EPA/540/R-95/526
September 1995
EMERGING TECHNOLOGY REPORT:
DEVELOPMENT OF A PHOTOTHERMAL DETOXIFICATION UNIT
by
John L. Graham
Barry Dellinger
Joseph Swartzbaugh
Environmental Science and Engineering Group
University of Dayton Research Institute
Dayton, Ohio 45469-0132
Cooperative Agreement No. CR8 19594-01-0
Project Officer
Chien T. Chen
National Risk Management Research Laboratory
Water Supply and Water Resources Division
(formerly Superfund Technology Demonstration Division)
Edison, NJ 08837
This study was conducted
in cooperation with
U. S. Environmental Protection Agency
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
TECHNICAL REPORT DATA
(Please read liutrucnoni on the reverse beiare comntttinti
1 REPORT MO.
12,
3, RECIPIENT'S ACCESS1ON«NQ.
riTLE AND SUBTITLE
Emerging Technology Report; Development of a Photothertnal
Detoxification Unit
S, REPORT OATE
August 1995
6, PERFORMING ORGANIZATION CODE
7 AUTMQB(S)
John L. Graham, Barry Dellinger, John Swartzbaugh
8, PERFORMING ORGANIZATION REPORT NO.
9, PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Science and Engineering Group
University of Dayton Research Institute
300 College Park
Dayton, OH 45469-0132
110. PROGRAM ELEMENT NO,
11. CONTRACT/GRANT NO,
CR819594-01-0
12. SPONSORING AGENCY NAME AND ADDRESS
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13, TYPE OF REPORT AND PERIOD COVERED
Final Report 1.0/92 - 12/94
14, SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Chien T. Chen, Telephone No. (908) 906-6985
Technology, 15 (2), 1995
1 o. ABo T RACT
There has long been interest in utilizing photochemical methods for destroying hazardous organic materials.
Unfortunately, the direct application of classic, low temperature photochemical processes to hazardous waste
detoxification are often too slow to be practical for wide spread use. Furthermore, low-temperature photochemical
processes often fail to completely convert targeted wastes to mineral products of complete conversion which are either
harmless to the environment or easily scrubbed from the system effluent. Researchers at the University of Dayton
Research Institute (UDRI) have developed a unique photothermal process that overcomes many of the problems
previously encountered with direct photochemical detoxification techniques. Specifically, it has been found that there
are numerous advantages to conducting photochemical detoxification at relatively high temperatures. Under the
conditions of simultaneous exposure to heat and ultraviolet (UV) radiation the rate of destructive photothermal
reactions can be greatly increased and that these reactions result in the complete mineralization of the waste feed.
Furthermore, it has been demonstrated that at the elevated temperatures used In this process the efficiency of UV
radiation absorption also increases resulting in an overall improvement in process efficiency. These features (i.e., fast,
efficient, and complete destruction of organic wastes) makes this process a promising technique for destroying
hazardous organic wastes in the gas-phase. The authors present the theoretical foundation for the photothermal
detoxification process along with a summary of the results from a bench-scale flow reactor system. The basic design,
capital cost, and operating cost for a full-scale flow reactor system using currently available industrial illumination
equipment is also presented.
17,
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTlFIERS/QPEN ENDED TERMS
c, COSATI Field/Croup
Trichloroethylene, Tetrachloroethylene,
Benzene, Toluene, Ethylbenzene, Xylene,
Photothermal Reaction
11. DISTRIBUTION STATEMENT
Release to the public
19, SECURITY CLASS tTHu
unclassified
21. NO. OF PAGES
20. SECURITY CLASS
unclassified
22, PRICE
EPA Form 1120-1 «t-73)
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency through Cooperative Agreement No.: CR819594-01-0 to the
University of Dayton Research Institute. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or recommendation for use.
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the Agency
strives to formulate and implement actions leading to a compatible balance between human activities
and the ability of natural systems to support and nurture life. To meet this mandate, EPA's research
program is providing data and technical support for solving environmental problems today and
building a science knowledge base necessary to manage our ecological resources wisely, understand
how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for investigation
of technological and management approaches for reducing risks from threats to human health and the
environment. The focus of the Laboratory's research program is on methods for the prevention and
control of pollution to air, land, water and subsurface resources; protection of water quality in public
water systems; remediation of contaminated sites and ground water; and prevention and control of
indoor air pollution. The goal of this research effort is to catalyze development and implementation
of innovative, cost-effective environmental technologies; develop scientific and engineering
information needed by EPA to support regulatory and policy decisions; and provide technical support
and information transfer to ensure effective implementation of environmental regulations and
strategies.
This publication has been produced as part of the Laboratory's strategic long-term research
plan. It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
-------
ABSTRACT
There has long been interest in utilizing photochemical methods for destroying hazardous
organic materials. Unfortunately, the direct application of classic, low temperature photochemical
processes to hazardous waste detoxification are often too slow to be practical for wide spread
use. Furthermore, low-temperature photochemical processes often fail to completely convert the
targeted wastes to mineral products of complete conversion which are either harmless to the
environment or easily scrubbed from the system effluent. Researchers at the University of Dayton
Research Institute (UDRI) have developed a unique photothermal process that overcomes many
of the problems previously encountered with direct photochemical detoxification techniques.
Specifically, it has been found that there are numerous advantages to conducting photochemical
detoxification at relatively high temperatures. Under the conditions of simultaneous exposure to
heat and ultraviolet (UV) radiation the rate of destructive photothermal reactions can be greatly
increased and that these reactions result in the complete mineralization of the waste feed,
Furthermore, it has been demonstrated that at the elevated temperatures used in this process the
efficiency of UV radiation absorption also increases resulting in an overall improvement in process
efficiency. These features (i.e., fast, efficient, and complete destruction of organic wastes) makes
this process a promising technique for destroying hazardous organic wastes in the gas-phase. The
authors present the theoretical foundation for the photothermal detoxification process along with
a summary of the results from a bench-scale flow reactor system. The basic design, capital cost,
and operating cost for a full-scale flow reactor system using currently available industrial
illumination equipment is also presented.
This report was submitted in fulfillment of Cooperative Agreement CR819594-01-0 by the
University of Dayton Research Institute, Environmental Science and Engineering Croup, under
the (partial) sponsorship of the U. S. Environmental Protection Agency. This report covers a
period from 1 October 1992 to 30 December 1994, and work was completed as of 30 December
IV
-------
CONTENTS
FOREWORD iii
ABSTRACT »v
FIGURES v»
TABLES xvi
ABBREVIATIONS AND SYMBOLS xx
ACKNOWLEDGMENT xxii
1. INTRODUCTION l
1.1 System Description 1
1.2 Theory of Photothermal Detoxification 3
1.3 Experimental Design 5
1.4 Laboratory Systems 7
1.4,1 Trie High-Temperature Absorption Spectrophotometer 8
1.4.2 The Laboratory-Scale Photothermal Detoxification Unit 10
2 EXPERIMENTAL PROCEDURES AND QUALITY ASSURANCE/ 12
QUALITY CONTROL
2.1 High-Temperature Absorption Spectrophotometer 12
2.2 Laboratory Scale Photothermal Detoxification System 15
2.3 LS-PDU Carbon and Chlorine Balances 20
2.4 LS-PDU Illumination System 24
3 ABSORBANCE AS A FUNCTION OF TEMPERATURE 25
3,1 Hydrocarbons 27
3.2 Chlorinated Methanes 27
3.3 Chlorinated Alkenes 23
3.4 Aromatics and Arenes 32
3.5 Chlorinated Aromatics 32
3.6 Chlorinated Dibenzo-p-Dioxins 37
3.7 Gasoline 37
3,8 Summary 39
4 LABORATORY-SCALE PHOTOTHERMAL DETOXIFICATION 41
4.1 Alkanes and Chlorinated Alkanes 42
4.2 Chlorinated Alkenes 45
4,3 Aromatics and Arenes 48
4.4 Chlorinated Aromatics ^4
4.5 Chlorinated Dibenzo-p-dioxins 57
4.6 Gasoline ^°
-------
CONTENTS (continued)
4.7 Mixtures of TCE, DCBz, and Water Vapor 64
4.8 Benzene and Hydrogen Peroxide 72
4.9 Summary Remarks 76
5 PRODUCTS OF INCOMPLETE CONVERSION 77
5.1 Alkanes and Chlorinated Alkanes 77
5.2 Chlorinated Alkenes 80
5.3 Aromatics and Arenes 83
5.4 Chlorinated Aromatics 87
5.5 Chlorinated Dibenzo-p-dioxins 90
5.6 Mixtures of TCE, DCBz, and Water Vapor 90
5.7 Summary 86
6 BASIC DESIGN FOR THE PROTOTYPE PDU 96
6.1 Lamp Selection 96
6.2 Basic Reactor Vessel Design 99
6.3 Predicted PDU Reactor Performance 106
6.4 Estimated Cost of Prototype PDU System 113
7 CONCLUSIONS 116
8 REFERENCES 118
VI
-------
FIGURES
Number Page
1.1 Conceptual schematic of a prototype Photothermal Detoxification 2
Unit (PDU) showing the basic elements of the system including the
reactor vessel and illumination system.
1.2 General energy versus reaction coordinated diagram illustrating the 3
relative energy requirements for thermal (curve So) and photothermal
(curves Si and T,) decomposition processes.
1.3 General schematic of the High Temperature Absorption 8
Spectrophotometer (HTAS) showing the principal elements of this
system.
1.4 General schematic of the Laboratory Scale-Photothermal Detoxification Unit 10
(LS-PDU) showing the principal elements of this system.
2.1 GC/FID calibration curve for trichloroethylene illustrating the 19
linearity of the quantitative transport through the LS-PDU.
2.2 GC/TCD calibration curve for CCb using the discontinuous technique 21
of collecting the LS-PDU effluent in a Tedlar sample bag for analysis
with a stand-alone gas chromatograph.
2.3 GC/TCD calibration curve for CO using the discontinuous technique 22
of collecting the LS-PDU effluent in a Tedlar sample bag for analysis
with a stand-alone gas chromatograph.
3.1 The spectral irradiance spectrum for the LS-PDU's xenon arc lamp showing 26
the spectral distribution of radiation impinging on the reactor with an overall
intensity of 18.1 W/cm2.
3.2 The gas-phase absorption spectra for chloroform at200,300,400, 28
and 500°C.
3.3 The gas-phase absorption spectra for carbon tetrachloride at 100,200,300, 28
400,500, and 600°C.
VII
-------
FIGURES (continued)
Page
3.4 Summary of the photon absorption rate constants for chloroform and carbon 29
tetrachloride showing the increase in the strength of absorption with
temperature and chlorine substitution,
3.5 The gas-phase absorption spectra for trichloroethylene at 100, 200, 300, 30
400, 500, and 600°C.
3.6 The gas-phase absorption spectra for tetrachloroethylene at 200,300, 31
400,500, and 600°C
3,7 Summary of the photon absorption rate constants for trichloro- 31
ethylene and tetrachloroethylene showing the increase in the strength
of absorption with temperature and chlorine substitution.
3.8 The gas-phase absorption spectra for benzene at 100, 200, 300, 400,500, 33
and 600°C
3,9 The gas-phase absorption spectra for toluene at 200, 300,400,500 and 33
60G°C
3.10 The gas-phase absorption spectra for ethyl benzene at 200,300, 400,500 34
and 600°C,
3.11 The gas-phase absorption spectra for m-xylene at 200,300,400, 500 and 34
600°C.
