INCINERABILITY RANKING OF HAZARDOUS ORGANIC COMPOUNDS
by
Robert E. Mournighan
Marta K. Richards
Howard Wall
Technology Research Section
Thermal Destruction Branch
Waste Minimization Destruction and
Disposal Research Division
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio
For Presentation at the 15th Annual Research Symposium on
Remedial Action, Treatment and Disposal of Hazardous Waste
Cincinnati, Ohio
April 10-12, 1989
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
INTRODUCTION
Since the regulation of hazardous waste by the USEPA, both the Agency and
the regulated community have been searching for means of evaluating incinerator
performance. The concept of a "trial burn,"- a series of tests designed to
evaluate the ability of an incinerator to process waste to certain standards,
was developed; and surrogates were introduced (1) for hazardous and solid waste
and for Principal Organic Hazardous Constituents, or POHCs. In essence, the
incinerator's ability to destroy waste was based on its performance in destroying
the POHCs.
Destruction and Removal Efficiency (DRE) is used by EPA to express
incinerator performance. The DRE of a POHC is calculated as follows:
ORE = (^in - Uput) x 100 (2)
^in
Where win and W0ut are the mass flow rates of the POHC
input and output (at the stack) respectively.(2)
As regulations and policies regarding trial burns developed, it became
apparent to all concerned that DRE determinations were rapidly becoming extremely
expensive and time consuming. It is not unusual to see trial burn costs
approaching 2% of the capital costs of the facility. Largely responsible for
these costs was the requirement that several types of POHCs be evaluated,
resulting in the use of multiple sampling trains. The process of obtaining
representative samples could take a week, or longer. Sample processing and
analysis, also expensive, would drag out the time-span between trial burn and
results to over three months. The objective for regulators and regulated alike
-------
was to search for a less expensive approach. The approach was similar to that
taken for POHCs and solid waste: find a surrogate for the waste being burned
and develop a cheap, quick analytical method for the analysis of the surrogate.
Several compounds, e.g., Sulfur hexafluoride (SFg) and Freons^ fluorocarbons,
were put forth as candidates.(3)
Because of its high thermal and chemical stability and the fact that it
is inexpensive and non-toxic, SFg received much attention as a surrogate. Quick,
relatively inexpensive, and reliable methods were developed for its analysis,
and initial evaluation began in 1984.(4) A system employing a gas chromatograph,
equipped with an electron capture detector (GC/ECD), was used to obtain SFg
concentrations in stack gas every 2 to 4 minutes, a much shorter time than the
sampling time required for sampling trains. Equipment set-up, operation, and
calculation of results could be done in 1-1/2 days (4), a far cry from three
months.
Since that time, a great deal of effort, time and money has been spent,
both in the United States and Canada (5), to exhaustively evaluate SFg as a POHC
or waste surrogate.
During the same time period, the concept of using a standard POHC mixture
(a "POHC Soup") for trial burns, by employing a group of compounds with a range
of chemical and physical characteristics, was developed. A paper describing the
rationale and results of that effort is being given at this conference.(6)
Both of these efforts have been researched extensively at the laboratory
stage (5, 6, 7). An evaluation program at pilot scale for the USEPA Combustion
-------
Research Facility (CRF) was proposed in 1987. The CRF has two pilot-scale
(3 x 10^ Btu/hr) incinerators, a liquid injection system and a rotary kiln.
The SFg/"POHC Soup" program was performed in the liquid injection system. This
paper describes the ORE results and discusses the implication of this work and
of other researchers.
Interest in assuring and demonstrating compliance with RCRA permit
conditions has also concerned the hazardous waste community. Trial burns are
prohibitively expensive to carry out on a routine or even annual basis, with
not much gained in the process. This approach still leaves "gaps in coverage,"
where incinerator performance may be unknown. It became apparent, after initial
research at EPA, that residence time, temperature and turbulence were not the
only factors to consider in the evolution of incineration criteria. The Toxic
Substance Control regulations (8) for PCB incineration reflected that
realization, and stipulated not only time and temperature, but also a minimum
stack 02 concentration as well as a minimum combustion efficiency of 9,9.9% are
required.
This paper concentrates on the evaluations of SFg and the "POHC Soup" as trial
burn surrogates.
