experiment Ag
the afterburner propane feed will be shut
off for 2 minutes out of every 15 minutes.
AIQ W1'H involve the same procedure for
the kiln burner.

     In the event that the high concentra-
tion feed stock, 62, shows compliance
under all conditions,  it will be appropri-
ate to conduct an additional series of
experiments using Mix 62 but at a feed
rate of 5.6 1/hr (,f.5 ga/hr) at conditions
s imilar to AS and AS-

     In all cases (sets of parameters  in
Table 1) the kiln residual pressure will
be adjusted to -0.05  inches of water,
gauge, "in order to  keep the excess air
approximately constant.
CONCLUSIONS

     The following  conclusions, may be
drawn from recent studies  on  the  incinera-
tion characteristics of  selected  chlor-
 inated benzenes  in  toluene solution:

     o  The  CRF  incinerator can produce
        the  required ORE of greater than
        99.99% for  the selected POHCs.

     o  The  procedures for hot zone sampl-
         ing  that have been implemented
        appear to yield  consistent results.

     o  Comparison  of absolute quantities
        of several  chlorinated benzenes  to
        soot weight in kiln samples shows
        ri correlation.

     o  The  behavior of  the afterburner  at
         low  injection rates is apparently
        dependent on the presence of PICs
        that are derived from the POHC.

     o  There is apparently a direct
        relationship between  DE/\  and the
         injection  rate  at  low injection
         rates.
       As expected, the existence of a
       residence time distribution  in the
       afterburner results  in the failure
       of first order kinetics to describe
       the destruction of hexachloroben-
       zene and 1,2,4-trichlorobenzene.

       The proposed residence time  dis-
       tribution and the method of  its
       measurement apply quite well to
       the CRF  incinerator.

       The use  of helium dilution for the
       measurement of combustion gas flow
       yields consistent results.

       There apparently exists a strong
        interrelationship between 1,2,4-TCB
       pentachlorobenzene  and HCB,  a
       relationship  suggesting that any
       two will be PICs derived from the
       third.

       The proposed  scrubber blowdown
       model appears  to be  directly
       applicable to  the CRF incinerator.

       The poor turndown of the  Iron
       Fireman  burners  1imits the  feed
       rate  that can  be used,  indicating
       the need to replace these burners.

       The castable  refractory  has  stood
        up remarkably well  even  under the
        severe  thermal  stress of  intermit-
       tent  operation.   Engineering
        tests conducted  without  POHC feed
        always  show  the  POHC and  PIC
        from  previous analytical  tests
        which were  conducted with  POHC
        feed.
RECOMMENDATIONS
        The successful  demonstration of 3
        method of residence time distribu-
        tion measurement in the CRF after-
        burner opens the door  to an on-
        1ine measurement of retention time
        which could then become a monitor-
        ing and control method.
                                             153

-------
The successful  application of
simple models to such aspects of
the CRF incinerator as the after-
burner retention time and the
blowdown discharge suggest that
modeling of other aspects of the
system can lead to enhanced under-
standing of the incineration
process and to  better control
thereof.
Attempts should be made to  allow
the measurement of mixing  in the
afterburner and the effects of
such mixing on the temperature
distribution as well as on  the
retention time distribution and
thereby on the performance  of the
incinerator.

The fact that a new feed face
configuration has been designed
with the capability of pumpable
fluids, solids and containerized
solids feed indicates that  by the
time the RCRA permit has been
issued, the CRF will be in  posi-
tion to study complex or problem
wastes.
                                      ACKNOWLEDGEMENTS

                                            The  author  wishes  to  acknowledge the
                                      on-site contractor,  Versar,  Inc.,  for
                                      their diligent performance under very
                                      trying conditions;  in particular,
                                      Dr. Frank  C.  Whitmore,  Program Manager;
                                      Mr. Robert Ross,  Senior Chemist; Mr.  Tony
                                      Reynolds,  Senior  Engineer; and Dr.  Fred
                                      Fowler, Health and  Safety  Officer.
                                  154

-------
            SAMPLING METHODS FOR EMISSIONS FROM HAZARDOUS WASTE COMBUSTION

                                   Larry D. Johnson
                     Industrial  Environmental  Research Laboratory
                          US Environmental Protection Agency
                           Research Triangle Park,  NC 27711

                                       ABSTRACT
     Incineration as a means of hazardous waste disposal has received increasing
attention during the last several years.  Since it is necessary to characterize
the stack emissions as well as the waste in order to determine how well the unit
is operating, research into adequate methods of sampling and analysis has been
necessary.  This paper presents an overview of the sampling methods which are
resulting from current and previous research programs and which are being recommended
to EPA regulatory programs to EPA engineering research and development projects,
and to interested parties in the industrial community.  The methods discussed
are generally applicable to incineration and also to processes closely related
to incineration, such as co-firing of waste in industrial boilers, and burning
of contaminated heating oil.

     Although methods for inorganic hazardous compounds are very briefly outlined, the
primary emphasis of the paper is on organic compounds which are likely to be chosen
as principal organic hazardous constituents (POHCs) for a trial burn.  Methods
receiving major attention in this paper include:  the Modified Method Five Train  (MM5)
which includes an XAD-2 sorbent module, the Source Assessment Sampling System (SASS),
the recently developed Volatile Organic Sampling Train  (VOST), and assorted containers
such as glass bulbs and plastic bags.
INTRODUCTION

     Incineration as a means of hazardous
waste disposal, especially for organic
materials, has received increasing atten-
tion during the last several years.
Although there are many measurements that
may be made in order to characterize the
operation of an incinerator, virtually all
of them require sampling and analysis of
the flue gas a's well as the waste feed
material.  Although sampling and analysis
of incinerator emissions have been accom-
plished in the past, a number of advances
in the state of the art have been made
within the last few years.  The range of
chemical compounds likely to be of interest
has also widened considerably.

     A program to'provide and document
adequate methods for characterization of
hazardous waste incinerators is being
carried out by EPA's Office of Research
and Development (ORD).  A reference
document(l) for use by ORD incineration
engineering R&D programs, EPA's Office
of Solid Waste and Regional Offices,
and the regulated community has been
released, following peer and policy
review.  The methods in Reference 1 are
state of the art, and are scientifically
sound.  Extensive programs of careful
and thorough validation, which are under-
way at present, will greatly benefit the
users of these methods.  In the meantime,
the methods to be described are best used
by those with considerable experience in
stack testing and chemical analysis: as
much expert advice as possible should be
sought when preparing or carrying out an
incineration characterization project.

 This paper presents an overview of
the methods proposed in Reference 1 with
special emphasis on stack sampling of
organic compounds likely to be chosen as
principal organic hazardous constituents
(POHCs).
                                            155

-------
 Stack Sampling Methods

    Although numerous sampling methods
 might be applied to a given incinerator,
 this discussion will concentrate on four
 approaches that are being utilized to the
 greatest extent, not only because of their
 generally high performance, but because of
 their flexibility and usefulness for a
 wide range of materials.

    The Source Assessment  Sampling System
 (SASS),  Figure 1, was developed for envi-
 ronmental assessment programs and is still
 the train of choice when  large amounts of
 samples  are necessary for extended chem-
 ical analysis or biotesting.   The SASS
 includes:  cyclones for particle sizing, a
 glass or quartz fiber filter  for fine
 particle collection, a sorbent module for
 collection of semivolatile organics,  and
 impingers for collection  of volatile
 metals.   The SASS operates at 110 to 140
 Lpm (4 to 5 cfm)  and is usually operated
 long enough to collect 30 m3  of flue gas.
 It  is possible,  but not convenient,  to
 traverse a stack with an  SASS,  so it is
 usually  operated at a single  point in the
 stack under pseudo-isokinetic conditions.
 For  further details as to the meaning and
 potential  ramifications of this mode  of
 sampling,  see Reference 2.

   Potential corrosion of stainless  steel
 in the sorbent module  of  the  SASS has
 prompted development of glass sorbent
 modules  which appear to perform adequately.
 One  of these  is  described in  Reference 3.

   The Modified  Method Five Train (MM5) is
 conceptually very  similar  to  the  SASS but
 operates at a  lower  flow  rate,  usually 14
 to 28 Lpm  (0.5 to  1 cfm).  The MM5, shown
 in Figure  2, does  not  include particle siz-
 ing  cyclones,  and  is usually  constructed of
 glass rather  than  stainless steel.  The
MM5  results  from a  very simple modification
of any of  the commercial  sampling  trains
 available which  conform to the requirements
of EPA Method  5.  The  sorbent module  with
cooling  capability  is  simply  inserted
between  the  filter and  the first  impinger.
The  sorbent module must be positioned
vertically  so the gas  and any condensed
liquids  flow downward  through it.