3,12 Summary of the photon absorption rate constants for benzene, toluene, ethyl 35
benzene, and rn-xylene showing the increase in the strength of absorption
with temperature and the extent of alkane substitution,
3.13 The gas-phase absorption spectra for monochlorobenzene at 200,300,400, 35
500 and 600°C.
3.14 The gas-phase absorption spectra for o-dichiorobenzene at 200, 300,400, 36
500 and 600°C
3.15 Summary of the photon absorption rate constants for monochloro-benzene 36
and o-dichlorobenzene showing the increase in the strength of absorption
with temperature and chlorine substitution.
VIM
-------
FIGURES (continued)
Number
3.16 The gas-phase absorption spectra for 1,2,3,4-tetrachlorodibenzo-p-dioxin at 38
300,400, and 500°C.
3.17 Summary of the photon absorption rate constants for 1,2,3,4-tetrachloro- 38
dibenzo-p-dioxin showing the increase in the strength of absorption with
temperature.
3.18 The gas-phase absorption spectra for gasoline using an assigned mean 39
molecular weight of 140 g/g-mol and a liquid phase density of 0.75 g/ml at
200,300,400,500, and 600°C.
4.1 Summary of thermal and photothermal data for chloroform exposed to 0 and 44
17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.2 Summary of thermal and photothermal data for trichloroethylene exposed to 46
0 and 18.1 W/cm2 of xenon arc radiation for 10 sec in air.
4.3 Summary of thermal and photothermal data for tetrachloroethylene exposed 47
to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.4 Summary of the photothermal quantum yields for trichloroethylene and 47
tetrachl oroethy 1 ene.
4.5 Summary of thermal and photothermal data for the benzene component of 51
BTEX exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.6 Summary of thermal and photothermal data for the toluene component of 51
BTEX exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.7 Summary of thermal and photothermal data for the ethyl benzene component 52
of BTEX exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in
air.
4.8 Summary of thermal and photothermal data for the m-xylene component of 52
BTEX exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.9 Summary of the photothermal quantum yields for components of BTEX. 53
4.10 Summary of thermal and photothermal data for monochlorobenzene exposed 55
to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
IX
-------
FIGURES (continued)
Number Page
4.11 Summary of thermal and photothermal data for o-dichlorobenzene exposed 56
to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.12 Summary of the photothermal quantum yields for monochlorobenzene and 56
o-dichlorobenzene.
4.13 Summary of thermal and photothermal data forl,2,3,4-tetrachlorodibenzo-p- 58
dioxin exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.14 Summary of the photothermal quantum yields for 1,2,3,4-tetrachlorodibenzo- 59
p-dioxin.
4.15 Example GC/FID chromatogram of gasoline exposed to non-destructive 60
conditions (300°C for 10 sec in air) illustrating the complex nature of this
sample.
4.16 Overall decomposition (sum of all integrated GC/FID peaks) for gasoline 61
exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.17 Summary of the number of GC/FID peaks observed from the analysis of 62
gasoline exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.18 Summary of the decomposition of the 2-methyl butane (identification 63
assigned by GC/MS spectral library) component of gasoline exposed to 0 and
17.6 W/cm2 of xenon arc radiation for 10 sec in air.
4.19 Summary of the decomposition of the toluene (identification assigned by 63
GC/MS spectral library) component of gasoline exposed to 0 and 17.6 W/cm2
of xenon arc radiation for 10 sec in air.
4.20 Summary of thermal and photothermal data for the trichloroethylene 66
component of a TCE:Water: DCBz (60:20:1) mixture exposed to 0 and 17.6
W/cm2 of xenon arc radiation for 10 sec in air.
4.21 Summary of thermal and photothermal data for the o-dichlorobenzene 66
component of a TCE:Water:DCBz (60:20:1) mixture exposed to 0 and 17.6
W/cm2 of xenon arc radiation for 10 sec in air.
-------
FIGURES (continued)
Number Page
4.22 Summary of the photothermal quantum yields for the trichloroethylene and o- 67
dichlorobenzene components of a TCE:Water:DCBz (60:20:1) mixture.
4.23 Summary of thermal and photothermal data for the trichloroethylene 69
component of a TCE:Water:DCBz (1:20:1) mixture exposed to 0 and 17.6
W/cm2 of xenon arc radiation for 10 sec in air.
4.24 Summary of thermal and photothermal data for the o-dichlorobenzene 70
component of a TCE:Water:DCBz (1:20:1) mixture exposed to 0 and 17.6
W/cm2 of xenon arc radiation for 10 sec in air.
4.25 Summary of the photothermal quantum yields for the trichloroethylene and o- 70
dichlorobenzene components of a TCE:Water:DCBz (1:20:1) mixture.
4.26 Summary of thermal and photothermal data for the benzene component of a 73
Ez'.HiOi- Water (1:3:13) mixture exposed to 0 and 17.6 W/cm2 of xenon arc
radiation for 10 sec in air.
4.27 Summary of the photothermal quantum yields for the benzene component of a 74
Bz:H202:Water (1:3:13) mixture.
4.28 Absorption spectrum for hydrogen peroxide in water (referenced against 74
water) at 20°C illustrating that this compound is a very weak absorber of UV
radiation.
5.1 Summary of LS-PDU data for chloroform and it's major PICs exposed for 10 79
sec in air.
5.2 Summary of LS-PDU data for chloroform and its major PICs exposed to 17.6 80
W/cm2 of xenon arc radiation for 10 sec in air.
5.3 Summary of LS-PDU data for trichloroethylene and its major PIC exposed for 82
10 sec in air.
5.4 Summary of LS-PDU data for trichloroethylene and its major PICs exposed to 83
18.1 W/cm2 of xenon arc radiation for 10 sec in air.
XI
-------
FIGURES (continued)
Number Page
5.5 Example GC/FID chromatograms from BTEX exposed to 300,600, and 84
700°C for 10 sec in air showing that benzene is the most stable component. 1)
Benzene, 2) Toluene, 3) Ethyl Benzene, 4) m-Xylene, 5) PIC PI, 6) PICP2,
7) PIC P3, 8) PIC P4.
5.6 Example GC/FID chromatograms from BTEX exposed to 300,600, and 85
700°C and 17.6 W/cm2 for 10 set in air showing that benzene is the most
stable component. 1) Benzene, 2) Toluene, 3) Ethyl Benzene, 4) m-Xylene, 5)
PIC PI, 6) PIC P2, 7) PIC P3, 8) PIC P4.
5.7 Summary of LS-PDU data for the benzene component of BTEX and the major 86
PICs when BTEX is exposed for 10 sec in air.
5.8 Summary of LS-PDU data for the benzene component of BTEX and the major 87
PICs when BTEX is exposed to 17.6 W/cm- for 10 sec in air.
5.9 Summary of LS-PDU data for o-dichlorobenzene and its major PICs exposed 89
for 10 sec in air showing the relatively small yield of the organic products.
5.10 Summary of LS-PDU data for o-dichlorobenzene and its major PICs exposed 89
to 18.1 W/cm2 for 10 sec in air showing the relatively small yield of the
organic products.
5.11 Example GC/FID chromatograms from TCDD exposed to 300 thermal, 600 91
thermal, and 600°C photothermal (17.6 W/cm2) for 10 sec in air showing that
the photothermal process destroys the complex mixture of PICs as well as the
parent compound.
5.12 Summary of LS-PDU data for the trichloroethylene and o-dichlorobenzene 93
components of a TCE/DCBz/water mixture exposed for 10 sec in air showing
the relative yield of carbon tetrachloride, the only major organic product from
this mixture.
XII
-------
FIGURES (continued)
Number
5.13 Summary of LS-PDU data for the trichloroethylene and o-dichlorobenzene 94
components of a TCE/DCBz/water mixture exposed to 17.6 W/cm2 of xenon
arc radiation for 10 sec in air taking the initial concentration of TCE as the
basis for comparison showing the relative yield of carbon tetrachloride, the
only major organic product from this mixture.
6.1 Comparison of the radiant intensity as a function of wavelength for 1.0 W/cm2 98
of high pressure xenon and medium pressure mercury arc radiation illustrating
that the radiation from the latter is concentrated in several high intensity bands
in the near-UV.
6.2 Radiant intensity as a function of radial and axial position from a 200 W/in. 100
medium pressure mercury arc lamp with a nominal arc length of 200 cm (79
in.) illustrating that the radiant intensity is nearly constant over the length of
the lamp and decreases rapidly off the ends of the arc.
6.3 Mean radiant intensity as a function of vessel radius for a PDU chamber 101
measuring 250 cm long enclosing a single 200 W/in. medium pressure mercury
arc lamp with a nominal 200 cm arc showing that the mean radiant intensity
decreases as the chamber radius increases.
6.4 Mean radiant intensity as a function of the number of lamps for a PDU 103
chamber measuring 250 cm long and 120 cm in diameter enclosing 200 W/in.
medium pressure mercury arc lamps with a nominal 200 cm arc mounted on a
30 cm radius from the vessel axis showing that the mean radiant intensity
increases approximately linearly with the number of lamps.
6.5 Mean radiant intensity as a function of the radius on which the lamps are 104
mounted for a PDU chamber measuring 250 cm long and 120 cm in diameter
enclosing four 200 W/in. medium pressure mercury arc lamps with a nominal
200 cm arc showing that the mean radiant intensity decreases as the lamps are
mounted closer to the vessel walls which are assumed to be non-reflective.
6.6 Basic design for a prototype PDU chamber based on the laboratory tests with 105
the LS-PDU and assuming the use of six 200 W/in. medium pressure mercury
arc lamps with a nominal arc length of 200 cm.
Xlll
-------
FIGURES (continued)
Number Page
6.7 Predicted time to achieve 99% destruction of trichloroethylene using a series 108
of PDU chambers as illustrated in Figure 6.6 operated in series assuming each
chamber may be modeled as 2 CSTRs in series.
6.8 The predicted processing capacity of a PDU achieving 99% destruction of 109
trichloroethylene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases rapidly with temperature, and nearly linearly with the number of
chambers.
6.9 The predicted processing capacity of a PDU achieving 99% destruction of 109
tetrachloroethylene as a function of mean operating temperature and number
of chambers using the basic design illustrated in Figure 6.6 showing the
capacity increases significantly with temperature, and nearly linearly with the
number of chambers.
6.10 The predicted processing capacity of a PDU achieving 99% destruction of 110
benzene in BTEX as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above SOOT, and nearly linearly with
the number of chambers.
6.11 The predicted processing capacity of a PDU achieving 99% destruction of 110
toluene in BTEX as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above SOOT, and nearly linearly with
the number of chambers.
6.12 The predicted processing capacity of a PDU achieving 99% destruction of 111
ethyl benzene in BTEX as a function of mean operating temperature and
number of chambers using the basic design illustrated in Figure 6.6 showing
the capacity increases significantly at temperatures above SOOT, and nearly
linearly with the number of chambers.
6.13 The predicted processing capacity of a PDU achieving 99% destruction of m- 111
xylene in BTEX as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above SOOT, and nearly linearly with
the number of chambers.
XIV
-------
FIGURES (continued)
Number Page
6.14 The predicted processing capacity of a PDU achieving 99% destruction of 112
monochlorobenzene as a function of mean operating temperature and number
of chambers using the basic design illustrated in Figure 6.6 showing the
capacity increases significantly at temperatures above 400°C, and nearly
linearly with the number of chambers.