EXPERIMENTAL BACKGROUND
The experimental program was executed at the USEPA Combustion Research
Facility's Liquid Injection System (LIS). Figure 1 is a sketch of the unit,
showing each of the elements that make up the incinerator and the air pollution
control system.
-------
Stack
Building wall
Afterburner chamber _~
(unfired) \
Sampling port
'Quench chambei
Main chamber
Aux. propane
Liquid feed'
Atomizing steam
Combustion air
HEPA Carbon bed
filter/^ absorber
..Ionizing wet
/scrubber
-4-Pack
column
scrubber
Venturl scrubber-/
Scrubber liquid tank
Figure 1. CRF liquid Injection system.
-------
The LIS is fired with liquid waste, and propane is used to maintain
temperature control. Combustion air is supplied by forced-draft fan. The waste
mixture containing the "POHC Soup" was stored in a stainless steel, vessel and
pumped continuously to the LIS.(9) SFg was injected into the liquid feed stream
as a gas, where it dissolved into the liquid phase just prior to incineration.
Sampling of the stack gas for both SFg and the POHCs was done at the exit
of the air pollution control systems, just after the ionizing wet scrubber (IWS).
Even though the gas passes through additional gas cleanup, sampling further
downstream was not conducted. The final gas cleanup is unique to this facility
and is not standard for commercial and private incinerators. Table 1 shows the
experimental conditions and the ORE results for the SFg and the POHC components.
Oxygen and incinerator temperature were varied as part of a designed
experiment. All other variables, such as waste composition and flow rates, were
kept constant.
The compounds used in the test mixture were toluene (70%), chlorobenzene
(10%), pentachlorobenzene (10%) and tetrachloroethylene (10%). This combination
of compounds resulted in a mixture which contained both volatile components and
semivolatiles. Stack sampling of these components was accomplished with the
VOST (USEPA Method 0030) and Modified Method 5 (USEPA Method 0010). The SF6
sampling method was proportional gas sampling, with the sample being fed directly
to GC/ECD.
-------
Table 1. Data Table - Temperature, Oxygen and
DREs (number of nines)
Afterburner ORE
Exit Tetra-
Tempera- Exit chloro- Chloro- Pentachloro-
Experiment ture, °C Oxygen % SFg* ethylene* Toluene* benzene* benzene*
]
2
3
4
5
6
7
8
9
10
Calculated
1030
1114
943
1274
1091
945
1310
1077
1175
1105
ORE by
2.3
3.3
5.6
1.3
5.1
8.2
4.5
9.2
8.0
4.7
4.35
3.59
3.33
7.00
5.31
4.17
5.33
3.44
5.70
3.96
the formula ORE
5.40
5.52
6.14
5.51
5.74
6.66
5.34
5.85
6.70
5.16
= -Log
6.19
6.21
7.10
6.11
5.85
6.68
5.59
5.74
7.55
6.39
C-DRE \
•m)
6.11
6.04
6.38
6.40
4.14
6.66
5.80
6,28
6.54
5.68
7.80
7.52
7.34
7.30
6.98
7.41
7.49
7.35
7.43
7.44
ANALYSIS OF RESULTS - SFg
To gain as much information about the behavior of the POHC/SFg OREs with
respect to the two variables, a regression analysis was performed for each data
set, one for each compound. Table 2 lists the results of the regression
analysis, the model used and the correlation coefficient.
Figure 2 is a contour plot vrith independent variables, temperature and
afterburner exit oxygen concentration, on the Y and X axis respectively. The
SFg ORE (expressed in the number of nines) is displayed as contours of the
regression surface. This figure demonstrates why it is so hard to make solid
judgments about the relationship between ORE and a single variable. As the
figure clearly illustrates, whether ORE goes up or down with increasing §2
-------
SF6 ORE VS T, O2
1400
T
^1300H
P
E
R 1200
A
T
U 1100-
R
E
1000 H
900
^
I I i I I I I I I
0123456789 10
O2 CONCENTRATION
• 3-9's
D 6-9's
ORE
+ 4-9's
x 7-9'8
* 5-9's
Figure 2. SF6 ORE vs. T,
-------
depends on the operating temperature. At 950°C, ORE increases from about
3-nines to nearly 4-nines as oxygen increases. At 1350°C, the same change
results in the opposite effect, dropping the ORE from greater than 7-nines to
les§ than 5-nines ORE as oxygen increases.