   The sorbent of choice for most sampling
jobs for both the SASS and MM5 is XAD-2.
For  further discussion of the reasons be-
hind this choice, as well as sorbent
 module placement in the train see
 Reference 2.

      Because  of its more convenient
 size  and ready availability,
 the MM5 is usually chosen over
 the SASS for  incinerator  sampling
 unless larger samples  are needed
 for lower detection limits or
 extensive  analysis requirements.

      Either the SASS or MM5
 provides collection ability for
 particulate  material,  acid
 gases such as HC1,  gaseous
 metal compounds (if appropriate
 collection liquids are  chosen),
 medium boiling organics (b.p.
 100C to 300C),  and high boiling
 organics (b.p.  greater  than
 300C).  Organics  with boiling
 points between 100C and 120C
 require  individual  attention
 during the sampling planning
 stage and may require decreased
 sampling times to prevent volumetric
 breakthrough.   Volumetric break-
 through  is related  to the migration
 rate  of  sorbed material through
 unsaturated sorbent beds.   For
 further  discussion  of this
 important concept,  see  Reference
 2.

      Considerable field experience
 has been gained with the SASS
 and MM5  trains  even though
 formal validation procedures
 have  not been  completed.   Much
 of  the confidence in these
 trains'  ability to  collect
 organics rests  on knowledge of
 the behavior of sorbents  with
 respect  to qollection and  recovery.
 Reference  2 discusses this
 point  in more detail.   A  particularly
 effective  demonstration of
 these trains'  performance on a
 dioxin injected into combustion
 gas is given in Reference  4.
 Similar  research  with other
 compounds  is underway,   and a
 validation project  has  begun.

     The MM5 and  SASS are not
 generally quantitative  collection
 trains for organics with b.p.
 less than  100C.  For these low
 boiling compounds,  the  recommended
methods are plastic sampling
                                           156

-------
bags, glass sampling bulbs, or the newly
developed volatile organic sampling train
(VOST).  The ambient air at many inciner-
ator sites exhibits relatively high levels
of volatile organics.  This greatly
increases the difficulty of obtaining an
uncontaminated sample of low concentration
and requires a great deal of care as well
as adequate blanks.  All of the above
methods have shortcomings, but they are
much less severe than the .shortcomings and
limitations of alternate approaches  <2) .

  The VOST is shown in Figure 3.  This
train and the analysis approach applied to
the samples resulting from its use were
developed in order to address stack con-
centrations as low as 0.1 ng/L.  A sorbent
tube containing 1.6 g of Tenax is position-
ed early in the train in order to remove
organics from the gas and liquid stream as
soon  as possible.  A second sorbent  tube
containing 1 g of Tenax and 1 g of charcoal
follows the condensate collector in  order
to act as a backup  in case of breakthrough.
The charcoal provides added stopping  power
for  the very low boilers  such as vinyl
chloride.  The train was  designed to use
six  pairs of sorbent tubes sequentially,
each  operating for  20 minutes at  1  L/min.
The  ability to concentrate the organics
from all  six sets  of tubes onto one  ana-
lytical  tube and  subsequently heat  desorb
into a GC  or GC/MS makes  it possible to
achieve  very low  levels  relative  to the
stack gas  concentration.   For higher stack
gas  concentrations,  it  appears  to be pos-
sible (perhaps even necessary)  to operate
the  VOST at lower  flow  rates  and  longer
 sampling times.   There  is no  fundamental
 reason why this  piece of equipment should
 not be useful  in a number of  operation
modes as long  as excessive volumes of gas
 are not pulled through  a single tube,
 since low boilers break through the sorbent
 after relatively low volumes.

   The development and use of the VOST has
 been discussed in several publications(5,6,
 7).   Reference 7 is particularly useful to
 users of the train.  Preliminary evidence
 that the train is effective is given in
 these references, and an EPA validation
 .project for the VOST is underway.

   Various types of plastic sampling bags
 have been used in the past, with very
 mixed results.  It is quite possible  to get
 good results with this approach, but  it  is
 essential that the sampling and storage
characteristics of the specific POHC,
relative to the specific types of bags to
be used, be well known.  For example,
organics such as alcohols usually exhibit
poor storage characteristics in bags (> .
It is also essential that field blanks be
included in the sampling strategy, since
all known methods for volatile organics
from incinerators are subject to potenti-
ally severe contamination in the field and
during transit.

     When sampling relatively high con-
centrations of volatiles, glass sampling
bulbs with secure seals may be the best
choice.  Although somewhat inconvenient
and lacking in sample concentrating ability,
the glass bulbs do show better sample stor-
age characteristics than plastic bags (8>.

     In addition to the methods discussed,
special procedures may be necessary for
certain pollutants.  Formaldehyde is an
example of compounds that require special
handling  C*).  Reference 1 gives guidance
in such cases.  A number of other less
satisfactory general sampling approaches
exist, but have not been recommended.   For
further discussion of  these techniques  and
their shortcomings, see Reference 2.

Summary

Sampling methods have  been discussed which
are generally  applicable to incineration
and to processes closely related  to  incin-
eration,  such  as cofiring of waste in
industrial boilers, and burning of contami-
nated heating  oil.  Although some of the
methods are relatively new and all require
a great deal of care and attention,  it  is
possible  to produce excellent results
through their  application.
                                             157

-------
REFERENCES

1. Harris, J.C., Larsen, D.J., Rechsteiner,
   C.E.- and Thrun, K.E., "Sampling and
   Analysis Methods for Hazardous Waste
   Combustion," EPA-600/8-84-002, PB84-
   155845, February 1984.

2. Johnson, L.D., and Merrill, R.G., "Stack
   Sampling for Organic Emissions,"
   Toxicological and Environmental Chemistry,
   6, 109 (1983)

3. Berg, S., Williamson, A.D., and Miller,
   B.C., "Design and Fabricate a Corrosion
   Resistant Organic Sorbent Module,"
   Appendix D, in "Modified Method 5 Train
   and Source Assessment Sampling System
   Operators Manual" Draft Final Report,
   Task 16, Contract 68-02-3627, August 1983.

4. Cooke, M., DeRoos, F., Rising, B.,
   Jackson, M.D., Johnson, L.D., and Merrill,
   R.G., "Dioxin Collection from Hot Stack
   Gas Using Source Assessment Sampling
   System and Modified Method 5 Trains - An
   Evaluation," Presented at Ninth Annual
   Research Symposium on Land Disposal,
   Incineration, and Treatment of Hazardous
   Waste, Ft. Mitchell, KY, May 1983.

5. Jungclaus, G.A., Gorman, P.G., Vaughn,
   G., Scheil, G.W., Bergman, F.J., Johnson,
   L.D., and Friedman, D., "Development of
   Volatile Organic Sampling Train (VOST),"
   Presented at Ninth Annual Research
   Symposium on Land Disposal, Incineration,
   and Treatment of Hazardous Waste.
   Ft. Mitchell, KY May 1983.

6. Jungclaus, G.A., Gorman, P.G., and
   Bergman,  F.J.,  "Sampling and Analysis
   of Incineration Effluents with the
   Volatile Organic Sampling Train (VOST),"
   In Proceedings:  National Symposium on
   Recent Advances in Pollutant Monitoring
   of Ambient Air and Stationary Sources,
   Raleigh,  NC, May 1983, EPA-600/9-84-001,
   January 1984.

7. Hansen,  E.M., "Protocol for the Collect-
   ion and Analysis of Volatile POHCs
   Using VOST," EPA-600/8-84-007, PB84-
   170042,  March 1984.