6.15 The predicted processing capacity of a PDU achieving 99% destruction of o- 112
dichlorobenzene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above 500°C, and nearly linearly with
the number of chambers.
6.16 The predicted processing capacity of a PDU achieving 99% destruction of 113
1,2,3,4-tetrachlorodibenzo-p-dioxin as a function of mean operating
temperature and number of chambers using the basic design illustrated in
Figure 6.6 showing the relatively high system capacity for this compound and
that the processing rate increases nearly linearly with the number of chambers.
XV
-------
TABLES
Number Page
2.1 Summary of Sample Purity as Specified by the Suppliers 14
2.2 Carbon Balances from Thermal Exposures at 700°C for 10 sec in Air 23
2.3 Carbon Balances from Photothermal Exposures with 17.6 W/cm2 23
Xenon Arc Radiation at 700°C for 10 sec in Air
3.1 Oscillator Strengths (X>230 nm) Relative to Benzene at 100°C 25
3.2 Photon Absorption Rate Constants with 18.1 W/cnv* Xenon Arc 26
Radiation
4.1 Summary of LS-PDU Exposure Conditions for Samples Analyzed as 42
Pure Compounds
4.2 Summary of LS-PDU Exposure Conditions for Samples Analyzed as 43
Mixtures
4.3 Summary of LS-PDU Results for Chloroform Exposed 10 sec in Air 44
to 0 and 17.6 W/cm2 Xenon Arc Radiation
4.4 Summary of LS-PDU Results for Trichloroethylene Exposed 10 sec 46
in Air to 0 and 18.1 W/cm2 Xenon Arc Radiation
4.5 Summary of LS-PDU Results for Tetrachloroethylene Exposed 10 46
sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
4.6 Summary of LS-PDU Results for the Benzene Component of BTEX 49
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
4.7 Summary of LS-PDU Results for the Toluene Component of BTEX 50
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
4.8 Summary of LS-PDU Results for the Ethyl Benzene Component of 50
BTEX Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
xvi
-------
TABLES (continued)
Number Page
4.9 Summary of LS-PDU Results for the m-Xylene Component of 50
BTEX Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
4.10 Summary of LS-PDU Results for Monochlorobenzene Exposed 10 54
sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
4.11 Summary of LS-PDU Results for o-Dichlorobenzene Exposed 10 55
sec in Air to 0 and 18.1 W/cm2 Xenon Arc Radiation
4.12 Summary of LS-PDU Results for 1,2,3,4-Tetrachlorodibenzo-p- 58
dioxin Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
4.13 Summary of LS-PDU Results for Gasoline Exposed 10 sec in Air to 61
0 and 17.6 W/cm2 Xenon Arc Radiation
4.14 Summary of Contaminants at an Example Site in Western Nevada 65
4.15 Summary of LS-PDU Results for the Trichloroethylene Component 65
of Mix #1 Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
4.16 Summary of LS-PDU Results for the o-Dichlorobenzene 65
Component of Mix #1 Exposed 10 sec in Air to 0 and 17.6 W/cm2
Xenon Arc Radiation
4.17 Summary of LS-PDU Results for the Trichloroethylene Component 68
of Mix #2 Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
4.18 Summary of LS-PDU Results for the o-Dichlorobenzene 69
Component of Mix #2 Exposed 10 sec in Air to 0 and 17.6 W/cm2
Xenon Arc Radiation
4.19 Summary of LS-PDU Results for the Trichloroethylene Component 71
of Mix #3 Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc
Radiation
XVII
-------
TABLES (continued)
Number Page
4.20 Summary of LS-PDU Results for the o-Dichlorobenzene 71
Component of Mix #3 Exposed 10 sec in Air to 0 and 17.6 W/cm2
Xenon Arc Radiation
4.21 Summary of LS-PDU Results for Benzene in the Presence of 72
Hydrogen Peroxide Exposed 10 sec in Air to 0 and 17.6W/cm2
Xenon Arc Radiation
4.22 Summary of LS-PDU Results for Benzene in the Absence of 73
Hydrogen Peroxide Exposed 10 sec in Air to 0 and 17.6W/cm2
Xenon Arc Radiation
4.23 Summary of Photothermal Quantum Yields for Samples Tested as 75
Pure Compounds
4.24 Summary of Photothermal Quantum Yields for Samples Tested as 75
Mixtures
4.25 Summary of Pseudo First-Order Thermal Oxidation Kinetic 76
Parameters Measured for the Test Compounds Used in This Project
5.1 Summary of LS-PDU Data For Chloroform Exposed For 10 sec in 78
Air
5.2 Summary of LS-PDU Data for Chloroform Exposed to 17.6 W/cm2 78
Xenon Arc Radiation For 10 sec in Air
5.3 Summary of LS-PDU Data for Trichloroethylene Exposed for 10 81
sec inAir1
5.4 Summary of LS-PDU Data for Trichloroethylene Exposed to 18.1 W/cm2 82
Xenon Arc Radiation for 10 sec in Air
5.5 Summary of LS-PDU Data for o-Dichlorobenzene Exposed for 10 sec in Air 88
5.6 Summary of LS-PDU Data For o-Dichlorobenzene Exposed to 18.1 88
W/cm2 Xenon Arc Radiation for 10 sec in Air
XVIII
-------
TABLES (continued)
Number Page
5.7 Summary of LS-PDU Data for TCE/DCBz/Water Mixture #1 92
Exposed for 10 sec in Air
5.8 Summary of LS-PDU Data for TCE/DCBz/Water Mixture #1 92
Exposed to 17.6 W/cm2 Xenon Arc Radiation for 10 sec in Air
6.1 Photon Absorption Rate Constants Using 1 W/cm2 Medium Pressure 98
Mercury Arc Illumination
6.2 Photon Absorption Rate Constants Using 1 W/cm2 Xenon Arc 99
Illumination
6.3 Estimated Costs for A PDU Chamber Fitted with Six 200 W/in 114
Lamps
XIX
-------
ABBREVIATIONS AND SYMBOLS
°C
°K
A
AMU
BTEX
Bz
C
cal
cfm
dfrm
an
Ctet
d
DCBz
Ea
exi
ESE
EthBz
exp(n)
F
no
fr
-------
ABBREVIATIONS AND SYMBOLS (continued)
Definition
ml Milliliter
Hi Microliter
mm Millimeter
mol Mole
MS Mass Spectrometer
N Number of tanks
N1ST National Institute of Standards and Technology
nm Nanometer
OMA Optical Multichannel Analyzer
p Power
pe Power at a volume element
PCB Polychlorinated Biphenyl
PCDD Polychlorinated Dibenzo-p-dioxin
PCDF Polychlorinated Dibenzo-p-furan
PCE Teuachloroethylene
PDU Photothermal Detoxification Unit
PIC Product of Incomplete Conversion
Pn Unidentified organic product n
PNA Polynuclear Aromatic
ppm Part Per Million
psig Pounds per Square Inch Gauge
R Universal Gas Constant
R Radius
RSD Relative Standard Deviation
D Summation
scfm Standard Cubic Feet per Minute (dry @ 20°C)
sec Second
Sn Singlet state n
SVE Soil Vapor Extraction
T Temperature
TPDU Temperature of the PDU
Tref Reference Temperature
t Time
•r
Mean residence time
t.
199 Time required for 99% conversion
TCDD 1,2,3,4-Tetrachlorodibenzo-p-dioxin
TCE Trichloroethylene
Tn Triplet state n
To! Toluene
UDRI University of Dayton Research Institute
UV Ultraviolet
V Volume
W Watts
Xyl m-Xylene
XXI
-------
ACKNOWLEDGMENT
We would like to gratefully acknowledge the United States Environmental Protection
Agency for funding this project under Cooperative Agreement No: CR819594-01-0. We would
also like to express our appreciation to Norma M. Lewis, Chief, Emerging Technology Section,
for her support and encouragement. We would also like to acknowledge the Project Officer for
this program, Dr. Chien T. Chen, for his assistance throughout this project. And finally, we would
like to thank the reviewers who did an excellent job helping us prepare this final report
XXII
-------
SECTION 1
INTRODUCTION
The University of Dayton Research Institute's (UDRI) Environmental Science and
Engineering Group (ESE) has developed a new process based on a photochemical technology
that is well suited for treating the dilute gas-phase waste streams that are associated with several
types of Superfund site cleanup operations (e.g., soil vapor extraction, thermal desorption, etc.)
and may also find other applications such as a retrofit on conventional hazardous waste
incinerators to assist in the destruction process or even as a primary destruction technology in
special cases such as mixed wastes. Specifically, previous work on the use of concentrated solar
energy to destroy hazardous organic wastes has demonstrated that a photothermal process (a
high-temperature photochemical process) can destroy toxic organic materials far more efficiently
and cleanly than conventional methods such as thermal oxidation. [1,2] Given the limited
quantity of short-wavelength ultraviolet (UV) radiation in sunlight as compared to artificial
sources,[3,4] it was felt that the results of the solar research only hinted at the potential for a
photothermal treatment process. Therefore, the solar-based photothermal technology was
generalized to include the use of nearly any source of thermal and W energy to make the
concept of a Photothermal Detoxification Unit (PDU) available for Superfund site remediation.
In this specific application, the PDU will serve as an off-gas cleanup device used in conjunction
with various conventional soil and debris decontamination systems.
1.1 SYSTEM DESCRIPTION
The PDU is conceptually a relatively simple device that can be included as a new unit
operation on existing treatment technologies requiring little or no change in the overall design of
these systems. Specifically, the PDU could conceivably be designed to be retrofitted to existing
gas handling systems with the subsequent elimination of activated carbon filters or other
treatment devices such as condensers and separators. This eliminates the capital and operating
-------
costs of this equipme nt and the transport and disposal of the collected wastes. Indeed, it is
possible that the inclusion of a PDU in the design of a treatment system may result in significant
savings by eliminating storage transportation, and treatment costs.
Thermally Insulated
Reaction Vessel
Mounting
Flange x^
\
External Lamp
Assembly (3)
Gas Inlet
tilet I
Preheater
\
Exhaust
S upport/Transporta tion
Palette
Sampling Ports (4)
Figure i, 1. Conceptual schematic of a prototype Photothermal Detoxification Unit (PDU)
showing the basic elements of the system including the reactor vessel and
illumination system.
The basic concept of the PDU is fairly simple; a flowing gas stream is heated to a
relatively high temperatur (>200°C) and exposed to intense UV radiation for a duration long
enough to destroy the hazardous organic components. While the exact configuration of the PDU
may take many forms depending on the specific application of a given unit, a general schematic
of a conceptual prototype unit is shown in Figure 1.1. In this Figure, the PDU is shown as a
thermally insulated, cylindrical vessel with high intensity UV lamps mounted around the
perimeter so that the interior of the reactor can be evenly illuminated. Regardless of the final
form of the PDU, all designs share the same central concept; exposing a flowing gas-phase
stream to a relatively high temperature and intense UV radiation long enough to achieve an
acceptably complete mineralization (i.e., decomposition to mineral products of complete
conversion such as water, carbon dioxide, hydrogen chloride, etc.) of the hazardous organic
components. It is the problem of determining the necessary conditions of time, temperature, and
-------
radiant intensity as well as the optimization of these parameters that present a unique design
challenge.