Table 2. Data Table
Correlation
Compound Model Coefficient
SFg 1st Order 0.77
Toluene 1st Order 0.45
Tetrachloroethylene 1st Order 0.75
Chlorobenzene 2nd Order 0.76
Pentachlorobenzene 2nd Order 0.68
At this point, it is stressed that the regression analysis and the
resulting figures developed from it should only be used in general, not
predicting absolute DRE figures for specific compounds. Until additional data
and developed, these relationships should only be applied to the CRF incinerator
and can only be used to describe the relationship between dependent (DRE) and
independent (Temp, and 02) variables within the range of the experiment.
Extrapolation of the results outside the range is risky.
In the laboratory work done on the thermal decomposition of SFg at the
University of Dayton Research Institute (UDRI) (7), it was determined that SFg
destruction was independent of oxygen concentration, indicating the mechanism
to be unimolecular decomposition. Although Figure 2 shows an effect of $% on
SFg DRE, it may not mean that 02 specific!ally causes the changes. What is not
depicted is that as the D£ concentration increases, air to fuel ratios and
incinerator residence times change concurrently. The variation in SFg DRE may
-------
be related to those effects and is not necessarily inconsistent with the UDRI
results.
Figure 3 is a plot of the SFg DREs versus temperature with 3 data sets
Illustrated. The data shown as diamonds were the SFg DRE-temperature
relationship calculated from the data supplied in References 5 and 7. This was
calculated for a 2-second residence time and represented the ORE temperature
relationship for SFg exclusive of any other physical or chemical effects.
The data plotted as Xs were data from the work conducted at the Alberta
Environmental Centre (5) in which the effect of the presence of refractory on
SFg ORE was evaluated. As one can see, the presence of refractory in the reactor
had a marked effect on the DRE-temperature relationship, reducing the required
temperature by some 200°C.
The third data set plotted in Figure 3 was the CRF data, and shows the
regression line for those data points. Here, as with the Canadian data, there
is a marked difference in SFg DRE-temperature behavior. While some of the
difference could be attributed to the presence of refractory and residence time
changes in the incinerator, the authors feel that there is more to it than that.
Since the laboratory data (UDRI) were taken with non-flame systems, in which the
substances underwent an extremely narrow temperature distribution, SFg ORE should
not be expected to be similar, since in flame conditions, SFg would "see" or
experience a wide range of temperatures in the incinerator environment. At any
rate, the behavior is not the same, and more investigation is warranted.
-------
8
7
6
5
4
3
SF6 THERMAL DECOMPOSITION
PILOT/LAB COMPARISON
ORE (-LOG(1-DRE/100))
D
D
D
800 900 1000 1100 1200 1300
TEMPERATURE, DEG C
1400
REF 5
CRF REGRESSION
-&- REF 6 and 7
° CRF DATA
Figure 3. SF6 Thermal decomposition pilot/lab caparison.
10
-------
POHC RANKING
Figure 4 is a contour plot of the relationship between conditions necessary
to achieve 6-nines SFfi ORE and that of the individual POHCs. It also illustrates
why looking at one variable at a time can be confusing. At low 02
concentrations, tetrachloroethylene requires only 900°C for 6-nines while at 4.5%
02, 1250°C is required, surpassing that for chlorobenzene, which was more stable
at the lower 02 concentrations. Pentachlorobenzene DREs were never below
6-nines and therefore are not plotted.
Ranking the POHCs is shown in Figure 4. This figure illustrates that
rankings can change, depending on combustion conditions. This is also true
depending on the nature of the experimental device.
Table 3 is a list of rankings of POHCs derived from Figure 4 and from data
supplied by UDRI in Reference 9. Although not definitive, the research into
POHCs and POHC rankings have produced results which seem to make surrogate use
even more questionable than when it was first suggested.
Table 3. POHC Ranking With 02 Present3
LIS Incineration
UDRI Low 02 (1.5%) High 02 (5%)
Tetrach 1 oroethy 1 ene
Toluene
Pentachlorobenzene
Chlorobenzene
1
2
5
3
4
1
4
2
5
3
2
3
1
5
4
al being most stable; 5 being least stable.