8. Thrun,  K.E., Harris, J.C., and Beltis,
   K.,  "Gas Sample Storage," EPA-600/7-
   79-095,  PB 298-350, April 1979.
Beltis, K.J., DeMarco, A.J., Grady, V.A.,
and Harris, J.C., "Stack Sampling and
Analysis of Formaldehyde," Presented at
Ninth Annual Research Symposium on Land
Disposal, Incineration, and Treatment
of Hazardous Waste, Ft. Mitchell, KY.,
May 1983.
                                           158

-------
                                           1
                                           OT.
                                            o>
                                            v>
                                            $
                                            o
                                            co
159

-------
                                                 c

                                                 2
                                                 *-

                                                 tn

                                                 a
                                                 o
                                                g


                                                CM'

                                                 
160

-------
Ifl
3
                                                                to
                                                                O
                                                                 c

                                                                '(0
                                                                 O)
                                                                 (O

                                                                 S"
                                                                 O
                                                                 _a>


                                                                 1
                                                                 O



                                                                 n

                                                                 

                                                                 >
                                                                 u.
                       161

-------
                    INCINERATION OF HAZARDOUS WASTE IN POWER  BOILERS:

                EMISSIONS PERFORMANCE STUDY RATIONALE  AND  TEST  SITE MATRIX

                                    Robert  A. Olexsey
                               Incineration Research Group
                            Energy  Pollution Control Division
                        Industrial  Environmental  Research  Laboratory
                           U.S.  Environmental  Protection Agency
                                    Cincinnati,  OH  45261
 INTRODUCTION

 Purpose

      The  purpose  of this document  is to
 summarize the  background,  rationale, and
 current study  approach for the  U.S.
 Environmental  Protection Agency  (EPA)
 research  and assessment program  relating
 to the incineration of hazardous waste in
 Industrial boilers.  The overall objective
 of the boiler  program is to evaluate the
 technical and  environmental acceptability
 of the use of  hazardous wastes as  fuels in
 industrial boilers, to aid in an Agency
 determination  on  the need  for the  nature
 of regulations of this practice under the
 Resource  Conservation and  Recovery Act
 (RCRA).

 Background

      Under RCRA,  regulations and specific
 process performance standards have been
 developed and promulgated  governing the
 issuance  of permits for the operation of
 incinerator facilities for hazardous waste
 control  (January  23, 1981).  These
 standards, however, do not currently apply
 to the use of combustible  hazardous wastes
 as fuels  in,energy recovery operations
 such as power boilers, process heaters,
 and high temperature industrial  processes
 (cement kilns, steel furnaces, etc.).

     The use of hazardous waste as fuels
 in power boilers  is known to be a wide-
 spread practice.  As much as 19 million
metric tons of hazardous waste may be
 disposed  of this way.   EPA, however, has
 chosen to defer a  regulatory decision on
 this  practice  because specific data on the
 extent of the  practice  were not available
 (i.e., number  of sites, type of wastes,
 operating conditions),  nor was there data
 on the destruction efficiency or the
 emissions performance of boilers employing
 hazardous waste as a fuel supplement.
 Other factors  in this decision are docu-
 mented in the  Federal Register, Volume
 45, No. 98, May 19, 1980.

 Problem Definition

      While an  initial decision was made to
 not regulate hazardous waste incineration
 in boilers, it was also recognized that a
 potentially significant loophole existed
 in the regulatory structure.  The same
 combustible hazardous waste which could be
 incinerated only under strict RCRA regula-
 tory  standards at the 300+ incineration
 facilities in the United States could also
 be incinerated as a fuel supplement in any
 of the over 400,000 boilers in the United
 States with virtually no environmental
 controls, save those involving the storage
 of the wastes and their transport to the
 boiler.   The EPA, through the Office of
 Solid Waste (OSW) and the Office of
 Research and Development (ORD), initiated
 a program of research and performance
testing to evaluate the engineering and
environmental  acceptability of hazardous
waste incineration  in boilers and to
determine the need  for and nature of
 regulatory standards to control  the
practice.
                                            162

-------
     Because of the large number of
facilities, boiler types and operating
conditions, and wastes which potentially
need to be evaluated with limited resources,
a considerable amount of thought is
necessary to carefully select a research
and performance testing approach which
optimizes the usefulness of the informa-
tion collected.

     The purpose of the balance of this
document is to summarize the stepwise
process, important assumptions, and
rationale which form the basis for
the boiler testing, program now underway.

Limiting Assumptions

     With limited funding available for
field testing, any such test program must
be directed toward obtaining only essen-
tial information in an economical manner.
A test program that set out to attack the
450+ listed RCRA hazardous wastes across
the full spectrum of boiler types in an
unordered fashion would lack credibility
and accomplish little in resolving the
issue of whether boiler co-disposal of
hazardous waste can be an acceptable
practice.   In 1975, there were about
400,000 industrial boilers of all sizes
and types in operation.  In order to be
realistic,  the test program must be
directed toward boilers that- are actually
used to burn wastes, and combinations of
wastes and  boilers that truly represent
test cases  and will allow generalizing
extrapolations to be made across the
entire waste/boiler matrix.

     Some simplifying assumptions can be
made:

1.  Each owner/operator of an industrial
boiler will use his boiler only  for
disposal of in-plant generated wastes and
will not likely accept waste from any
other plant.. The logic behind this
assumption  is that the primary advantage
of  boiler disposal  (aside from energy
conservation) is that the practice is
exempt from RCRA regulations if  the waste
is  not transported, stored, or disposed of
off-site.   Any  facility which accepts
wastes that are not produced onsite must
be  permitted  as a Treatment, Storage  and
Disposal  Facility  (TSDF) under RCRA.  To
accept  "imported" wastes could jeopardize
the "exempt"  status of the  boiler  facility.
2.   Most hazardous wastes are generated
by industry.  This will exclude the large
number of "industrial" size boilers that
,are located at commercial and institutional
facilities such as shopping centers and
hospitals.

3.   It is unlikely that wastes will be
burned in boilers smaller than 10 MM
Btu/hr.  This size (7,000 Ib/hr of steam,
1.0 megawatts of electricity) is probably
the minimum size at which combustion units
can be retrofitted to accept significant
quantities of wastes without negatively
impacting the thermal operating charac-
teristics of the units.  Boilers below
this size range are generally natural-gas
or oil-fired firetube or cast iron units
used for space heating.  The 10 MM Btu/hr
cutoff point is also convenient in the
most boiler inventory reports (NEDS, EADS,
etc.) do not report on boilers below that
capacity.  There are roughly 43,000
industrial boilers in the United States
with capacities larger than 10 MM Btu/hr
(1).

 4.  Most hazardous wastes that will be
burned in boilers are generated by the
organic chemicals industry.  Table 1
describes the number of boilers with
capacities larger than 10 MM Btu/hr for
the top 10 hazardous waste generating SIC
codes (2).  As can be seen from Table 1,
the largest hazardous waste generation is
from the chemicals industry.  The chemicals
industry also has a large number of
boilers.  Therefore, the opportunity for
disposal of hazardous wastes in boilers is
greatest in the chemicals industry.  The
primary metal industries generate about 10
percent of the hazardous wastes, but
wastes from this industry are less likely
to be combustible than are wastes in the
chemicals industry.  The largest sources
of organic hazardous wastes that are
available to boilers for disposal are the
chemicals, petroleum refining, and paper
industries, in that order.

     Through the application of these
limiting assumptions, the universe of
industrial boilers of concern can be
reduced to less than 5,500.  However,
these 5,500 boilers are  accessible to
about 70 percent of total hazardous waste
generation and over 95 percent of all
potentially combustible  hazardous wastes
generated in the United  States.
                                            163

-------
WASTE MATRIX DEVELOPMENT

Haste Property Consideration

     Limiting consideration (for the
purpose of scoping the proposed field
testing program) to the three target
industries greatly reduces the boiler
population of concern.  However, these
three industries are known to generate.
over 400 of the 456 RCRA listed wastes' in
the F, K, P, and U Series.  Of course,
funding provisions do not allow for
testing all 400 waste compounds against
all 5,500 individual boilers.  The wastes
and boiler types must be ordered in such a
manner that the limited test program
answers the crucial question of which
waste and boiler combinations are techni-
cally suitable and which are to be avoided.

     As a first cut, wastes can be sorted
into Good, Potential, and Poor candidates
for combustion in any operation.  This
process was conducted in development of
the Engineering Handbook for Hazardous
Waste Incineration (3).  The categoriza-
tion process was approached as follows:
       Waste Containing

o  Carbon, hydrogen, and/or
   oxygen

o  Carbon, hydrogen, 30
   percent by weight,
   chlorine and/or oxygen

o  Carbon hydrogen, and/or
   oxygen, <30 percent by
   weight chlorine, phos-
   phorous, sulfur, bromine,
   iodine, or nitrogen

o  Unknown percent of
   chlorine

o  Inorganic compounds

o  Compounds containing
   metals
Combustion
 Category

   Good
   Good
 Potential
 Potential


    Poor

    Poor
                   TABLE 1.  BOILER  POPULATION AND HAZARDOUS WASTE
                             GENERATION BY SIC CODE (Reference 2)
Hazardous Waste Generation
103 Metric TPY
28
33
29
34
40
26
37
36
31
35
Chemicals and Allied Products
Primary Metal Industries
Petroleum Refining
Fabricated Metals Products
Non-Manufacturing Industries
Paper and Allied Products
Transportation Equipment
Electrical and Electronic Equipment
Leather and Leather Products
Machinery
25510
4061
2118
1997
1971
1296
1241
1092
474
323
Number of Boilers
% of Total w/Capacity >10? Btu/hr
62.0
9.8
5.1
4.8
4.8
3.1
3.0
2.6
1.1
0.8
3346
1417
1177
399
No Data
1584
775
318
No Data
432
                                           164

-------
     Of course, mixed wastes can contain
compounds that represent each category.
An appropriate testing matrix would
encompass wastes from each category.