1.2 THEORY OF PHOTOTHERMAL DETOXIFICATION
Various photophysical and photochemical processes that may take place in a
photochemical reactor operating at a relatively high temperature can be described by the energy
versus reaction coordinate diagram presented in Figure 1.2. This figure illustrates the potential
energy surfaces for the thermochemical and photochemical reaction pathways available in this
E?
o
c
LU
Reaction Coordinate —>
Figure 1.2, General energy versus reaction coordinated diagram illustrating the relative energy
requirements for thermal (curve So) and photothermal (curves S, and T,)
decomposition processes.
In Figure 1.2 the hazardous organic molecules present in a process stream are considered
to be initially near the lowest vibrational energy state of the ground electronic state (i.e., near the
-------
bottom of curve SQ in Figure 1.2). The combustion literature demonstrates that molecules on the
ground state surface require approximately 30-100 kcal/mol of thermal energy to initiate
oxidation reactions,[5] It is this relatively high activation energy barrier that makes it necessary
for purely thermal processors (i.e., incinerators) to operate at such high temperatures, often in
excess of 1,000°C. In the conventional thermal treatment of hazardous wastes there is little that
can be done to improve the efficiency of the fundamental thermal process. Specifically, exposure
time, temperature, and excess air are the only process variables that can be readily changed in an
effort to achieve more efficient destruction.
When the option of introducing intense UV radiation to the system is considered, we find
that we are no longer restricted to ground state, thermal chemistry. If organic molecules are
exposed to UV radiation of an appropriate wavelength, they can be promoted to electronically
excited states as shown by the curve S\. Molecules absorbing a photon of UV radiation store
that energy by promoting an electron to a higher energy state. This new electronic configuration
tends to weaken the molecular bonds, thereby reducing the energy required to induce chemical
reactions. Hence smaller activation energy barriers to destructive reactions are the result (cf.
Figure 1.2).
The excited state most readily accessible from the ground state is the lowest energy
singlet state, labeled Sj in Figure 1.2, in which the spin of the promoted electron is preserved.
When using sunlight as the radiation source, which has a UV cutoff of about 300 nm, only S j is
accessible in most hazardous waste molecules. Literature values suggest the activation energy
from this state is typically on the order of 2-10 kcal/mol,[6] so molecules in Si should react far
more rapidly than those in the parent ground state at a given temperature. Furthermore, if UV
radiation sources are used which generate shorter wavelength UV photons, higher energy single'
states (§2, SB, etc.) may be reached. In these higher energy states, activation energy decreases
still further, and the states may even be dissociative, requiring no additional thermal energy for
reactions to proceed.
One consequence of the preservation of electron spin in excited singlet states is that the
deactivation of the state through rapid processes, such as re-emission of the energy as
-------
fluorescence, is readily allowed. Therefore, the lifetime of these states can be quite short, often
on the order of nanoseconds. By necessity, then, only very fast reactions can occur from these
states before they return to the ground state. However, the spin of the excited electron may
change and the molecule converts to what is referred to as triplet state, shown as TI in Figure
1.2. Like their associated singlet states, the activation energies from triplet states is quite small,
often less than 5 kcal/mol. However, unlike molecules in excited singlet states, molecules in
excited triplet states relax through much slower processes, such as phosphorescence so there is
far greater opportunity for reactions to occur from excited triplet states than excited singlet
states, interestingly, one of the dominant promoters of conversion of a singlet to a triplet state is
the presence of heteroatoms such as chlorine, bromine, sulfur, nitrogen, and oxygen. [6]
Therefore, molecules that tend to be of environmental concern (chlorinated solvents, pesticides,
PCBs, PCDDs, PCDFs, etc.) also have the tendency to reside in excited triplet states upon
irradiation and should be particularly susceptible to photothermal destruction.
In addition to the radiative relaxation processes described above (fluorescence and
phosphorescence) molecules in either excited singlet or triplet states can relax back to the ground
state through radiationless paths These pathways may take several forms and depend upon the
particular molecule involved. Consequently, the relative rate and efficiency of these processes
cannot be generalized. However, the possibility that a molecule will absorb UV radiation and
still remain refractory must be considered. The existence of these rapid radiationless processes
further exacerbates the need far high-temperature to accelerate the photochemical reactions to
the point that they successfully compete with deactivation.
1.3 EXPERIMENTAL DESIGN
The photothermal detoxification process described above and illustrated in Figure 1.2
may also be described mathematically as; [I]
fr = exp[-(kgnd + fckabJtr J ( l ' l )
where fr is the fraction of reactant(s) remaining, kgn(j is the rate constant of ground state
(purely thermal) oxidation (sec"1), r is the photochemical quantum yield (mol/Einstein), k8t> is
the UV photon absorption rate constant (sec*1), and tr is the mean residence time in the reactor
-------
(sec). Equation 1.1 represents the global plug flow reactor performance model for the
photothermal detoxification process In order to use Equation 1,1 to predict the performance of
a prototype PDU, and hence aid in the design of the prototype unit, we must have knowledge of
typical values of kgnr, and kat>. Therefore, the experimental portion of this project will
concentrate on obtaining these values for selected example compounds.
It has been shown that the rate of thermal oxidation of many organic compounds can be
adequately described by simple pseudo-first-order kinetic models,[5} Specifically,
fr = exp(- kgnd tr) (1.2)
so,
kgnd = ln(l/fr)/tr (1.3)
Therefore, k^ may be found by measuring the fraction remaining following a purely
thermal exposure of known duration Furthermore, the temperature dependence of kgnd usually
follows the Arrhenius, thermal activation expression;
kgnd = AexpC-Ea/RT) (1.4)
or,
ln(k|nd) = In(A)-Ea/RT (1.5)
where A is the frequency factor (sec"1), Ea is the molar activation energy (cal mol"1). R is the
universal gas constant (1.987 cal mol-1 "K"1), and T is temperature (°K), Therefore, by
measuring the reaction rate (Equation 1.3) at a minimum of two temperatures, Ea and A can be
calculated from Equation 1.5. Once these values are known, kgnd may be calculated at any
temperature using Equation 1.4.
In a fashion similar to the measurement of the rate of purely thermal (i.e., ground state)
oxidation, has been shown that the photochemical quantum yield may be found from; [1]
tr (1.6)
-------
where fr(0) is the fraction remaining following a purely thermal exposure, fr(Io) is the fraction
remaining following an identical exposure but with the reactor being illuminated with radiant
intensity IQ (W cm*2). The rate coefficient of UV photon absorption, kat>» may be calculated as;
kab=1.92X10-*IejtiXjIoxi (1.7)
where exi is the molar extinction coefficient(L moHcnr1) of the waste at wavelength A., (nm),
and low is the radiant intensity (W cm'2) between XM and A.,. The summation is carried over the
wavelength region of overlap between the absorption spectrum of the waste and the emission
spectrum of the illumination source.
Reviewing Equations 1.1 through 1.7 shows that the important experimental values are
temperature, time, radiant intensity, spectral distribution, molar extinction, and fraction
remaining. It should also be noted that these models were developed for use in oxidizing
systems (though not necessarily limited to oxidation) so initial concentrations must be such that
excess air is available for complete oxidation. Therefore, the overall experimental plan is to
measure the molar extinction spectra (the molar extinction as a function of wavelength) of
individual selected organic compounds and measure the fraction remaining following carefully
controlled laboratory experiments, and then use this data to obtain the fundamental parameters of
Ea, A, kab» and §r. Once this information is available for several example compounds, the
Performance of a pilot scale PDU may be estimated thereby providing guidance on the system
specifications.
1.4 LABORATORY SYSTEMS
To conduct the experiments required to investigate high temperature photothermal
destruction of hazardous organic wastes, two dedicated instrumentation systems have been
constructed. Shown in Figures 1.3 and 1.4, respectively, these are the High-Temperature
Absorption Spectrophotometer (HTAS), and the Laboratory Scale-Photothermal Detoxification
Unit (LS-PDU). With these systems the most important aspects of photothermal destruction can
be studied under controlled, laboratory conditions. Specifically, the HTAS is used to directly
measure how strongly molecules absorb light by taking high-temperature, gas-phase absorption
spectra, and the LS-PDU is used for thermal and photothermal studies using various radiation
-------
sources such as simulated solar radiation, xenon, xenon/mercury, and mercury arc lamps, and
lasers.
1.4.1 The High-Temperature Absorption Spectrophotometer
The HTAS, shown in Figure 1.3, is a custom built, single-beam spectrophotometer
capable of operating at temperatures as high as 1,000°C Since the organic molecules for which
the HTAS was designed to study tend to decompose with prolonged exposures at elevated
temperatures, the system is fitted with a flow cell rather than a static cell found in most
commercial spectrophotometers. Furthermore, an inert carrier gas (i.e., nitrogen) is used for
sample transport to eliminate the possibility of sample oxidation. By flowing a carrier gas laden
with the sample of interest through the cell, the length of exposure to elevated temperatures can
be kept short (typically 1 second) to limit destruction of the sample at very high temperatures.
OMA Workstation
Focussing
Optics
ffi
Absorption coliimation
Furnace / Optics
OMA
Monochrometer Vent
1 I
y
Sample Inlets
Deuterium
Lamp
•Inlet Heaters
Figure 1.3, General schematic of the High Temperature Absorption Spectrophotometer
(HTAS) showing the principal elements of this system.
Referring to Figure 1.3, the absorption cell assembly consists of a thermally insulated
enclosure which houses a pair of sample inlet chambers and the absorption cell itself. The two
-------
heated inlets include a low volume inlet specifically designed for gasses and liquids, and a high
volume inlet which may be fitted with a variety of probe inserts for admitting gases, liquids,
solids, and even mixed phase samples. During normal operation carrier gas is admitted to the
system through both chambers to prevent back-diffusion into the unused channel. The flow
from the two inlets join together into a single transfer line which is fused to the sidewall of the
absorption cell, The cell is in the form of a cylinder, 1.2 cm in diameter by 20 cm long, and is
mounted along the centerline of a conduit which passes completely through the housing. Heat is
provided by half-cylinder ceramic heaters which are closed at either end with flat quartz
windows which prevent convection currents from passing through the heated zone. The exhaust
from the cell exits the housing through a heated transfer line, then passes through a paniculate
filter, an activated carbon filter, and finally to a fume hood. For the sake of chemical inertness,
structural integrity, and UV transparency, the entire flow system, from the inlets to the exhaust,
is fabricated from fused quartz. The cell is illuminated with collimated radiation from a
deuterium lamp and the light leaving the cell is dispersed with a 0.25 m monochrometer and
detected with a 512 channel optical multi-channel analyzer (OMA).
Operation of the HTAS is described in detail in Section 2. Briefly, molar extinction
spectra were obtained by comparing reference (system blank) and sample spectra taking into
account the cell length and the calculated molar concentration of sample in the cell.
1,4.2 The Laboratory Scale-Photothermal Detoxification Unit
The LS-PDU, shown in Figure 1.4, is a dedicated flow reactor system capable of
obtaining thermal and photothermal decomposition data on a great variety of compounds.