11
-------
POHC RANKING
Figure 4 is a contour plot of the relationship between conditions necessary
to achieve 6-nines SFg ORE and that of the individual POHCs. It also illustrates
why looking at one variable at a time can be confusing. At low 02
concentrations, tetrachloroethylene requires only 900°C for 6-nines while at 4.5%
02, 1250°C is required, surpassing that for chlorobenzene, which was more stable
at the lower 02 concentrations. Pentachlorobenzene DREs were never below
6-nines and therefore are not plotted.
Ranking the POHCs is shown in Figure 4. This figure illustrates that
rankings can change, depending on combustion conditions. This is also true
depending on the nature of the experimental device.
Table 3 is a list of rankings of POHCs derived from Figure 4 and from data
supplied by UDRI in Reference 8. Although not definitive, the research into
POHCs and POHC rankings have produced results which seem to make surrogate use
even more questionable than when it was first suggested.
Table 3. POHC Ranking With 02 Present3
US Incineration
UDRI Low 02 (1.5%) High 02 (5%)
Tetrachloroethylene
Toluene
Pentachlorobenzene
Chlorobenzene
1
2
5
3
4
1
4
2
5
3
2
3
1
5
4
al being most stable; 5 being least stable.
12
-------
RELATIVE POHC STABILITY
SF6
6-9'S POHC DRE
+ TCE * TOL
T
E
M
P
E
R
A
T
U
R
E
C
1,**UU
1,300-
1,200-
1,100-
1,000-
900-
onn -
#
*.-•'
..**
. • *
.•••''!**
» \Lf f^ ^^
•' ** i.-f
^#*** . ++
\L>H|u^ J»
^pr^ 4-
+
++
+
oi
4-+ °D°°°0D
+. '-'DDDDODDDDODDDDD^'
1 23456789 10
O2 CONCENTRATION
CLBZ
Figure 4. Relative POHC stability.
13
-------
CONCLUSIONS
o SFg is a limited surrogate. Data showing reactivity with refractory
and difficulty with using it may have reduced its apparent value
o POHC ranking is not absolute and depends on combustion conditions
o Toluene is the most stable of the organics at elevated temperature
and 02 levels, while SFg is most stable at intermediate and lower
levels.
RECOMMENDATIONS
o Modify the use of surrogates for trial burns, and specify
minimum combustion conditions
o Choose a minimum number of POHCs for trial burns
o Use the most stable POHCs at low oxygen values, since that condition
i.s present in most incinerator failures
o Develop method for continuous monitoring of toluene and benzene,
for performance monitoring.
14
-------
REFERENCES
1. Oppelt, E.T. Incineration of Hazardous Waste - A Critical Review. JAPCA.
37: 558, 1987.
2. 40 CFR Part 264.343.
3. Tsang, W.M. and Shaub, W.M. Surrogates as Substitutes for Principal
Organic Hazardous Constituent Validation of Incinerator Operation. In:
Proceedings of the Second Conference on Municipal, Hazardous and CoaT~
Wastes, Miami, FL 1983. p. 241.
4. England, W.G., Rappolt, T.J., Teuscher, L.H., Kerrin, S.L. and Mournighan,
R.E. Measurement of Hazardous Waste Incineration Destruction and Removal
Efficiencies Using Sulfur Hexafluoride as a Chemical Surrogate. In:
Proceedings of the 79th Annual APCA Meeting, 1986. 106:162097W.
5. Pandompatam, B., Liem, A.J., Frenette, R. and Wilson, M.A. Effect of
Refractory on the Thermal Stability of SF6. JAPCA. 39: 310, 1989.
6. Dellinger, B. Testing and Evaluation of a POHC/PIC Incinerability
Mixture. U.S. EPA 15th Annual Research Symposium, Cincinnati, Ohio, April
April 10-12, 1989.
7. Taylor, P.H. and Chadbourne, J. Sulfur Hexafluoride as a Surrogate.
JAPCA. 37: 729 1987.
8. 40 CFR 761.70a.
9. Waterland, I.E. and Staley, L.J. Pilot Scale Listing of SFr As A
Hazardous Waste Incinerator Surrogate. To be Presented at the 82nd
National Meeting of the Air and Waste Management Association, Dallas, TX,
1989. Paper 89-23B.4.
15
------- |