Incinerability

     The wastes of concern can be sorted
into good, potential, and poor candidates
for incineration as a function of gross
physical and/or chemical parameters as
described above.   However, once inorganic
and metallic compounds are dismissed as
incineration candidates, the remaining
organic materials must be further categor-
ized in some hierarchy of thermal stabil-
ity to ensure that a variety of thermal,
as well as physical conditions are exam-
ined.  Each of the numerous approaches
that have been taken to define "inciner-
ability" to date can be best described as
representing either one of two basic
premises:  kinetics or equilibrium.  The
kinetics premise assumes that the energy
required to ignite a given molecule is the
primary determinant of how difficult it is
to "destroy" the molecule in combustion.
The equilibrium premise holds that the
energy required to complete the combustion
process  (to produce carbon dioxide and
water) is the  dominant concern.  The
kinetics approach  is best represented by
^99.99-  The equilibrium approach is
best represented by the heat of combustion
AHc/MW.  For reasons presented in several
forums,  neither method  is an unquestion-
ably accurate  description of "incinerabil-
ity."  However, EPA has chosen to utilize
heat of  combustion (in  Kcal/gram) as the
currently recommended  Principal Organic
Hazardous Constituents  (POHC) selection
criteria.  The AHc approach will be the
primary  incinerability  parameter used to
select wastes  for  test  firing in this
boiler test program  (4).

     The main  advantages  of using the
AHc method for boiler  waste test firing
are that the method  is  comprehensive
 (there  exists  at  least  a  computed AHc  for
each waste in  Appendix  VIII) and that the
rfaste/boiler  selection  can  be fitted  into
the waste/incinerator  permitting network.
The AHc  method is, as  described  earlier,
an  inexact incinerability measure.  To
subdivide the  AHc  listing  too finely  would
be  imputing more  accuracy  to the inciner-
ability index than can be  supported.
Therefore, it  is  proposed  here  to  divide
the AHc list into three groups in describ-
ing difficulty to incinerate:
POHC Group
II
III
AHc Range (Kcal/g)
0 - 3.99
4.00 - 6.99
7.00 - 10.00+
     This split results in three subgroups
of roughly equal size.  Examination
of the POHC constituents within the
categories lends some empirical credibil-
ity to the system.  Materials ip Group I
tend to be highly halogenated ethylenes,
benenes, and ethanes.  Group II contains
lesser halogenated materials and process
production bottoms.  Group III contains
the nonhalogenated solvents and solvent
recovery bottoms.  In essence, Group I
contains materials that would not normally
be considered for burning as fuels in the
absence of RCRA.  Group III contains
materials that  have historically been
considered as candidate fuels for combus-
tion operations.

Waste Classification Matrix

     Three characteristics of a given
waste material  are important in mapping
out a combustion testing program:

     1.  Thermal stability of the waste
         (incinerability)
     2.  Volume of waste .generated in
         the-United States
     3.  Physical state of the waste
         at 25C.

     The thermal stability of the waste is
important in  aiding waste selection that
will provide  information on the relative
ease of destroying various waste types.
The volume of waste produced is important
in that a goal  of the test program is to
test realistic  waste/boiler combinations.
A low volume  waste is less likely to be
burned in a boiler than is a high volume
waste.  The physical  state of the waste
will impact the probability of boiler type
selection.  A waste that is a  solid or a
sludge will not likely be burned directly
in a gas- or  oil-fired boiler.

     Table 2  provides an example matrix of
waste  POHC, waste identification code for
some wastes that contain that  POHC, waste
description,  annual waste production rate,
and physical  state of the waste at 25C.
                                            165

-------
Table 3 summarizes the information  in
Table 2 in terms of EPA waste identifica-
tion numbers.  Implicit in  this  analysis
is the assumption  that  wastes  in mixtures
retain their discrete properties.
                         TABLE 2.   WASTE  CLASSIFICATION MATRIX

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
1.
2.
3.
4.

POHC Group I:
Trichl orof 1 uoromethane
Tetrachl oromethane
(Carbon Tetrachl on' de)
Hexachloroethane
Chloroform
Tetrachl oroethyl ene
(Perch! oroethylene)
Methyl ene Chloride
(Dichl oromethane)
Trichl oroethyl ene
Hexachl orobenzene
Tr,ichloroethane
Hexachl orobutadiene
Pentachlorophenol
Dichl oroethane
(Ethylene Dichloride)
Trichl orobenzene
2,4 Dichl orophenol
POHC Group II:

Vinyl idine Chloride
Dichl orobenzene
(1,2 Dichl orobenzene)
Phthalic Anhydride
Nitrobenzene
AHC
(Kcal/g)
0.1JL
0.24
0.46
0.75
1.19
1.70
1.74
1.79
1.99
2.12
2.19
3.00
3.40
3.81
4.45
4.57
5.29
5.50
Representative
Waste I.D.
F002
K073
K085
K009
F002
F002
F002
K030
F002
K016
- K001
K020
K015
K001
K020
F002
K022
K025
Generation
Rate
MT/YR
9,072
12,500
12,500
100,000
255,800
213,000
181,000
40,000
181,000
*
*
30,000
1,800
*
*
11,800
*
1,000
State at
25C
Li qu i d
Liquid
Solid
Liquid
Liquid
Liquid
Liquid
Solid
Liquid
Solid
Solid
Liquid
Solid
Solid
Liquid
Liquid
Solid
Liquid
                                          166

-------
Table 2 (continued)

5
6
1
2
3
4
5
6
7
8
9
10
*

POHC Group II (cont.)
. Toluene Diisocyanate
. Chlorobenzene
POHC Group III:
. Acetonitrile
. Phenol
Pyridine
. Acrylonitrile
. Methyl Ethyl Ketone
. Cresylic Acid
. Cresols
Naphthalene
Benzene
. To! uene
Undetermined Quantity
AHC
(Kcal/g)
5.92
6.60
*
7.37
7.78
7.83
7.93
8.07
8.09
8.18
9.62
10.03
10.14

Representative
Waste I.D.
K027
F002
K014
K022
K026
KOI 3
F005
F004
F004
K060
K001
K027

Generation
Rate State at
MT/YR 25 C
6,000 Solid
77,000 Liquid
6,000 Liquid
2,000 Solid
*
*
* Liquid
* Liquid
* Liquid
* Sludge
* Solid
* Liquid

          167

-------
                   TABLE 3.  WASTE MATRIX BY RCRA WASTE I.D. NUMBER
WASTE I.D.
POHC GROUP
PHYSICAL FORM
GENERATION RATE (MT/YR)
F001
F002
F004
F005
K001
K009
K011
K013
K014
K015
K016
K018
K019
K020
K021
K022
K023
K025
K026
K027
K029
K030
K036
K037
K042
K073
K085
K095/96
I
I, II
III
III
I, III
I
III
III
III
I
I
I
I
I, II
I
II, III
II
II
III
III
I
I, II
.III
III
III
I
I
I
Organic, Sludge, Solid
Liquid
Liquid
Liquid
Solid
Aqueous Liquid
Aqueous Liquid
Aqueous Liquid
Organic Liquid
Organic Liquid
Organic Liquid
Organic Liquid
Organic Liquid
Organic Liquid
Aqueous Metallic Sludge
Organic Sludge .
Liquid
Organic Sludge
Aqueous Liquid
Organic Sludge
Aqueous Liquid
Organic Liquid
Organic Sludge
Aqueous Sludge
Organic Solid
Organic Liquid
Organic Sludge
Organic Liquid
100,000*
100,000
-
-
65,000
-
-
6,000
47,000
50
3,200
35,000
8,000
125,000

100,000
-
1,000
1,000
6,000
1,200
61,000
-
-
-
125,000
6,600
34,000
* Estimated 10 percent of Production.
- Undetermined Quantity
     Aside from thermal stability, selected
waste characteristics may be desirable
from a testing perspective.  Two charac-
teristics that may be complicating factors
from a combustion perspective are the
                             presence  of  significant  amounts of water
                             and  metal  contaminants in the wastes.  A
                             hierarchy of wastes  in descending order of
                             difficulty to  destroy in boilers can be
                             envisioned as  follows in Table 4:
                                            168