Structurally, the LS-PDU shares many features with the HTAS, For example, the LS-PDU
includes dual sample inlets connected to a cylindrical vessel through a single sample transfer
line. However, whereas the HTAS was designed to prevent reactions from taking place in the
system, the LS-PDU was designed to conduct reactions under carefully defined conditions and
analyze the products of those reactions. For this purpose the vessel connected to the inlet line is a
small cylindrical reactor measuring 1.2 cm in diameter by 8.4 cm long. The exhaust from the
reactor flows through a heated transfer line to a trapping system which collects all of the
condensable materials from the flowing gas, This trap is a single tube-in-shell design similar
-------
a laboratory condenser. The shell side is cooled with nitrogen gas which has in turn been cooled
by a liquid nitrogen bath. This allows the trap to operate at temperatures as low as -180°C,
though -160°C is routinely used. During the condensate collection phase of operation the exhaust
from the trap is vented to a fume hood. In preparation to analyze the collected condensate, this
vent is closed which directs the flow of gas to an inline analytical system consisting of a
programmed temperature, capillary column gas chromatograph (GC) fitted with dual columns
and an inlet splitter. One of the columns is connected to a scanning quadrapole mass
spectrometer (MS), the other to a hydrogen flame ionization detector (FID). The LS-PDU may
be used with nearly any UV radiation source and it is currently configured with a pulsed dye
laser, solar simulator, and a high pressure xenon arc lamp.
MSD Workstation
h,
H
H
HFID Workstation
Furnace Xenon Arc
„ . Lamp
Chilled N- \ Reactor
/
Inlet Heaters
MSD
Gas Chromatograph
Sample inlets
Figure 1.4. General schematic of the Laboratory Scale-Photothermal Detoxification Unit (LS-
PDU) showing the principal elements of this system.
The operation of the LS-PDU is described in detail in Section 2. Briefly, the fraction
remaining (the fraction of organic feed surviving the exposure in the reactor) is calculated by
comparing integrated chromatographic peak areas measured under destructive conditions with
values obtained from non-destructive tests (e.g., 300*C purely thermal). Quantitative organic
10
-------
product yields were calculated by comparing the integrated peaks with those from known
amounts of analytical standards. Qualitative organic product yields were calculated by
comparison with the initial peak area of the parent species from non-destructive tests.
11
-------
SECTION 2
EXPERIMENTAL PROCEDURES
AND QUALITY ASSURANCE/QUALITY CONTROL
Before a detailed design could be considered for the prototype PDU it was necessary to
obtain fundamental spectroscopic, thermochemical, and photochemical data on the types of
compounds which would be considered candidate for photothermal detoxification. This data
collection was one subject of this research project, and Section 6 presents the design for a
prototype PDU.
As described below, the spectroscopic information was collected using the HTAS, while
the LS-PDU was used to obtain the thermochemical and photochemical data. In addition to the
general operating procedures, the quality assurance and quality control procedures are described.
2.1 HIGH TEMPERATURE ABSORPTION SPECTROPHOTOMETER
Referring to the general schematic of the HTAS as shown in Figure 1.3, operation of this
system began with checking the gas supply pressure and making sure the supply valves were
fully open. Next, the deuterium lamp (Oriel Model 6316), data system (Apple He), and
computer interface (Tracer Northern Model 6100/6200) were switched on to allow them to
stabilize while the absorption cell system was set up for a specific analysis. Once the deuterium
lamp was started, the beam exiting the absorption cell assembly was inspected to ensure it was
properly centered on the entrance slit of the monochrometer (Jarrel-Ash Model 82-410).
With the subsystems on and stabilizing, the temperature of the inlet chamber, absorption
cell, and exhaust line were set as needed. These were typically already set at 100°C from
previous operations in what is referred to as an idle mode. The temperature sensors (type K
thermocouples) and readouts (dedicated thermocouple readers) were calibrated using a National
Institute of Standards and Technology (NIST) traceable thermocouple calibrator (Omega Model
CL23A). The temperature of the absorption cell was considered critical, so a calibration table
12
-------
was assembled for the cell's temperature controller. The measurement of the cell temperature
was repeated 5 times to ensure the precision (2% relative standard deviation) and accuracy (+/-
5°K absolute difference) was met. The inlet chamber and exhaust line temperatures were
considered noncriticai so the indicated temperature was used. The initial temperature was set to
100,200, or 300°C depending on the boiling point of the sample,
Once the system temperatures were set, the flow rate of gas through the absorption celi
was measured with a soap film flow meter (Alltech Model 4047). The operation and typical
accuracy and precision of this device has been described in the literature.[7] The indicated
volume of this device was calibrated by water displacement using volumetric flasks (0.05%
accuracy) and found to be within the accuracy of the measurement, therefore the indicated
volume was taken as the actual volume, The timer used to make the flow measurement (Fisher
Model 14-649-7) was checked against time signals from WWV, the NIST broadcast time
standard. The measurement of five replicates of 60 seconds each showed the timer to be accurate
to within 0.005 sec. and precise to within 0.084% relative standard deviation (RSD), The flow
through the test cell was measured 5 times to insure it was within the acceptance criteria of +/-
0.2 nil/sec and 5% RSD, A nominal flow rate of 12 ml/s at 20°C was set (with the actual value
being measured and recorded) to give a nominal mean residence time in the cell of 1 second at
300°C.
Along with the system flow rates, the ambient pressure was recorded using a vacuum-
over-mercury barometer (Sargcmt-Welch Model S4565). The pressure in the absorption cell was
taken as ambient.
With the flow of the carrier gas through the system set, the monochrometer wavelength
selector was set to the desired value. The wavelength indicator was calibrated using a mercury
vapor lamp and was found to be within the required acceptance criteria of 5% relative standard
deviation and +/-1 nm absolute difference.
With clean carrier gas (i.e., dry nitrogen) flowing through the cell a reference spectrum
was taken and stared on the data system's floppy disk. The test sample was then admitted to the
system at a controlled rate using a syringe pump (Sage Instruments Model 341B). The samples
used in this project were kept in a special storage area and restricted to use for this project. A
13
-------
manifest accompanied each sample to record when samples were drawn from the stock, how
much, and by whom. The purity of each sample as specified by the manufacturer is summarized
in Table 2.1. The flow rate of sample into the system was determined by measuring the time
required for the syringe plunger to admit a predetermined volume. This measurement was
repeated 5 times to insure the acceptance criteria of 5% RSD and +/-1 ppm absolute difference
in concentration was met. With sample flowing through the cell, a second spectrum is taken and
stored on disk.
Table 2.1
Summary Of Sample Purity As Specified By The Suppliers
Name Purity Supplier
Benzene 99,9+% Sigma-Aldrich
Carbon Tetrachloride 99,9+% Sigma-Aldrich
Chloroform 99,9% Sigma-Aldrich
o-Dichlorobenzene 99% Sigma-Aldrich
Ethyl Benzene 99+% Sigma-Aldrich
n-Hexane 99+% Sigma-Aldrich
Monoehlorobenzene 99.99% Sigma-Aldrich
1,2,3,4-Tetrachlorodibenzo-p-dioxin 98+% Ultra Scientific
Tetrachloroethylene 99.9+% Sigma-Aldrich
Toluene 99,8% Sigma-Aldrich
Trichloroethylene 99+% Sigma-Aldrich
2,2,4-Trimethylpentane 99% Sigma-Aldrich
m-Xylene 99+% Sigma-Aldrich
From the reference spectrum, the sample spectrum, and the calculated concentration of
sample in the cell, the molar extinction spectrum is determined by;
ex = logioOox/IO/OQ (2,1)
where EX is the molar extinction coefficient (L/mol-cm) at wavelength X (nm), IQX is the
reference intensity at X (detector counts), Ix is the measured intensity with sample flowing
14
-------
through the cell with a molar concentration of C (moI/L), and with a cell, path length of l(cm).
The molar concentration of sample in the cell is calculated from the measured flow rates of
sample and carrier gas (L/s), the measured molar flow rate of sample (mol/s), and the measured
cell temperature (to correct for the temperature difference between the point of measurement and
the ceil). The temperature of the flow meter was taken as ambient, a condition confirmed by
direct measurement" The overall procedure was checked using a NIST UV absorption standard
and procedure (Standard Reference Material 935a, potassium dichromate) and was found
to be within the acceptance criteria of 5% relative standard deviation and +/-0.05 absorbance
units,
The sequence described above was repeated for each wavelength region of interest and
for each sample. A complete spectrum was assembled by combining the data from across the
spectral region of interest,
2.2 LABORATORY SCALE-PHOTOTHERMAL DETOXIFICATION SYSTEM
Referring to the general schematic of the LS-PDU shown in Figure 1.4, operation of this
system began with checking the gas supply pressure and making sure the supply valves were
fully open, A nominally slow flow rate of dry nitrogen gas was set through the cold trap to make
sure this component was dry before the nitrogen gas chiller was filled with liquid nitrogen
coolant. Next, the GC/FID and GC/MS data systems were turned on and the GC/MS data system
configured to autotune the quadrupole mass spectrometer. Normally, the LS-PDU was left
pressurized with approximately 5 psig helium from previous runs in what is referred to as an idle
mode. If this was not the case, the system was flushed with helium, the system vent closed, and
the system pressure set to approximately 5 psig. With the LS-PDU pressurized with
approximately 5 psig of helium an auto-tune was performed according to the manufacturer's
specifications.
While the auto-tune procedure was executing, the ambient conditions (temperature and
pressure) were recorded on a daily log and a stock sample for the day was prepared. This stock
sample was prepared by injecting a predetermined volume of the liquid phase sample into a glass
sampling vessel which had been cleaned, evacuated, and filled with air from the same
15
-------
compressed air cylinder used for the LS-PDU reactor carrier gas. The internal volume of these
gas sampling vessels had been previously measured by water displacement and recorded on each
vessel. The sample injected into the glass vessel was allowed to evaporate resulting in a gas-
phase sample of known concentration.
While the sample was evaporating in the sampling bulb, the temperature of the inlet
chamber, exhaust line, and cold trap enclosure were set. These were typically initially set at
300°C from previous operations in an idle mode similar to that described for the HTAS. The
temperature of the cold trap enclosure (the temperature controlled compartment which houses
the cold trap) was nominally set at 350°C to prevent cold spots in the lines leading to and from
the trap. The temperature sensors (type K thermocouples) and readouts (dedicated thermocouple
readers) were calibrated using an NIST traceable thermocouple calibrator (Omega Model
CL23A). The temperature of the reactor vessel was considered critical, so a calibration table was
assembled to assist in setting the actual reactor temperature. Furthermore, the measurement of
the reactor temperature was repeated 5 times to ensure the precision (2% RSD) and accuracy (+/-
5°K absolute difference) was met. The inlet chamber, exhaust line, and cold trap enclosure
temperatures were considered noncritical so the indicated temperature was used. Since the LS-
PDU could not be operated at a temperature of less that 300°C due to the heat input from the
illumination system, the inlet, reactor, and exhaust line temperatures were set to this value. The
cold trap enclosure temperature was set slightly warmer to approximately 350°C to promote
rapid heating of the trap itself as described below.
At this point the results from the auto-tune procedure were available and the procedure
complete. The system exhaust vent was opened to depressurize the system and the inlet gasses
switched from helium to air. While the system was being purged with air the coolant reservoir
was filled with liquid nitrogen and the cold trap cooled to approximately 20°C to condition the
carrier gas for measuring the system flow rate.
With the system temperatures set, the flow rate of air through the reactor was set to give
a mean residence time of 10 sec. The flow rate of gas through the system was measured at the
system exhaust vent using a soap film flow meter (Alltech Model 4045). The indicated volume
of this device was measured by water displacement using volumetric flasks (0.05% accuracy)
16
-------
and found to be within the accuracy of the measurement, therefore the indicated volume was
taken as the actual volume. The timer used to make the flow measurement was checked against
time signals from WWV, the NIST broadcast time standard, and was found to be accurate to
within 0.005 sec and precise to within 0.084% RSD. The flow through the test cell was
measured 5 times to insure it was within the acceptance criteria of +/- 0.2 ml/s of the desired
value and 5% RSD.
