Paper for presentation at 2005 Scientific Conference on Chemical and Biological Defense Research,
Timonium, MD, November 14-16, 2005.

THERMAL DESTRUCTION OF CB CONTAMINANTS BOUND ON BUILDING MATERIALS:

EXPERIMENTS AND MODELING

P. Lemieux, J. Wood, C. Lee, S. Serre
U.S. Environmental Protection Agency
Office of Research and Development
Research Triangle Park, NC 27711

M. Denison, M. Bockelie, A. Sarofim, J. Wendt
Reaction Engineering International
Salt Lake City, UT 84101

ABSTRACT

An experimental and theoretical program has been initiated by the U.S. EPA to investigate issues of
chemical/biological agent destruction in incineration systems when the agent in question is bound on
common porous building interior materials. This program includes 3-dimensional computational fluid
dynamics modeling with matrix-bound agent destruction kinetics, bench-scale experiments to determine
agent destruction kinetics while bound on various matrices, and pilot-scale experiments to scale-up the
bench-scale experiments to a more practical scale. Finally, model predictions are made to predict agent
destruction and combustion conditions in two full-scale incineration systems that are typical of modern
combustor design.

INTRODUCTION

After a building has gone through decontamination activities following a terrorist attack with
chemical warfare (CW) agents, biological warfare (BW) agents, or toxic industrial chemicals (TICs),
there will be a significant amount of residual material and waste to be disposed. This material is termed
"building decontamination residue" (BDR). Although it is likely that the BDR to be disposed of will
have already been decontaminated, the possibility exists for trace levels of the toxic contaminants to be
present in absorbent and/or porous material such as carpet, fabric, ceiling tiles, office partitions, furniture,
and personal protective equipment (PPE) and other materials used during cleanup activities. It is likely
that much of this material will be disposed of in high-temperature thermal incineration facilities, such as
medical/pathological waste incinerators, municipal waste combustors, and hazardous waste combustors.

Although pathogens such as Bacillus anthracis (anthrax) present in BDR are killed at typical
incineration temperatures (> 800 °C), gas-phase residence times (> 2 s), and solid-phase residence times
(>30 min), it is possible for some of the pathogens to escape the incinerator due to bypassing the flame
zones, cold spots, and incomplete penetration of heat through the bed. In the early 1990's, EPA performed
testing of commercial hospital waste incinerators1 by introducing large quantities of Geobacillus
stearothermophilus (an anthrax surrogate) spores into the combustors and measuring the number leaving
in the stack emissions and in the incinerator bottom ash, in terms of Log reduction in spore concentration.
It was found that, in certain cases, only a 3-Log reduction in spore destruction was found, in spite of
acceptably high operating temperatures and sufficiently long residence times.

As a result of the 2001 anthrax attacks, the EPA instituted an experimental and theoretical research
program to investigate issues related to the thermal destruction of contaminated BDR2 initially including
carpeting, ceiling tile, and wallboard. Tests are being performed at bench- and pilot-scale, and a

1


-------
computer model is being developed to predict BDR-bound agent behavior in two commercial incinerator
designs. Contaminants that are being tested will include BW simulants (Geobacillus stearothermophilus)
and CW simulants (dimethyl methylphosphonate). These tests would examine time/temperature
requirements for agent destruction, issues related to facility compliance with relevant permits (e.g.,
emissions of nitrogen oxides), predicting potential operational difficulties in full-scale systems, and
understanding which facilities may or may not be appropriate to handle certain types of BDR.