-------
                         TABLE  4.   WASTE  CLASSIFICATION MATRIX
       CHARACTERISTICS
                                               WASTE  TYPE  EXAMPLE(s)
 cc
 o
Group I, Aqueous

Group I, Metals

Group I, Nonaqueous

Group II

Group III, Aqueous

Group III, Metals

Group III, Nonaqueous
  K001,  K009, K029

  K021

  F001,  F002, K015, K016, K018, K073, K085

  F002,  K020, K022, K023, K025, K036

  K011,  K013, K026, K037



  F003,  F004, F005, K001, K014, K022, K027, K036, K042
BOILER MATRIX DEVELOPMENT

Boilers In the Chemical  Industry

     The purpose of this testing program
is to demonstrate to what degree and which
boilers are capable of destroying which
waste.  Since the focus of the study has
been determined to be the chemicals
industry, we are interested in testing
boilers that are actually used in that
industry.
                                               Table 5 describes  the types  and
                                          capacities of boilers used in  the chem-
                                          icals industry.   By  far the largest number
                                          of boilers are oil  (46.2 percent) or  gas
                                          (42.0 percent) fired units.  Coal (10.3
                                          percent)  and waste  (1.5 percent)  fired
                                          boilers are far  less represented  in the
                                          chemicals industry.  Coal  boilers do  make
                                          up 13.8 percent  of total  boiler capacity,
                                          indicating that  the  average coal  boiler  is
                                          larger than either the  average gas- or
                                          oil-fired unit.
                      TABLE 5.  BOILERS IN THE CHEMICALS INDUSTRY
                                (SIC 28) (Reference 2)
FUEL
TYPE
Gas
Oil
Waste
Coal
NUMBER OF
10-50
542
679
13
41
50-100
276
351
13
74
BOILERS IN SIZE RANGE (106
100-200
261
299
12
128
200-400
206.
150
10
77
400-700
109
61
1
21
Btu)
700-1250
11
8
1
2
TOTAL NO.
BOILERS
1405
1548
50
343
TOTAL
CAPACITY
(109 Btu)
196.78
168.12
7.25
59.55
% TOTAL
CAPACITY
45.6
38.9
1.7
13.8
TOTALS 1,275    714
                    700
443
192
22
3346
431.70
100.0
                                            169

-------
Existing Data and Previous Studies

     Substantial data exists on emissions
of regulated criteria pollutants from
power boilers.  However, there exists no
data base on destruction of hazardous
chemical compounds in industrial boilers.
Source emissions data for regulated
pollutants may be useful in characterizing
a given boiler's ability to destroy
hazardous chemical compounds.  These will
be discussed in more detail in a subse-
quent section of this document.  Some
attempts have been made to directly assess
the capabilities for boiler destruction of
hazardous wastes through modelling and
engineering analysis.  These attempts will
be discussed in this section.

     In two separate studies, conducted in
1980 and 1981, Acurex Corporation examined
thermal conditions for distinct boiler
types (1, 5).  Time above temperature
predictive plots were developed for each
boiler type.  Fuel particle path analyses
were conducted at full and half loads.
The environment that each boiler presented
to a given fuel was described in terms of
conditions of time and temperature re-
quired for heating, volatilization,
thermal cracking, and oxidation of th,e
fuel.  Zonal heat balances were applied to
estimate conditions at each point in the
boiler.

     The Acurex analysis produced some
general conclusions.  It appeared that the
larger the boiler the greater the likeli-
hood that conditions for waste destruction
will exist.  The medium-sized watertube
units present mean conditions that are
conducive to destruction of most wastes
but have significant potential  for excur-
sion into conditions not favorable for
waste destruction at the margins of
operating conditions.  The firetube
boilers present high temperatures and good
flame exposure but the fuel residence time
appears to be of insufficient duration to
allow for volatilization and oxidation of
the waste fuel particles.

Haste Destruction Surrogates

     In the absence of specific waste
Destruction and Removal  Efficiency (ORE)
data for boiler firing of hazardous
wastes, examination of some boiler opera-
tional  parameters and stack emissions data
could prove helpful in assessing a boiler's
ability to destroy hazardous wastes.
None of these candidate ORE "surrogates"
have been verified in experimentation or
testing but discussion of the various
available parameters is of-some value.

     ORE is primarily a function of
temperature.  Flame exposure, residence
time, and mixing are also factors that
affect the degree of oxidation of a
particular molecule.  For a boiler with
all other factors held constant, waste
destruction will more than likely occur as
described below:

     ORE will increase as the following
     items increase:

        Boiler load
        Mean gas temperature
        Mean gas residence time
        C02 emissions
     ORE will increase as the following
     items decrease:

        Excess air
        02 emissions
        Boiler surface-to-volume ratio
        CO emissions
        Total HC emissions

Examination of the above listing leads to
a winnowing out process.  While it is true
that boiler combustion efficiency in-
creases with load tip to optimum (generally
nameplate) output conditions, it is also
true that the same boiler can operate at
different load conditions.  Boiler load is
not a machine characteristic but an
operating characteristic and cannot be
used to distinguish between boiler types.
The same statement can be made concerning
excess air, C02 emissions, and QZ
emissions.  Excess air is largely a
function of load conditions.  C02 and
02 emissions are excess air- and load-
dependent.

     Table 6 describes pertinent machine
specific parameters and emissions that can
be used to classify boilers with respect
to their estimated ability to destroy
hazardous compounds (1, 6, 7, 8).  Each of
the parameters presented in Table 6 will
be described in detail.
                                            170

-------
     Mean gas temperature and residence
times are significant from a waste -kine-
tics perspective and were utilized exten-
sively in the Acurex analysis.  However,
at temperatures above the kinetics operat-
ing zone and in the presence of a flame,
time-temperature paths are not sufficient
to fully describe compound destruction and
require detailed knowledge of the kinetic
behavior of the target compounds to be
effective.
     Table 6 does show that while the
temperatures in the firetube boilers
are very high, the bulk gas residence
times are so short that there is less
certainty that all particles of a given
material will "experience" the complete
kinetics process (e.g., volatilization,
pyrolysis, ignition).
           TABLE 6.  BOILER CLASSIFICATION PARAMETERS (References 1,  6,  7,  8)
BOILER
TYPE
Firetube
Fi retube
Watertube
Wate'rtube
Watertube
Watertube
Watertube
Watertube
Watertube
BOILER
FUEL
Natural Gas
Oil
Natural Gas
Oil
Stoker Coal
Stoker Coal
Natural Gas
Oil
Pulverized
Coal
BOILER
SIZE
(MM Btu/hr)
10-50
10-50
10-100
10-100
10-100
100+
100+
100+
150+
EXIT
T (F)
2400
2400
2300
2300
2060
2000
2200
2000
2000
RESIDENCE
t (sec)
0.23
0.22
0.50
0.50
1.50
1.90
0.82
0.77
1.82
S/V )
RATIO
1.60
1.60
0.80
0.80
0.30
0.22
0.52
0.52
0.27
CO
ng/J
8
15
15
14
146
86
17
14
10
HC
ng/J
3.4
1.7
2.4
3.7
3.6
1.7
0.7
3.0
1.7
NOX
ng/J
42
59
156
156
155
240
232
190
361
      Surface-to-volume ratio (S/V) is
 important  as  a descriptor for a potential
 ORE failure mechanism -- gas quenching.
 The larger the boiler tube surface with
 respect  to the total furnace volume, the
 more  likely that  a  particle of waste
 material will experience something less
 than  mean  furnace temperature.  As ex-
 pected,  S/V generally decreases with
 boiler  size.  The principal exception to
 this  rule  is  the  case of coal fired
 boilers.  For coal  boilers, the fuel
 temperature must  be kept below the ash
 fusion  temperature  to avoid slagging and
 fouling.  This lower temperature operation
 requires that, in order to produce a given
 pound  of steam,  a  given  coal  boiler must
 be much  larger than  an equivalent oil- or
 gas-fired boiler.  Therefore,  furnace heat
 liberation rates for coal-fired  boilers
 are much lower than  for  other types of
 boilers  (by a factor of  from  2 to 10).
 This factor largely  explains  why coal
 boilers  tend to  be much  larger (capacity
 and size-wise) than  other boilers.  The
 net effect is that while larger  boilers
 have lower S/V ratios than smaller boil-
 ers, coal boilers  have even lower S/V
 ratios than either oil or gas boilers in a
 similar  size range.
                                            171

-------
      CO and  HC emissions may be machine
specific  and are a measure of the com-
pleteness of combustion.  Unfortunately,
these emissions are also fuel- and opera-
tional-dependent.  Two coals or oils may
produce different CO and HC emissions due
to different fuel and operational condi-
tions. Two different machines burning the
same  fuel  under "optimum" operating
conditions may yield meaningful CO and HC
data.  In a  similar manner, HC and CO
emissions from the same boiler may provide
useful surrogate data on compound destruc-
tion.  However, the AP-42 data presented
in Table  6 is compiled in an a posteriori
manner and is not useful in characterizing
ORE.