If the current test was a photothermal run, the xenon arc lamp (Spectral Energy Model
SS-lOOOx) was ignited and the lamp current set to 40 amps. With a lamp current of 40 amps the
lamp potential was nominally 24 Volts, giving an overall lamp energy of 960 Watts. The
illumination system was allowed to stabilize for two minutes and the lamp current and voltage
checked to ensure that they were at expected values and stable.
The coolant level in the cold trap reservoir was checked and refilled with liquid nitrogen,
if necessary. The flow of nitrogen gas to the cold trap was then increased, reducing the
temperature to below -160°C.
Once the cold trap temperature was stable (approximately 2 minutes) a predetermined
volume of stock sample (typically 200 ^1) was injected into the LS-PDU at a controlled rate
using a syringe pump (Sage Instruments Model 341B). The flow rate of sample into the LS-PDU
was determined by measuring the time required for the syringe plunger to admit a predetermined
volume. This measurement was repeated 5 times to insure the acceptance criteria of 5% RSD
and +/-1 ppm absolute difference in concentration was met. As the sample was being admitted to
the reactor, the system conditions (temperatures, lamp current, lamp voltage, etc.) were
monitored to ensure they were stable.
After all the sample had been introduced to the system the syringe was removed and a
brief period of time allowed to pass (typically 2 minutes) to ensure all of the sample had passed
through the reactor and cold trap. If the current test was a photothermal run the lamp was
switched on at this time.
At this point the reactor phase of the run was complete and the analysis phase was ready
to begin. Specifically, the system gases were switched from air to helium in preparation of
17
-------
running the gas chromatograph. While the system purge was being conducted the GC/FID and
GC/MS data systems are configured to perform the analysis. Furthermore, the GC/MS
workstation was set to control the temperature pragram of the GC oven and the oven was cooled
to its initial temperature. Once the GC oven was stable at it's initial temperature the system
exhaust port was closed, the system pressurized to approximately 5 psig, the GC/FID and
GC/M S data acquisition programs placed in acquire mode, and the cold trap coolant turned off,
The cold trap would fhen rapidly heat to 350°C (approximately 3 minutes) releasing the
collected products into the GC for analysis,
The analytical subsystem of the LS-PDU was a Hewlett-Packard Model 5890 GC main
frame fitted with dual columns and detectors, and a heated split/splitless interface to the cold
trap. The GC/FID channel was fitted with a 328 (im x 15 m fused silica column coated with a
0.25 jim dimethylpoiysiloxarte stationary phase (J&W Scientific DB-1). Similarly, the GC/MS
channel was fitted with a 200 ^m x 20 m fused silica column also coated with a 0.40 \im
dimethylpolysiloxane stationary phase. Furthermore, the split/splitless interface was operated in
a split mode using a 320 \im x 1 m split restrictor. The temperature program consisted of a 2
minute bold at -80°C (time zero was taken as the moment the coolant to the cold trap was
switched off) followed by heating at 10°C/min to 0°C, then 15°C/min to 260°C. The temperature
was held at 260°C for 5 minutes or until all solute peaks were observed on the GC/MS trace
(which lagged, the GC/FID trace by about 3 minutes), which ever was longer. This combination
of GC columns and temperature program was found to be suitable for all the organic compounds
used in this program. The GC/MS channel used a Hewlett-Packard Model 5970A Mass Selective
Detector (MSD) operating in a scanning mode. The mass range was scanned from 35 to 350
atomic mass units (AMU),
Quantitative results for parent compounds were recorded as fraction remaining by
normalizing the integrated GC/FID peak areas with data obtained under non-destructive
conditions (i.e., 300°C thermal exposures). Peaks were identified by relative -tention time
(GC/FID and GC/MS) and mass spectral identification (GC/MS), In the case of parent species
identification was made using the GC/MS data system's mass spectral library. The identity of
products were assigned by a combination of the mass spectral library identification followed by
18
-------
confirmation using analytical standards. GC/FID response factors were determined by analyzing
known quantities of standards. Typical GC/FID response factors are shown in Figure 2.1. These
data, for trichloroethylene, illustrate that the LS-PDU was a very linear system with respect to
the quantitative transport of material.
200-
o
5
c
c
S»
CD
O
o
*c
150-
100-
50-
0
nMol TCE = 0,2397 + 0.00059582(Peak Area)
R= 0,99999
5 104 1 105 1.5 105 2 105 2.5 105 3 105 3.5 105
GC/FID Area, counts
Figure 2, i. GC/FID calibration curve for trichloroethylene illustrating the linearity of the
quantitative transport through the LS-PDU,
The overall procedure described above was conducted twice for each temperature and
exposure condition (thermal and photothermal) to ensure that the acceptance criteria of 5% RSD
and +/-0.01% absolute difference. The data for fraction remaining reported in the following
sections represent the average value of two runs.
19
-------
2.3 LS-PDU CARBON AND CHLORINE BALANCES
On the outset of this program it was anticipated that carbon and chlorine balances would
be conducted on the LS-PDU. Unfortunately, difficulties were encountered which prevented
conclusive results from being obtained from these efforts.
In the case of chlorine, it was assumed that the major products of HC1 (melting point
-114°C) and Ch (melting point -101°C) would be quantitatively captured by the LS-PDU's cold
trap and analyzed by the GC/MS channel (the GC/FID channel would not respond to inorganic
gases). In practice this proved to be correct. Indeed HC1 and Ch were the sole chlorinated
products observed in the GC/MS traces at high temperatures where the GC/FID traces showed
all organic species had been destroyed. Unfortunately, attempts to obtain GC/MS response
factors were unsuccessful, so the amount of HC1 and C\2 produced could not be quantified.
Specifically, it was observed that the sample handling equipment (e.g., syringe needles, sample
bulbs, and septa) were rapidly and heavily corroded while attempting to prepare analytical
standards of HC1 and C\i from bottled gasses and it was found that it was not possible to prepare
a reliable calibrant for either gas.
The difficulties with the carbon balance was of a different nature. In this case it was not
possible to use the LS-PDU's cold trapping system as it was not capable of quantitatively
capturing CO (melting point -205°C). The trap was easily capable of capturing COi (melting
point -56°C) and indeed a large COi peak was observed in all GC/MS analysis (the GC/FID
channel would not respond to CO or CCh). However, since the reaction atmosphere (dry air)
contained COi it was not possible to distinguish COi produced as a product versus C02 from the
atmosphere.
In an effort to overcome the problems associated with CO and COi analysis an
alternative technique was developed. Specifically, during tests in which analysis of CO and COi
were to be conducted the cold trap on the LS-PDU was maintained at 350°C instead of the
customary -160°C. In this condition materials would flow through the trap and exit the system
through the exhaust port. To collect these materials a 1 L Tedlar sample bag was attached to this
port during the sample introduction phase of a run conducted for the sole purpose of obtaining
20
-------
carbon balance data. Prior to each test the sample bag was evacuated and the sample collection
time was carefully controlled to ensure consistency in the collected sample volume. After the
sample collection was complete, the contents of the bag were analyzed for light gases using an
isothermal, packed column gas chromatograph fitted with a thennoconductivity detector (TCD).
O
o
2500-
2000-
1500-
1000-
500-
0
nMol C02 =-808.45 + 15.41 9{PeakArea)
R= 0.99714
60
I ' ' ' I ' ' ' i ' ' i i ' ' ' I ' ' ' I '
80 100 120 140 160 180
GC/TCD Response, counts
200
Figure 2.2. GC/TCD calibration curve for CO, using the discontinuous technique of collecting
the LS-PDU effluent in a Tedlar sample bag for analysis with a stand-alone gas
cliromatograph.
Prior to conducting the carbon balance tests, the sample collection and GC/TCD system
was calibrated by injecting known amounts of CO and COi into the LS-PDU and collecting
samples of the exhaust and analyzing them as described above. The results are summarized in
Figures 2.2 and 2.3 for CO and CCh, respectively (the volume of calibrant was measured under
ambient conditions of 20°C and 745 mm Hg pressure). Note that the calibration curve for COa
has a negative intercept with the ordinate. This occurs because air is being used as the carrier
gas in the LS-PDU which has approximately 330 ppm of CCK Therefore, CCh is present in the
21
-------
Also, the GC/FID traces for light organic compounds (i.e. see 2-methyl-propane, melting point -
160°C, in Figure 4.15) illustrate the ability of the LS-PDUto capture and analyze light organic
compounds. These results suggest that no significant organic species remain unaccounted for in
the effluent analysis.
Table 2,2
Carbon Balances From Thermal Exposures At 700°C For 10 sec In Air
Name Co1 CO2 CO22 %CO3 %CO23 %Total4
TCE
DCBz
BTEX
TCDD
390
10700
435
177
411
13200
397
199
191
2000
154
113
105
123
91.3
112
49.0
18.7
35.4
63.8
154
142
127
176
'Initial amount of organic carbon, units of 10'9 moles.
2Final amount of carbon observed as carbon monoxide or carbon dioxide, units of 10"^ moles.
3Mole percent carbon recovered as carbon monoxide or carbon dioxide.
4Total mole percent carbon recovered.
Table 2.3
Carbon Balances From Photothermal Exposures
With 17.6 W/cm2 Xenon Arc Radiation At 700°C For 10 sec In Air
Name Co1 , CO2 COi2 %CO3 %CO23 %Total4
TCE
DCBz
BTEX
TCDD
390
10700
435
111
392
14300
392
197
195
2100
100
77
101
134
90
111
50.0
19.6
23.0
43.5
151
154
113
154
'Initial amount of organic carbon, units of 1Q~9 moles,
2Final amount of carbon observed as carbon monoxide or carbon dioxide, units of 10~9 moles.
3Mole percent carbon recovered as carbon monoxide or carbon dioxide.
4Total mole percent carbon recovered.
23
-------
Note that the carbon balance results do not suggest a bias in the LS-PDU data for organic
species. The LS-PDU data for the conversion of parent species and the yield of organic products
is conducted using a closed continuous process for which there is no opportunity for material
loss. The likely source of error in the carbon balance tests is thought to be a combination of the
discontinuous nature of the analysis and the small samples used. The problems with both the
carbon and chlorine balances will be addressed in any future work by using a larger continuously
fed reactor system which will permit the use of established methods for carbon and chlorine
analysis,
2.4 LS-PDU ILLUMINATION SYSTEM
The radiant intensity impinging on the LS-PDU reactor vessel was measured by
removing the illumination system from the instrument and characterizing the output from this
unit. Specifically, a thermopile power meter (Scientech Model 362) was placed at an
appropriate position with respect to the lamp to measure the radiation distribution impinging on
the reactor vessel. The power meter was calibrated as per the manufacturer's specifications using
an NIST traceable current source and voltage meter. Similarly, the spectral distribution of the
illumination system was calibrated as per the manufacturer's procedure using a scanning
monochrometer (Spex Model 1702/04) calibrated with a mercury vapor lamp. Following these
measurements the illumination system was re-installed on the LS-PDU and aligned with the
reactor.