Thermal processing of the BDR material in commercial incinerators is evaluated using an
incinerator-modeling tool provided by Reaction Engineering International (REI). The models provided to
the EPA are derived from models developed through funding from a DoD-Army Small Business
Innovative Research (SBIR) project in which REI created a simulator to predict the performance and
destruction of chemical agent in the US Army baseline Chemical Weapons Incineration facilities3. In that
work the simulator included a range of models, from time-dependent process models to detailed
Computational Fluid Dynamic (CFD) models. Using computational chemistry methods, detailed chemical
kinetic mechanisms were developed that describe the incineration of mustard blister agent and the nerve
agents GB and VX. The kinetic mechanisms describe the complete decomposition of the agent, including
products of incomplete destruction and final product speciation. Although reliable data are not available
for agent destruction, the kinetic mechanisms have been benchmarked against high quality experimental
data for similar substances (where available) and peer reviewed by an Advisory Panel of experts4. Agent
destruction in the furnaces is computed using finite-rate chemistry CFD models of the incinerator units.
The incinerator simulator was benchmarked to data from the Johnston Atoll Chemical Disposal Facility
(now de-commissioned). The simulator has been used to address a variety of operational questions for the
US Army and its system contractors3'5. The modeling techniques, combustion models and agent
destruction models from the Army funded project were re-configured to the specific incinerator units of
interest to the EPA, aspects of which have been benchmarked to data from the EPA rotary kiln incinerator
simulator (RKIS) (see below)6'7.

Although the methodology described in this paper will be use for both chemical and biological
contaminants, the experiments completed so far have been limited to biologicals. Bench-scale
experiments on CW agent simulants have been completed, but no pilot-scale experiments on CW agent
simulants have been completed, and because of that no model runs have been performed to compare to
the pilot-scale results.

APPROACH

The approach taken to develop the computer simulations for the current project will be as follows:

•	Bench-scale experiments were performed to investigate matrix effects on BW agent simulant
destruction at various temperatures

•	Bench-scale experiments wee performed to investigate matrix effects on CW agent simulant
adsorption/desorption at various temperatures

•	Pilot-scale experiments were performed to investigate operational issues, permit compliance
issues (e.g., air pollutant emissions), and to calibrate model scale-up efforts

•	Computer simulations will be conducted. These simulations will be based on bench- and pilot-
scale experimental data and CW agent combustion kinetic mechanisms already developed by REI
to predict agent behavior in commercial incinerators, including a medical/pathological waste
incinerator and a commercial hazardous waste-burning rotary kiln

The three combustion systems that are modeled include:

•	The EPA RKIS, located in RTP, NC - See Figure 1

•	A commercial dual-chambered modular medical pathological waste incinerator - See Figure 2

2


-------
• A commercial hazardous waste-burning rotary kiln - See Figure 3

The first unit, the EPA RKIS facility, is a simulated pilot-scale rotary kiln incinerator located at the
EPA research facilities in RTP, NC. It has a primary and secondary burner, each rated at 73 kW (250,000
Btu/hr). The RKIS can feed a variety of liquid and solid surrogate hazardous waste, and is described in
detail elsewhere \

Secondary Combustion Chamber	Afterburner

Main
Burner

Figure 1. EPA Rotary Kiln Incinerator Simulator (RKIS).

Continuous
Emissions
Monitors

Ram
Feed

The second combustion unit is a commercial dual-chamber starved-air modular medical/pathological
waste incinerator that is currently being operated jointly by the EPA and National Institutes of
Environmental Health Sciences (NIEHS) on their RTP, NC campus. This facility has a nominal firing rate
of 1 MW (3.4 MMBtu/hr) and is capable of processing about 400 kg/hr (900 lb/hr) of wastes which
mostly consist of animal bedding.

Exhaust gas (to waste heat
recovery and/or air pollution
control system)

Figure 2. Medical/Pathological Waste Incinerator

The third combustion unit is a commercial hazardous waste-burning rotary kiln system currently in
operation in East Liverpool, OH. This unit has a nominal firing rate of 35 MW (120 MMBtu/hr) and
processes about 8100 kg/hr (9 tons/hr) of hazardous waste from various sources.

3


-------
Liquid was
and/or
auxiliary fi

Solid
waste

Figure 3. Commercial Hazardous Waste Burning Rotary Kiln.