      Data on NOx emissions, on the other
hand, can be very useful in describing a
given boiler's potential to destroy waste.
The importance of NOX emissions as a
boiler classification tool is discussed in
the next  section.

Nitrogen  as  a POHC Surrogate

      Production of oxides of nitrogen
(NOX) from fuelborne nitrogen occurs
relatively easily in combustion operations.
However,   production of NOX from airborne
nitrogen  is  an extremely energy intensive
operation that occurs ony at elevated
temperature  (assuming atmospheric pres-
sure).  In the temperature regions in
which boilers operate far more NOX is
produced from air than from fuel.  There-
fore, in a boiler, the most likely process
for production of NOX is the cracking of
the N molecule to N* and the oxidation
of the N+ to form N02, Nos, etc.
This highly energy intensive process is
thermodynamically similar to that re-
quired to break a carbon = carbon double
bond or a carbon - chlorine bond and
produce CO and C02 from fuel.  Therefore,
all other things being equal, the tendency
of a given boiler to produce NOX should
be a rough measure of that boiler's
ability to destroy a waste compound.  Of
course, all things are never equal  but
this sort of macroapproach is well  suited
toward use of the AP-42 data presented in
Table 6.

     Rearranging the presentation of the
data in Table 6 so that boilers are
displayed in terms of ascending NOX
emissions and deleting the CO and HC daa,
we obtain Table 7.  Examination of Table 7
yields boilers ranked in ascending  order
of size (capacity).  Also, it should be
noted that within boiler fuel  types, the
ranking by size is rigidly adhered  to
(i.e., the smallest oil  boiler listed has
the lowest characteristic NOX emissions
of all oil  boilers; likewise,  the largest
oil boiler has the highest characteristic
NOX emissions).  Similar boiler types
(e.g., firetubes) have similar NOX
emissions.
                                           172

-------
             TABLE  7.   BOILERS  RANKED  IN'TERMS OF HISTORICAL NOX EMISSIONS
BOILER
TYPE
Fi retube
Fi retube
Watertube
Watertube
Watertube
Watertube
Watertube
Watertube
Watertube
BOILER
FUEL
Natural Gas
Oil
Stoker Coal
Natural Gas
Oil
Oil
Natural Gas
Stoker Coal
Pulverized Coal
BOILER
SIZE
(MM Btu/hr)
10-50
10-50
10-100
10-100
10-100
100+
100+
100+
150+
Boiler Classification Matrix
NOX
Ng/J
42
59
155
156
156
190
232
240
361
cut makes
S/V
RATIO
1.60
1.60
0.30
0.80 '
0.80
0.52
0.52
0.22
0.27
sense when
EXIT T
("F)
2400
2400
2060
2300
2300
2000
2200
2060
2060
we consider
RESIDENCE
(sec)
0.22
0.23
1.50
0.50
0.50
0.77
0.82
1.90
1.82
that, at
     Through utilization of Tables 5, 6
and 7, a boiler matrix can be developed.
The size and type breakdown in Table 7
conforms to AP-42 structure and need not
impose a constraint on matrix development
but should be used as a guide to boiler
selection.  However, function as well as
convenience is served by generally adher-
ing to these categories.

     From Table 7, firetube boilers
represent a suspected worst case for
waste destruction.  From Table 5, over
one-third of the boilers found in the
chemicals industry are in the size range
of 10 to 50 MM Btu/hr.  Not all of those
small boilers are firetubes, but all
firetube boilers are found in this size
range.  From Table 7, regardless of fuel
type, the firetube boilers perform and
appear to be identical.  Therefore, one
boiler matrix constituent will be "Fire-
tube Boilers."

     The oil- and gas-fired watertube
boilers in the 10 to 100 MM Btu/hr size
range are twins and should be grouped
together (Table 7).  The size range spread
may appear to be large.  However, such a
tice is to go from single or multiple
burner configurations.  The larger and
more uniform flame zone should provide for
more constant and harsher combustion
conditions with the boiler.

     The oil^ and gas-fired boilers in the
greater than 100 MM Btu/hr size range
should also be grouped together.  From
Table 5, those boilers account for approxi-
mately 30 percent of the population of
boilers in the chemicals industry.  Too
fine a breakdown would be of little merit,
particularly since this size range appears
to be biased toward a best case with
increasing boiler size.

     Coal boilers warrant little attention
due to their small presence in the
chemicals industry (Table 5).  However,
coal boilers have two characteristics
which mandate interest:  1) the ability to
accommodate waste solids and sludges and,2)
low S/V ratios, long residence times and
relatively high NOX emissions.  The
ability to handle solids is unique to coal
boilers.  The latter point makes coal
boilers a potential best case.  Examina-
tion of Table 5 yields that the critical
                                            173

-------
mass of coal boilers in the chemicals
industry is in the larger size range (100
to 400 MM Btu/hr), little difference
between pulverized coal and stoker boilers
with respect to physical and operating
characteristics.  The higher NOX emis-
sions for the pulverized coal  boilers
would be a function of size rather than
type of boiler.  A single coal-fired
boiler entry on the matrix will  be suffi-
cient.
                             Table 8 lists the boiler categories
                        on the test matrix in ascending order
                        of assumed ability to destroy wastes.
                        WASTE/BOILER TEST MATRIX

                        Waste/Boiler Matrix Format

                             By overlaying the boiler test matrix
                        from Table 8 onto the waste classification
                        matrix from Table 4, one can obtain a
                        waste/boiler test matrix format.  This
                        format is described in Table 9.
                   TABLE 8.   BOILERS RANKED IN ASCENDING ORDER  OF  ASSUMED
                             ABILITY TO DESTROY WASTES
         CASE I.D.
BOILER TYPE
FUEL TYPE
A
LU

O

k B

c
GC
8 D
Fi retube

Watertube

Watertube

Watertube
Oil, Gas

Oi 1 , Gas

Oil, Gas

Coal
10-50

10-100

100+
s
100+
                                           174

-------
TABLE 9.  ECONOMY WASTE/BOILER TEST PROGRAM
                      (12 Tests)
                               WORST CASE
                          BOILER DESCRIPTIONS
WASTE
POHC
GROUP
DESCRIPTORS
PHYSICAL
QUALIFIERS
 CASE I.D.
BOILER TYPE
FUEL TYPE
CAPACITY (106 Btu/hr)
A"
FIRETUBE
OIL, GAS
10-50
- 	 &
WATERTUBE
OIL,' GAS
10-100,
C
WATERTUBE
OIL, .GAS
100+
0
WATERTUBE
COAL
100+
AQUEOUS
METALS
I
SOLIDS
i
UI
CO
o
t/0
o
NON-AQUEOUS
II
. AQUEOUS
METALS
III
SOLIDS
NON-AQUEOUS
X X
N/A N/A N/A
X X X
X
X
N/A N/A N/A
XXX

X



X

                           NA  -  Not Applicable
     Utilizing Table 9, we can structure
a test program that comprehensively
covers the spectrum of waste and boiler
types or one that encompasses only boun-
dary conditions within a proscribed
budget.

     The wastes are described by POHC
Group and Physical Qualifiers that may
impact boiler type selection and waste
destruction.  "Aqueous" implies signifi-
cant water in the waste.  "Metals" refers
to the presence of inorganic or metallic
contaminants in the waste which could  -
present problems for the boiler.  "Solids"
refers to the waste containing solid
materials or sludges.  "Non-aqueous"
refers to.a waste with no significant
water, solids, or contaminants.  Of
course, a-given waste could contain a
combination of conditions.  All permuta-
tions are not shown but can be accommo-
                          dated within the matrix.   It may be argued
                          that a  highly aqueous  Group  III waste can
                          be  a worse  case than a non-aqueous Group I
                          waste.   The wastes  are arrayed from top to
                          bottom  in order of  assumed worst to
                          assumed best case.  Group  II wastes are
                          not broken  down further since these are
                          not assumed to be limiting conditions.