24
-------
SECTION 3
ABSORBANCE AS A FUNCTION OF TEMPERATURE
Reviewing the theoretical basis for the photothermal detoxification process
discussed in Section 1 illustrates that knowledge of the absorption spectra of typical
waste compounds is important to both the interpretation of data from laboratory scale
reactors and to predicting the performance of large scale systems. For this reason
spectroscopic data was obtained with the HTAS on a variety of test compounds ranging
from simple hydrocarbons to complex PNAs. These data consisted primarily of the molar
extinction spectrum measured at temperatures from 100 to 600°C. For the purposes of
comparing the intensity of the overall UV absorption of the test compounds the relative
oscillator strengths (the integrated molar extinction spectrum) were calculated at
wavelengths greater than 230 nm (the onset of the LS-PDU xenon arc illumination
system) and normalized by the value for benzene at 100°C.[8] These values are
summarized in Table 3.1. Furthermore, the photon absorption rate constants for 18.1
W/cm2 of xenon arc radiation were calculated as summarized in Table 3.2. The emission
spectrum for the LS-PDU's xenon arc source which was used in calculating the photon
absorption rate constants is shown in Figure 3.1.
Table 3.1
Oscillator Strengths (X>230 nm) Relative To Benzene At 100°C
Temp Clfrm Qet TCE PCE Bz MCBz DCBz Tol EthBz Xyl TCDD 89 Get
100*0 0.0713 3,80 LOO
200 0.0123 0.139
300 0.0314 0.246
400 0.0645 0,400
500 0.188 0,605
600 0.975
6,16
8.47
10.9
12.6
15.5
19.0
24.1
26.3
30.3
32.0
1.28
1.46
1.73
1.94
2.26
2.33
2.89
3.32
4.89
6.47
5.33
7.26
9.52
12.88
16.3
2.33
2.59
2.81
3.33
3.77
2.29
2.57
2.97
3.43
5.25
3.13
3.62
4.21
5.46
7.01
366
383
523
3.67
4.35
5.05
5.54
7.29
25
-------
Table 3,2
Photon Absorption Rate Constants With 18.1 W/cm2 Xenon Arc Radiation
Temp Clfrm Ctel TCE PCE Bz MCBz DCBz Tol EthBz Xylene TCDD
l,14e-i 1.76C-1
l,35e-l 2.04e-i l,56ei
i,56e-l 2.25e-I l,77ei
1.70e-l 2.68e-i 2,6iel
2.20e-2 3.02e-l 9.08e-l U6e-3 2.52e-l 5.00e-l i,93e-l 2.53e-l 3.17e-l
100°C
200
300
400
500
1.24e-4
4.20e4
9.88e-4
4.40e-3
9.20e-4
1.84e-3
3.71e-3
7,68e-3
1.23C-2
3.63e-2
7,72e-2
1.10e-l
1.64e-l
2.21e-l
3.28c-l
5.12e-l
6.00e-l
8,12e-l
3.81C-2
5,16e-2
6,28e-2
8.36e-2
9,48e-2
1,
1,
1,
2,
,24e-l
,49e-l
,64e-l
,14e-l
2.62e-l l,18e-l
2,
3,
4
.95e-l
,06e-l
.28e-l
1.36e-l
1.51c-l
1.77e-l
600
Key to abbreviations:
Clfrm Chloroform
Ctet Carbon tetrachloride
TCE Trichlorocthylene
PCE Telrachlorocthyknc
Bz Benzene
MCBz Monochlorobenzcnc
C>CBz o-DicWorobenzcnc
Tol Toluene
EthBz Ethyl Benzene
Xyl m-Xylene
TCDD 1,2,3,4-Tctrachlorodibenzo-p-dioxin
89 Oct 89 Octane Gasoline
E
c
-y
E
o
o
.5
£S
k»
1
"o
8L
CO
12-
10-
8-
6-
4-
2-
0-
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.1, The spectral irradiance spectrum for the LS-PDU"s xenon arc lamp showing
the spectral distribution of radiation impinging on the reactor with an
overall intensity of 18.1 W/cm2,
26
-------
3.1 HYDROCARBONS
The high temperature absorption spectra of n-hexane and iso-oclane were taken
from 100 to 600°C. These measurements showed that the test compounds were
completely transparent at wavelengths between 190 and 320 nm. This suggests that a
direct photothermal process using xenon or medium pressure mercury arc lamps may not
be effective in treating aliphatic hydrocarbon wastes, However, these types of materials
may prove treatable in the presence of other photosensitive wastes such as aromatic or
halogenated hydrocarbons or through the use of photo-initiators such as hydrogen
peroxide or ozone. Alternatively, other radiation sources (i.e. low pressure mercury, Xe2
excimer, etc.) which emit mare far UV radiation (wavelengths shorter than 190 nm) than
the xenon arc lamp used in the LS-PDU tests may prove effective.
3.2 CHLORINATED METHANES
The abso-tion spectrum of chloroform was obtained from 200 to 500°C (cf.
Figure 3.2) and of carbon tetrachloride from 100 to 600°C (cf. Figure 3,3). The photon
abso-tion rate data is summarized in Figure 3.4. The absorption spectra for these
compounds show the increase in absorption intensity and red-shift with increasing
temperature that is a common feature in this work, These dataillustrate that the
chlorinated methanes are very weak absorbers of UV radiation. As summarized in Table
3.1, the relative oscillator strength far chloroform ranged from 0.0123 at 200°C to 0.188
at 500°C, and for carbon tetrachloride from 0.0713 at 100°C to 0,975 at 600°C
Similarly, the photon abso-tion rate constants for chloroform increased from 1.24x10"4
ser1 at 200°C to 4,40xlO'3 sec4 at 500°C, and for carbon tetrachloride from 9.20xl(H
sec4 at 100°C to 2.2GxlO~2 sec4 at 600°C. Although the overall absorption is weak it is
clear that the chlorine content had a significant impact on the strength of the absorption
suggesting that larger chlorinated aliphatic compounds may be susceptible to the PDU
process. Furthermore, the strength of the abso-tion band far both compounds increases
with decreasing wavelength suggesting that these compounds may be effectively treated
with radiation sources that reach further into the UV than the xenon arc source used here,
27
-------
o
t
o
E
c
.Q
o
_c
+-*
X
LU
o
50-
40-
30-
20-
10
0
- 500°C
- 400°C
- 300°C
- 200°C
\
". \.
11« I' '"'•'» I ' « ' ' 11 ' * ' I * * * ' I ' * ' ' I " ' ' ' I ' " *
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.2, The gas-phase absorption spectra for chloroform at 200, 300,400, and
500°C.
150-
E
o
_
E
.2
...»
o
.E
"x
LJJ
j3
O
100-
50
......... 600°C
-- - - - 500°C
_ — . 300*0
. - . 200°C
"1 TT''|TTI^IIITT^' |"'|T|-|r-'r-t'-'- 'f | 1 I' 1"1 |' "fl t'-l1
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.3. The gas-phase absorption spectra for carbon tetrachloride at 100, 200, 300,
400, 500, and 600°C.
28
-------
JO
IS
0.025'
0.020
0.015
0.010
0,005-1
0.000 j in i i i t ri
0 100 200
Chloroform
Carbon Tetrachloride
TV i i i i i i } i i i i j i T
300 400 500 600
Temperature,°C
700
Figure 3.4, Summary of the photon absorption rate constants for chloroform and carbon
tetrachloride showing the increase in the strength of absorption with
temperature and chlorine substitution.
3.3 CHLORINATED ALKENES
The absorption spectra for trichloroethylene (100-600°C) and tetrachloroethylene
(200-600°C) are summarized in Figures 3.5 and 3.6, respectively. The photon absorption
rate constants are summarized in Figure 3.7. These data illustrate that the chloroalkenes
are much more efficient absorbers of UV radiation than the chloromethanes. Indeed, the
relative oscillator strengths given in Table 3.1 show that the chloroethenes are among the
strongest UV absorbers studied in this program. Specifically, the relative oscillator
strength for trichloroethylene (TCE) varied from 3.80 at 100°C to 15.5 at 600°C and
from 19.0 at 200°C to 32.0 at 600°C for tetrachloroethylene (PCE). The photon
absorption rate constant for TCE varied from 0.0363 sec4 at 100°C to 0.302 sec4 at
600°C and from 0.328 at 200°C to 0.908 at 600°C for PCE. Coupling of the non-
29
-------
bonding electrons on the chlorine atoms with the K*<-n transition shifts the absorption
towards the red as compared to hydrocarbons. This shifted predominantly n*<-n
absorption renders these molecules susceptible to the PDU process as will be shown in
Section 4. Unfortunately, the strength of the UV absorption rapidly fades at longer
wavelengths suggesting that lamps which generate radiation deeper into the UV than
standard xenon arc lamps may have a strong impact on the ability of the PDU to
successfully treat these types of wastes.
o
t
o
c
x
yj
i~
JS
o
600°C
... - 500°C
— ~3QO°C
100°C
1' ' I * « ' ' I ' '^^ 1 I' • I J • • • » I • « • • I • ' « • I ' • '
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3,5. The gas-phase absorption spectra for trichloroethylene at 100, 200, 300,
400, 500, and 600°C,
30
-------
2,500
E 2,000
o
15
E
c
,2
™.
o
c
1,500
x 1,000-1
til
o
600°C
- 500°C
_ . _ . . 4QO°C
500H
0
230 240 250 260 270 280 290 300 310
Temperature, °C
Figure 3.6. The gas-phase absorption spectra for tetrachloroethylene at 200,300, 400,
500, and 600°C.
1.00-
Cft
X
J€*
0.80-
0.60H
0,40-
0.20-
0.00-
Trichloroethylene
Tetrach loroethy le ne
100 200 300 400 500
Temperature,°C
T
600 700
Figure 3.7, Summary of the photon absorption rate constants for trichloroethylene and
tetrachloroethylene showing the increase in the strength of absorption with
temperature and chlorine substitution.
31
-------
3.4 AROMATICS AND ARENES
High temperature absorption spectra for benzene (100-60G°C), toluene (200-
600°C), ethyl benzene (200~600°C), and m-xyiene (200-600°C) are summarized in
Figures 3.8 through 3.11, respectively. The photon absorption rate constants are
summarized in Figure 3.12, The relative oscillator strength (see Table 3.1} for benzene
varied from 1,00 at 100°C (by definition) to 2.26 at 600°C, for toluene from 2,33 at
200°C to 3.77 at 600°C, for ethyl benzene from 2,29 at 200°C to 5.25 at 600°C, and for
m-xylene from 3.13 at 200°C to 7.01 at 600°C. The photon absorption rate constant (see
Table 3.2) for benzene varied from 0.0381 sec4 at 100°C to 0.116 sec4 at 600°C, for
toluene from 0,118 sec4 at 200°C to 0,193 sec4 at 600°C, for ethyl benzene from 0.114
sec4 at 200°C to 0.253 sec4 at 6QO°C, and for m-xylene from 0.176 sec4 at 200°C to
0.317 sec4 at 600°C. These data suggest the alkane substitution had a significant impact
on the strength of the UV absorption, though not as great as the effect from chlorine
substitution. Furthermore, multiple substitution on the benzene ring had a larger impact
on UV absorption than a single substitution of similar formula weight (e.g., ethyl
benzene versus m-xylene), In any event, these data suggest that the aromatic compounds
as a class should be good candidates for PDU treatment.