The geometries, operating parameters, and sample emissions data for the three combustors were
acquired from the operators of the facilities. Computational grids and model input parameters were
generated by REI based on this information. Currently only the BW simulant experiments have
progressed to the point of being able to compare experimental and model results. Bench-scale CW
simulant adsorption/desorption experiments have been performed, but the modeling and pilot-scale
experiments have not been completed yet.

Results have been obtained using bulk building materials consisting of carpet, ceiling tile, and
wallboard. Ceiling tiles that had been in use for approximately 1 year in a laboratory were made available
due to construction changes. These were Class A, standard-white, fire-retardant, textured-faced ceiling
tiles composed of wood fiber (0 - 60%) and fibrous glass (0 - 13%). New drywall was used for these
tests, which consisted of gypsum core wrapped with a paper lining. Carpet was nylon 6-6 carpeting
acquired directly from Shaw, the manufacturer.

For the bench-scale tests, the building materials were cut into sample sizes measuring 7.62 x 3.81
cm, weighed, individually wrapped in aluminum foil and steam sterilized by autoclaving. The sterile
samples were inoculated with either 1.0 ml of Bacillus subtilis spores for a final concentration of 10s
spores/ml or 1.0 ml of Geobacilhis stearothermophihis spores for a final concentration of 106 spores/ml.
Once inoculated the building material pieces were placed on a sterile tray and allowed to dry overnight
within a biological safety cabinet. Once dried, the samples were prepared by stacking two (7.62 x 3.81
cm) pieces with the inoculated sides facing each other and then tested in thermal destruction experiments.
In the thermal destruction experiments, the BDR samples were heated in a quartz reactor operating at
various temperatures, for various time intervals. The samples were recovered and analyzed for viable
spores. These tests are described in more detail elsewhere9. Figure 4 shows a sample set of results
illustrating the destruction of Bacillus subtilis inoculated onto ceiling tile.

It must be noted that the bench-scale tests on carpet had not been performed at the time this paper
was written. Because of the behavior of carpeting at elevated temperatures (melting, pyrolysis), the
carpet tests were put off to the end for the bench-scale experiments. However, for the very same reasons,
the carpet was tested first in the pilot-scale experiments - the carpet combusted completely leaving little
to no ash behind in the RKIS. For this reason, the bench-scale tests discuss only ceiling tiles, and the
pilot-scale tests and the modeling only discuss carpeting.

EXPERIMENTAL RESULTS

4


-------
10 15 20 25
Heating Time (minutes)

30

35

Figure 4. Sample Bench-Scale Results - Effect of Heating Temperature and Time on Reduction of

Bacillus subtilis Spiked on Ceiling Tile

Once the destruction studies were completed, the destruction of the spores bound on the building
materials (shown in Figure 4) were fitted to an Arrhenius-type expression analogous to a zero-order
reaction:

sporei.ving -> sporedead	( 1)

The results from this curve fit are shown in Figure 5, both in terms of the Arrhenius fit and the more
conventional Z value type approach.

149 °C

Data

204 °C

Data

260 °C

Data

316 °C

Data

149 °C

Z value

204 °C

Z value

260 °C

Z value

316 °C

Z value

149 °C

Arrhenius

204 °C

Arrheriius

260 °C

Arrhenius

316 °C

Arrhenius

4000

Figure 5. Log reduction versus time for B. subtilis on ceiling tile at various initial temperatures.

Tests were also performed on the pilot-scale RKIS facility, where biological indicator (BI) strips
containing 10" spores of Geobacillus stearothermophilus were enclosed in a sealed 0.95 cm (3/8 in) OD
steel tube with a thermocouple and embedded into a 454 g (1 lb) bundle of wetted carpeting. The carpet
bundles were fed manually into the RKIS operating at 850 °C (1562 °F), removed from the RKIS at
different times and quenched with water. The BI strips were recovered and viable spores were measured.
Figure 6 illustrates the temperature profiles and results from the tests that are described in more detail

5


-------
elsewhere10. Note that the profile for Run 4 was slightly different than the other runs because more water
was absorbed on the carpet, so the time for complete drying was longer. It was found that total spore kill
was achieved between 4 and 5 minutes.