                              The boilers are described by  firebox
                          type, fuel  type, and capacity.   The boil-
                          ers are arrayed from left  to right in
                          order of assumed worst to  assumed  best
                          case. .  Certain physically  constraining
                          cases can be assumed to exist.   For ex-
                          ample,  it is unlikely  that a solid-contain-
                          ing waste can be burned in an bil- or gas-
                          fired boiler unless that boiler  has been
                          converted to a liquid  fuel from  a  coal
                          operation or the waste fuel  has  been vola-
                          tilized in  some preprocessing application.
                       175

-------
      Additionally, air pollution control
 devices (APCDs) are not factored into the
 matrix.  Smaller industrial boilers (10 to
 50 MM Btu) generally have no APCDs.
 Larger coal-fired boilers may have multi-
 clones (stokers) or electrostatic precipi-
 tators (pulverized coal), but large oil-
 and gas-fired industrial  boilers generally
 have no APCDs.  It is felt that the
 addition of factors for presence, absence,
 or types of APCD would overly complicate
 the matrix discussion.

 Test Program Selection

      Ideally, of course,  a comprehensive
 test program would involve at least one
 test of each waste/boiler combination  and
 permutations of each  combination described
 in Table 9.  Such a test  matrix would
 require completion of at  least 60 tests.

      However, EPA currently has sufficient
 funds  for only an "economy" 12-test
 program.   Such a truncated program must
 focus  on  "worst case"  conditions.   Limited
 testing would be conducted on "best case"
 wastes and  only one Group  II  verification
 test could  be tested.   Comprehensive
 testing of  the most representative boiler
 type (B) would be emphasized.   Little
 exploration of qualifying  waste conditions
 (metals, aqueous)  would be conducted.
 Hopefully,  this  program, supported  by
 pilot-scale parametric testing,  would
 provide an  "envelope" around  acceptable
 waste/boiler  combinations.

 Utility and Flexibility of the  Matrix

     Almost every conceivable waste  and
 standard boiler  combination will fit into
 the  waste/boiler matrix.   Ideally, access
 to all types  of wastes and boilers can be
 secured.  In  the  "real" world of testing
 targets of  opportunity, the matrix can
 serve as a  guide to test selection.  For
 instance, if  eight test candidates are
 available,  the matrix would help ensure
 that five III B tests are not selected.

     Waste  volume is also a concern.  For
example, if there exists a choice between
two  IB combinations, the waste that has
the highest national generation rate would
be selected.
 Cautions  On the  Use of the Matrix

      The  matrix  is a tool for ordering a
 testing program.   Implicit in its use are
 certain assumptions.  The first assumption
 is  that all boilers are operated equally
 expertly.  It  is conceivable that all
 boilers,  if well designed and perfectly
 operated, can  achieve DREs of 99.99
 percent.  Within boiler types the over-
 riding factor  concerning waste destruction
 is  boiler operation.  It is entirely
 probable  that  a well operated "worst case"
 boiler will outperform a poorly operated
 "best case" boiler.

     The  second assumption is that addi-
 tion of the waste material to the boiler
 does not  alter boiler operating character-
 istics.   Any waste added in abundant
 quantities will indeed have an effect on
 boiler operating conditions.

     The  third assumption is that all
 wastes are equally likely to be burned in
 all boilers.   In fact, due to their
 potentially damaging impact on boiler
 hardware, certain wastes (particularly
 Group I highly chlorinated materials) may
 not be burned in boilers at all.
REFERENCES

1.  Castaldim", et al., "A Technical
    Overview of the Concept of Disposing
    of Hazardous Wastes in Industrial
    Boilers," Acurex Corporation for
    USEPA, lERL-Ci, October 1981.

2.  National  Environmental Data System,
    1982.

3.  Bonner, et al., Engineering Handbook
    for Hazardous Waste Incineration,
    Monsanto Research Corporation for
    USEPA, lERL-Ci, September 1981.

4.  Guidance Manual for Hazardous Waste
    Incinerator Permits, USEPA, QSW,
    and MITRE-'Corporation, September
    1982.

5.  McCormick, et.  al., "Engineering
    Analysis  of the Practice  of Disposing
    of Hazardous Wastes in Industrial
                                           176

-------
    Boilers," Acurex Corporation for
    USEPA, lERL-Ci, January 1982.

6.  Personal  Communication from Howard
    Mason, Acurex Corporation, to R.A.
    Olexsey,  USEPA, December 1982.
7.  "Compilation of Air Pollutant Emis-
    sions Factors," AP-42,  USEPA, 1982.

8.  Estimates and Interpolations by R.A.
    Olexsey, December 1982.
                                           177

-------
                       OBSERVATIONS OF FLUID DYNAMIC EFFECTS UPON
                    HIGH-TEMPERATURE DESTRUCTION OF ORGANIC COMPOUNDS
                             Wayne A. Rubey, John L. Graham,
                          Douglas L. Hall, and Barry Dellinger
                         University of Dayton Research Institute
                                       Dayton, OH
                                        ABSTRACT
     In the controlled high-temperature
Incineration of industrial organic wastes,
events that occur in the gas-phase are of
vital importance.  Within a rotary kiln-
afterburner incinerator the large volume
afterburner chamber is responsible for the
complete destruction of organic materials
that survive, or are emanated from the
previous exposures.

     The fluid dynamic behavior of the
high-temperature gas phase has been exam-
ined with special attention to the various
aspects of exposure time and fluid interac-
tion of gas-phase substances.  These
observations with respect to destruction
efficiency of organic substances have been
made on the laboratory-scale, and several
implications can be drawn with respect to
the gas-phase decompositions of both
hazardous organic compounds and their
thermal reaction products.
                                            178

-------
                  RESEARCH PROJECTS AT ERA'S LOUISIANA STATE UNIVERSITY
              HAZARDOUS WASTE RESEARCH CENTER - USEPA CENTER OF EXCELLENCE

                                     D.  P.  Harrison
                             Hazardous Waste Research Center
                               Louisiana State University
                              Baton Rouge,  Louisiana 70803
                                         ABSTRACT
     Now beginning its third year of opera-
tion, the Hazardous Waste Research Center
at Louisiana State University is conducting
fundamental and exploratory research in
three general areas:  chemical/material
interactions, incineration-, and alternate
methods of treatment/destruction.  The re-
search projects range from studies on the
rates of supercritical extraction of or-
ganics from solids to the destructibility
of pure chlorinated hydrocarbons in labora-
tory flames to an investigation of bonding
mechanisms associated with cement - stabi-
lized orgarncs.  Although administered
through LSU's College of Engineering, indi-
vidual research projects are being conduc-
ted by multidisciplinary groups represent-
ing a number of academic departments.
     In addition to its fundamental and
exploratory research efforts, the Center
has adtive and expanding programs in the
areas of applied research and technology
transfer.  The applied research component
will encompass projects of more immediate
interest to federal and state agencies and
to private industry.  The objective of the
technology transfer activities is to dis-
seminate state-of-the-art technology and
research advances to the professionals in
government and industry and also to the
nonprofessional public.  These latter ef-
forts are beyond the scope of the EPA-
funded exploratory research center and are
supported by the university, other govern-
mental agencies, and private industry.
                                             179

-------
                       HAZARDOUS WASTE CONTROL TECHNOLOGY DATA BASE
                                  Richard L. Holberger
                                  The MITRE Corporation
                                       McLean, VA

                                        C.C. Lee
                          U.S. Environmental Protection Agency
                                     Cincinnati, OH
                                        ABSTRACT
     The Hazardous Waste Control Technology
Data Base (HWCTDB) is an automated data
base sponsored by EPA's Incineration
Research Branch to store detailed technical
data relevant to the thermal destruction of
hazardous wastes.  The data base stores
design, operating, and performance data
obtained from incineration facility permit
applications, trial  burn reports, research
and development projects, and literature
sources.  The data base went on-line in
late 1982.
                                           780

-------
                   EVALUATION OF AN ADVANCED LOW NOX HEAVY OIL BURNER
                    FOR INCINERATION OF NITRATED HAZARDOUS COMPOUNDS
                            W.S. Lanier and Charles B. Sedman
                          U.S. Environmental Protection Agency
                               Research Triangle Park, NC
                                        ABSTRACT
     Among the RCRA-listed hazardous waste
compounds  is a variety of chemical species
which are highly nitrated.  During the
combustion of these compounds the bound
nitrogen atom is readily oxidized to form
nitrogen oxide (NO) in the incinerator or
boiler exhaust.  Limited test data has
shown that co-firing nitrobenzene with
natural gas  in an  industrial  boiler in-
creased NOx emission by a factor of 6 to 9
when the nitrobenzene feed rate represented
only 6 percent of the input heating value.
Accordingly, a concern exists that class-
ical incineration of such materials may
exchange destruction of the waste for a
potential NOX emission problem.