300

250

~ 200
u

a>

| 150
v

Q.

E
v

H 100
50

0

012345678

Time (min)

Figure 6. Internal Carpet Bundle Temperature vs. Time
MODEL RESULTS

The periodic loading of materials into the RKIS results in an inherently time-dependent operation
that must be adequately captured. Modeling the operation with a true transient computational fluid
dynamics (CFD) simulation would require excessive computing resources. To represent the time-
dependent nature of the RKIS in a computationally efficient manner, it is modeled using a combination of
a fast running, transient process model and a steady-state, reacting CFD model. The transient process
model and the CFD model contain accurate representations of the geometry details contained within the
RKIS (primary chamber, secondary chamber, connecting ductwork). The transient process model captures
the transient effects of building matrix combustion, on the chamber temperature, overall gas composition
and heat-up of the matrix pieces being processed. In addition, the transient model can mimic the use of
burner turndown, water (quench) sprays and variable air injection to maintain setpoints (e.g.,
thermocouple temperature, exit 02 concentration) used by the control system for the RKIS. The
conditions predicted by the transient model are subsequently used to define the boundary conditions used
by a steady-state, reacting 3D CFD model which computes the local mixing and CW/BW agent
destruction for a prescribed instant in time. The CFD models include the detailed chemistry and physics
required to analyze the RKIS. These models include the full coupling of turbulent fluid mechanics, all
modes of heat transfer (including radiation) and equilibrium combustion chemistry for agent and fuel.
After establishing the combustion flow field, a post-process calculation is used to evaluate CB agent
destruction.

The model of the RKIS is described in detail elsewhere6. Figure 7 shows example model outputs
compared to RKIS experimental results for the kiln exit temperature and the secondary combustion
chamber (SCC) exit oxygen (02) concentration, using carpet bundles as the model input feed. The model
does a reasonable job predicting the overall transient behavior while the RKIS is burning carpeting.

These example plots show kiln exit temperature, however, many output parameters are available from the
model results, including various temperatures, velocities, and species concentrations within the kiln.

6


-------
2000
u. 1900

| 1800

| 1700

m 1600

1500

Figure 7. Model Predictions vs. RKIS Measured Parameters.

The model was then used to predict BW agent destruction from spores bound on a carpet bundle.
Because the bench-scale experiment for the carpet samples was yet to be completed, the BW agent
destruction kinetics bound on ceiling tile were used for these preliminary runs. Figure 8 shows the model
predictions for spore destruction, using both the Arrhenius-type fit and the Z value type fit. This
prediction is in terms of the total number of initial spores that were converted to "dead" as per Equation 1.
Note that total spore destruction is predicted to occur between 4 and 5 minutes. This is entirely consistent
with the experimental results from the RKIS.

16.0
14.0

	Z value

Arrhenius

2.0

0.0	

0.0 1.0 2.0

Figure 8. Predicted Log Reduction of Spores

Figures 9 to 11 show visualizations of the gas temperature field from sample runs on the RKIS
model, the med/path incinerator model, and the hazardous waste burning rotary kiln models, respectively.
Data available for interrogation from the CFD model includes gas temperature, velocity, agent
concentration, combustion products (major and minor species), pressure as well as wall and equipment
surface temperatures and incident heat fluxes. These diagrams show temperature distributions of the
various combustors. Note the outlines of the burning carpet bundles in Figures 9 and 10.

12.0

c

I 10.0

8.0
6.0
4.0

3.0 4.0
Time, min

5.0

6.0

7


-------
Figure 9. A Sample Predicted Gas Temperature Field In The EPA RKIS.

Gas Temperature (K)

Figure 10. A Sample Predicted Gas Temperature Field In The Med/Path Incinerator Primary Chamber.

Gas Temperature (K)

¦ 2200

^ 300

f

I

Figure 11. A Sample Predicted Gas Temperature Field In The Hazardous Waste Kiln Primary Chamber.

CONCLUSIONS

A combined experimental and theoretical program has been initiated to understand and predict the
behavior of CB agents bound on building decontamination residue in thermal incineration systems.
Bench-scale experiments were used to develop kinetics of BW agent destruction. Pilot-scale experiments
were used to calibrate the models. Initial comparisons between model predictions and agent simulant
destruction show reasonable agreement.

Future work will include:

•	Pilot-scale CW agent simulant experiments and model comparisons

•	Inclusion of outdoor materials to the matrices being studied

•	Addition of stoker-fired incinerators to the modeling efforts

8


-------
• Full-scale testing to compare model results to the full-scale incinerators upon which the models
were based

REFERENCES

1	Wood, J.P.; Lemieux, P.M.; Lee, C.W., "Destruction Efficiency of Microbiological Organisms in
Medical Waste Incinerators: A Review of Available Data," International Conference on Incineration and
Thermal Treatment Technologies, Phoenix, AZ, May 10-14, 2004.

2	Lemieux, P. (2004), "EPA Safe Buildings Program: Update on Building Decontamination Waste
Disposal Area," EM, Vol. 29-33.

3	Bockelie, M.J., "Engineering Design Software for Military Incinerators", SBIR Phase II /II Plus
Final Report, Contract No. DAAD19-01-C0050, Program Manager Dr. Robert Shaw (ARO), December,
2004.

4	Montgomery, C. J., Denison, M. K., Bockelie, M. J., Sarofim, A. F., "Detailed Chemical Kinetic
Mechanisms for Decomposition and Oxidation of Chemical Warfare Agents and Application to CFD
Simulations," presented at the 2004 Scientific Conference for Chemical and Biological Defense Research,
Hunt Valley, MD, November 15-17, 2004.

5	Denison, M.K, Montgomery, C.J., Sadler, B.A., Sarofim, A.F., Bockelie, M.J., Gouldin, F., Magee,
R.S., Bozzelli, J.W., "Advanced Computer Simulations of Military Incinerators," Presented at the 24th
Army Sciences Conference, Orlando, Florida, November, 2004.

6	Denison, M.; Montgomery, C.; Zhao, W.; Bockelie, M.; Sarofim, A.; Lemieux, P., "Advanced
Modeling of Incineration of Building Decontamination Residue," Air and Waste Management
Association's 98th Annual Conference & Exhibition, Minneapolis, MN, June 21-24, 2005.

7	REI, "Configured Fireside Simulator User Guide - US EPA National Homeland Security Research
& Development Center, Modeling Support for Incineration of Building Materials", EPA Contract No:
4C-R010-NASX, June 30, 2005.

8	Lemieux, P.; Stewart, E.; Realff, M.; Mulholland, J.A. (2004), "Emissions Study of Co-firing
Waste Carpet in a Rotary Kiln," Journal of Environmental Management, Vol. 70, pp. 27-33.

9	Lee, C.W.; Wood, J.P.; Betancourt, D.; Linak, W.P.; Lemieux, P.M.; Novak, J., "Study of Thermal
Destruction of Surrogate Bio-contaminants Adsorbed on Building Materials," Air and Waste
Management Association's 98th Annual Conference & Exhibition, Minneapolis, MN, June 21-24, 2005.

10	Lemieux, P., "Pilot-Scale Combustion of Building Decontamination Residue," Air and Waste
Management Association's 98th Annual Conference & Exhibition, Minneapolis, MN, June 21-24, 2005.

9


-------