     A significant body of data on the
pyrolysis of nitrated compounds is avail-
able from the EPA NOX control program.
Species such as nitrobenzene, acryloni-
trile, and aniline were previously studied
as possible model compounds for nitrogen
bound within the structure of coal and
heavy residual fuel oils.  Those pyrolysis
studies clearly indicated that hydrogen
cyanide was the dominant nitrogen species
evolved.  Accordingly, incineration of
this class of waste raises concern that
HCN would be a potential  product of incom-
plete combustion (PIC) and that emission
of that PIC would be of equal or greater
concern than emission of the  principal
organic hazardous compound (POHC).

     Based upon considerations such as
those identified above, a need exists for
combustion technology which could assure
 high  destruction  efficiency  for  the  POHC
 and probable  PIC  and  simultaneously  mini-
 mize  the  emission of  NOX.  To  accomplish
 that  objective  requires  a  combustor  which
 establishes an  extremely high  temperature,
 fuel-rich primary combustion zone  followed
 by an excess  air  burnout region.   A  fuel-
 rich  primary  zone would  shift  nitrogen
 equilibrium almost  totally to  N2 whereas
 fuel-lean equilibrium  is substantially
 shifted toward  NO formation.   The  high
 temperature promotes  initial destruction
 of the POHC but also  accelerates chemical
 kinetic processes driving  fixed nitrogen
 species (such as  any  HCN PIC)  to N2-

      As part  of the EPA  NOX  control  pro-
 gram,  combustion  techniques  such as  that
 described above have  been  previously
 established.  A three million  Btu/hr
 version of this burner  is  available  at the
 in-house  research facilities of  IERL/RTP
 and a 60  x 106  Btu/hr version  is cur-
 rently being  field  tested  in Kern  County,
 California, on  an enhanced oil recovery
 steam generator.  The current  paper  re-
 ports on  an evaluation of  this low HO^
-burner technology for  incineration of
 three highly  nitrated hazardous compounds.
 Test  results  are  presented on  the  destruc-
 tion  efficiency of  the POHC, emission of
 HCN and NOX while burning  nitrobenzene,
 aniline,  and  acrylonitrile.  All waste
 compounds  were  doped  into  distillate fuel
 oil at concentrations of 5 and 25  percent
 by weight.  Test  results are being evalu-
 ated  prior to proceeding with  a full-scale
 field  verification  of this technology.
                                           181

-------
                    REGULATORY SUPPORT RESEARCH STUDIES SUPPORTED BY
                        EPA/CINCINNATI MSD COOPERATIVE AGREEMENT
                                      Boyd T. Riley
                                          RYCON
                                     Cincinnati, OH
                                        ABSTRACT
     A number of research studies have been
proposed and carried out to produce infor-
mation on various aspects of a publicly
owned, full scale hazardous waste incinera-
tor.  These studies are intended to produce
data to assist in the development, support
and revision of EPA's regulatory responsi-
bilities addressing hazardous waste incin-
erators.  Typical studies that have been
carried out include:  the preparation of a
design and economic history of the inciner-
ator complex, evaluation of options for
acid neutralization in combustion gases,
evaluation of DRE capabilities and poten-
tial monitors, and testing of alternative
modifications to particulate emission
control devices.  Anticipated studies
include:  developing a detailed understand-
ing of ash characteristics with regard to
hazardous properties, testing of various
pilot scale particulate control devices,
and preliminary evaluation of a colorimetric
DRE monitor.
                                            182

-------
                      FLAME-MODE HAZARDOUS WASTE DESTRUCTION RESEARCH

          Wm. Randall Seeker, John C. Kramlich, Michael P. Heap, Rachel K. Nihart
                       Energy and Environmental Research Corporation
                                 Irvine, California 92714

                                     Gary S. Samuel sen
                           Department of Mechanical Engineering
                                University of California
                                 Irvine, California 92717
                                         ABSTRACT
     The results of a laboratory-scale
study of flame-mode hazardous waste in-
cineration are reported.  The primary goal
of the work was to compare the various
incinerability ranking procedures with
rankings generated under flame conditions
typical of liquid-injection incinerators.
A secondary objective was the comparison
of flame and non-flame thermal destruction
behavior.  We previously reported (Ninth
SHRD Symposium) that, when operated under
efficient combustion conditions, flames
nearly quantitatively destroy organic
waste compounds.  For incomplete
destruction to take place, the flame must
be perturbed into inefficient operation.
The compound rankings resulting from
these perturbed states (termed failure
conditions) were dependent on the cause of
the perturbation.  The results presented
here represent an extension of the pre-
vious work.

     The results were generated using a
65 Kilowatt (kW) turbulent liquid spray
flame.  The flame enclosure was cooled to
concentrate on flame-zone processes by
quenching post-flame reactions.  In the
present study the flame was fired on
No. 2 fuel  oil doped with four model waste
compounds (3% by weight):  Benzene,
chlorobenzene, chloroform and acryloni-
trile.  The flame was operated under high
efficiency conditions by high and low
excess air, by degraded fuel  spray atomi-
zation quality, and by impingement of the
flame on a cold surface.  The'destruction
efficiency was measured by Tenax trapping
of the stack effluent followed by thermal
desorption and gas chromatographic/flame
ionization detector analysis.  Simul-
taneous measurements of CO, C02, 02, CH4
and total unburned hydrocarbons were
obtained.

     The results sustained the previously
reported conclusions that flames are
capable of high DE, and that loss of this '.
high DE is necessarily due to perturbation
of the flame into a failure condition.
Waste DE rankings were found to be
dependent of flame condition.  The new
results also indicated that a linear
relationship exists between DE and un-
burned hydrocarbon emissions from the
flame zone.  Correlation of DE and CO
measurements showed that CO emissions
must increase to substantial levels
before DE begins to degrade.  It should
be emphasized that these flame-zone
results may be modified by realistic
post-flame time-temperature histories.

     These results suggest a performance
monitoring and control  methodology in
which the difference between a correctly
operating incinerator and a failure
mode can be detected.   The CO measure-
ment could be used as  an early indication
of the approach of a failure condition.
It could also be used  as a means  of
controlling operating  parameters.   Total
hydrocarbon measurement could be  used
to indicate the presence of a failure
condition.   Significant hydrocarbon
emissions could be used to shut down waste
flow.
                                            183

-------
       EVALUATION AND DEVELOPMENT OF INNOVATIVE INCINERATION TESTING PROCEDURES

                                  Jerome B. Strauss
                                     Versar, Inc.
                                  6850 Versar Center
                                Springfield, VA  22151
                                       ABSTRACT
  This study Investigated Innovative
real-time monitoring techniques for haz-
ardous waste Incineration.  The objec-
tives were to survey and select monitor-
Ing systems that measured (either
directly or Indirectly) destruction and
removal efficiency (ORE), and to demon-
strate the effectiveness of the selected
system 1n a test program.  Seven poten-
tial monitoring systems were evaluated
using the criteria of operating charac-
teristics, limitations, applications,
degree of development, and cost.  The
selected system was an 1n-l1ne flame
1on1zat1on detector (FID) that measures
total organic carbon In the stack.

  Tests were performed at the EPA's Com-
bustion Research Facility 1n P1ne Bluff,
Arkansas.  Six test runs were conducted
using hexachlorobenzene and/or toluene
as surrogate hazardous wastes.  Total
organic carbon concentrations corres-
ponded 1n the same direction and
magnitude as the ORE values as measured
by Method 5 stack tests.  The results
from these tests Indicate that the
selected approach 1s a promising system
for on-Hne, real-time monitoring of
Incinerator performance.

  From the results of these tests 1t
1s recommended that further study be
performed to quantify the relationship
between TOC concentration and POHC
destruction and removal.  Appropriate
relationships need to be rigorously
challenged and tested 1f TOC concen-
tration 1s to become an accepted method
for monitoring ORE values.
                                            184
                                                            * U.S. GOVERNMENT PRINTING OFFICE: 1984 - 759-102/10696

-------

-------
o
o
^
VJD

oo
-C-

O
to
1X3
         2  n c
         -? a.
    -
X  w o
3  o -^
TO  O (B
^  r n
            = -o
            -
            =  2
          H.5  3
                      " 2.

                                                                                                                                 -
                                                                                                                                 s
                                                                                                                                 0)
                                                                                                                                 .
                                                                                                                                 
                                                                                                                                .


                                                                                                                              D 0> I
                                                                                                                              01 rt |
                                                                                                                              NJ
                                                                                                                              C3>
                                                                                                                              00

-------