3.5 CHLORINATED AROMATICS
High temperature absorption spectra for monochlorobenzene (MCBz, 100-600°C)
and o-dichlorobenzene (DCBz, 100-600°C) are summarized in Figures 3.13 and 3.14,
respectively. The photon absorption rate constants are shown in Figure 3.15. The photon
absorption rate constants for benzene are included in Figure 3.15 for comparison. These
data illustrate that the chlorinated aromatic compounds as a class are relatively strong
absorbers of UV radiation, and that the strength of absorption rapidly increases with
increasing chlorine content. Specifically, the relative oscillator strength (ca. Table 3.1)
for MCBz varied from 2.33 at 200°C to 6.47 at 600°C, and for DCBz from 5.33 at 200°C
to 16.3 at 600°C. Similarly, the photon absorption rate constant (ca. Table 3.2) for
MCBz varied from 0.124 sec4 at 200°C to 0.252 sec4 at 600°C, and for DCBz from
0,262 sec4 at 200°C to 0.500 sec4 at 600°C These data illustrate that chlorine
32
-------
E
o
—
E
c
,2
**&£
o
.£
**3
X
LU
o
1,000-
800-
600
400-
200-
---- - - SOO'C
_ . _ . . 400°C
- 200°C
100°C
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.8. The gas-phase absorption spectra for benzene at 100,200, 300, 400, 500,
and 600°C.
E
o
»
O
E
o
*"*•-•
o
.£
'x
LU
jg
o
1,000-
800"
600H
400-
200-
- - - - - 500°C
. 300°C
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3,9, The gas-phase absorption spectra for toluene at 200, 300,400, 500 and
600°C,
33
-------
E
o
_
o
X
LU
ii
o
1,000-
800-
600-
400-
200
----- 600°C
500°C
400°C
0"'f t i i i "i i i-rr'i i i i i'|"i i » i ; t i iIPf
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.10, The gas-phase absorption spectra for ethyl benzene at 200, 300,400, 500
and 600°C.
1,000-
o
__
E
,o
o
'•*••*
X
LU
JS
o
800-
600-
400-
200-
0
1« «i«« >»r»« I^T^^
230 240 250 260 27v
Wavelength, nn>
Figure 3,11, The gas-phase absorption spectra for m-xylene at 200, 300, 400,500 and
600°C,
34
-------
E
o
"a
£
c
.2
"fS
o
.£•
"x
ULJ
%_
JJ
o
1-.000
800
600-
400-
200-
600°C
500SC
400°C
300°C
20CTC
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.12. Summary of the photon absorption rate constants for benzene, toluene, ethyl
benzene, and m-xylene showing the increase in the strength of absorption
with temperature and the extent of alkane substitution.
0.35'
0,30-
0.25-^
0.20-^
0.15^
0.10-:
0,05-
0.00-
-*-— Benzene
-•— Toluene
-*— Ethyl Benzene
-*--— m-Xylene
T '' > I "«"-» "i"""l |""t » l"T"|""t T'l "I-1]""! 1 FT'I I I" I "i |"i r-t-r
0 100 200 300 400 500 600 700
Temperature,°C
Figure 3,13. The gas-phase absorption spectra for monochlorobenzene at 200, 300, 400,
500 and 600°C.
35
-------
o
E
c
o
UJ
k_
"o
1,000-r
SOO-
600"
400
200
- - - • 600°C
500°C
0 "I I"""T I T { I T
230 240 250 260 270 280 290 300 310
Wavelength, nm
Figure 3.14. The gas-phase absorption spectra for o-dichlorobenzene at 200, 300,400,
500 and 600°C.
0,60
0.50-
0.40-
0.30
0.20-
0.10
~«— Benzene
HI—- Chiorobenzene
-*— 1,2-DichIorobenzene
0.00 ..... ...... i .......... i ....... i ......... i ........ ......... t
' ............ ........ ' ...... r"r ..... ''""
•r | i "!'•••> i "| i i"» i '"{ 'i i' i 'i 1' '
0 100 200 300 400 500 600 700
Temperature,°C
Figure 3.15, Summary of the photon absorption rate constants for monochlorobenzene
and o-dichiorobenzene showing the increase in the strength of absorption
with temperature and chlorine substitution.
36
-------
substitution on the benzene ring shifted the onset of the intense 82 absorption band (i.e.,
the strong absorption appearing at short wavelengths) to the red, a process further
enhanced by increasing temperature. The intensity of this absorption at short wavelengths
suggests once again that using illumination sources that provide radiation at wavelengths
shorter than pravided by standard xenon arc lamps may significantly increase the
potential performance of a PDU.
3.6 CHLORINATED DIBENZO-P-DIOXINS
The high temperature absorption spectra for 1,2,3,4-tetrachlorodibenzo-p-dioxin
(TCDD) from 300 to 500°C are summarized in Figure 3.16 with the photon absorption
rate constants shown in Figure 3.17, These Figures shows that TGDD is a very strong
absorber of UV radiation both in terms of absorption intensity and the breadth af the
wavelength region of absorption, which reaches to relatively long wavelengths. The
relative oscillator strength (ca. Table 3.1) varied from 366 at 300°C to 523 at 500°C
while the photon absorption rate constant increased from 15.6 sec1 at 300°C to 26,1 sec*1
at500°C. Comparing these values with those for the other test compounds shows that
TCDD is the mostintense absorber of UV radiation used in this program. The
combination of the intense 7t*<-n and 7t*<-n transitions with the photochemical lability
of carbon-chlorine bonds makes TCDD and related PNAs highly susceptible to the
photothermal process.
3.7 GASOLINE
Trie absorption spectra for a commercial 89 octane unleaded gasoline (200-
600°C) are shown in Figure 3,18. For the purposes of data reduction, this sample was
assigned a mean molecular weight of 140 g/g-mol and a mean liquid phase density of
0.75 g/ml. Surprisingly, this sample proved to be a moderately strong absorber of UV
radiation. As Figure 3.18 shows, this complex mixture exhibits an absorption spectrum
with a strong aromatic character, though the transition between the Si and 82 absorption
bands is poorly defined compared to that of the spectra for other aromatic
37
-------
50000-
E
9
"o
13
c
40000-
30000-
1 20000-
UJ
10000-
500°C
400°C
300°C
240
I • i • ,
280 300 320
Wavelength, nm
340
360
Figure 3,16. The gas-phase absorption spectra for 1,2,3,4-tetrachlorodibenzo-p-dioxin at
300, 400, and 500°C.
30-
25-
20-
15-
10-:
5-
0
0 100 200 300 400 500 600 700
Temperature, °C
Figure 3,17. Summary of the photon absorption rate constants for 1,2,3,4-
tetrachlorodibenzo-p-dioxin showing the increase in the strength of
absorption with temperature.
38
-------
E
o
i
o
E
c"
,2
.»w
O
.£
*x
LU
05
1
0)
cc
1,000-
800-
600"
400
200-
•. 600°C
- 500°C
- 400°C
- 300°C
0 ""{"""t l"1 1 |"I l""t""f } I I t'1"| | l-f"""!"
230 240 250
:iftw?
260 270 280
Wavelength, nm
290 300 310
Figure 3.18. The gas-phase absorption spectra for gasoline using an assigned mean
molecular weight of 140 g/g-mol and a liquid phase density of 0.75 g/rnl at
200, 300,400,500, and 600°C.
compounds. The strength of absorption suggests that gasoline may be a candidate for
PDU treatment.
3.8 SUMMARY
From the standpoint of absorption spectroscopy, several classes of compounds are
suggested as good candidates for photomermal detoxification. Specifically, aromatic and
arene compounds would generally be classified as good candidates based on their
moderately strong absorption of UV radiation; chlorinated alkenes and aromatic
compounds would be considered very good candidates; and, dioxins/furans and related
PNAs may be very susceptible to the photothermal process. In contrast, systems
composed exclusively of alkanes, even chlorinated alkanes, would likely pose a
significant challenge to the photothermal process. However, it is possible that radiation
sources with very short wavelength UV radiation (i.e., low-pressure mercury,
39
-------
excimer, etc.) may be able to address these materials. Finally, complex mixtures like
gasoline may be considered treatable based on their spectroscopy, but the final degree of
applicability may depend on which compounds in the mixture prove photo-active and
how they interact with the other mixture components.
40
-------
SECTION 4
LABORATORY-SCALE PHOTOTHERMAL DETOXIFICATION
The spectroscopic data discussed in Section 3 illustrated how the absorption of UV
radiation of different classes of compounds behave as a function of temperature and provided the
molar extinction information required as part of the PDU reactor performance model. In this
Section the photothermal decomposition resulting from the absorption of UV radiation is
discussed along with the thermal data needed to extract the fundamental kinetic and
photochemical parameters needed to complete the reactor performance model.
Thermal and photothermal decomposition data has been obtained on six pure compounds
and six mixtures of compounds with varying degrees of complexity. Specifically, data was
obtained using the LS-PDU on chloroform, trichloroethylene (TCE), tetrachloroethylene(PCE),
monochlorobenzene (MCBz), o-dichlorobenzene (DCBz), and 1,2,3,4-tetrachlorodibenzo-p-
dioxin (TCDD). With respect to mixtures, tests were run with one formulation of benzene,
toluene, ethyl benzene, and m-xylene (BTEX) and three variations of a mixture of TCE, DCBz,
and water vapor. Tests were also run with gasoline and a trial mixture of benzene, hydrogen
peroxide, and water vapor. The pure compounds were considered representative of chlorinated
alkanes (chloroform) which weakly absorb ultraviolet radiation (UV), chlorinated alkenes (TCE,
PCE) which absorb UV relatively well and represent a class of solvents commonly found in
Super-fund sites, chlorinated aromatic compounds (MCBz, DCBz), and chlorinated dibenzo-p-
dioxin compounds (TCDD) which are representative of the semi-volatile compounds which may
be encountered in some Superfund remediation sites. The mixtures presented an opportunity to
quickly assess the effectiveness of the photothermal process on several aromatics and arenes
(BTEX), simulate the principal threats from an actual Superfund site (TCE/DCBz mixtures) for
which data is available, and allow some very preliminary tests of a possible indirect
photothermal process through the use of a photo-initiator (hydrogen peroxide). The overall
41
-------
exposure conditions for each pure compound and mixture are summarized in Tables 4.1 and 4.2,
respectively. The fraction remaining (where fr(n) is the fraction remaining with n W/cm2
radiation) for each compound and mixture is summarized in Tables 4.3 through 4.22. The
photothermal quantum yields (the ratio of the rate of photochemical reaction to the rate of
photon absorption) for the pure compounds are summarized in Table 4.23 and for the mixture
components in Table 4.24. The pseudo first-order rate parameters for selected data sets are
given in Table 4.25.
TABLE 4.1
Summary of LS-PDU Exposure Conditions1 For
Samples Analyzed as Pure Compounds
Name ; Initial. Ccmc.-2 Radian.t..In.t...
Chloroform
Trichloroethylene
Tetrachloroethylene
Monochlorobenzene
o-Dichlorobenzene
1,2,3,4-Tetrachlorodibenzo-p-dioxin
2,650 pprn
157
6.15
37,0
1,913
481
17.6 W/cm2
18.1
17.6
17.6
18.1
17,6
samples were run with a mean exposure time of 10 sec in dry air.
Expressed at the initial conditions of 300°C and a mean atmospheric pressure of 745 torr.
4.1 ALKANES AND CHLORINATED ALKANES
It was anticipated that the most challenging class of wastes for the photothermal process
would be those that either don It absorb UV radiation at all, or absorb it only weakly. Alkanes
were considered examples of the former, and chlorinated alkanes examples of the later. For the
purposes of this project chloroform and carbon tetrachloride were taken as example chlorinated
alkanes and with hexane and iso-octane as example alkanes. High temperature absorption spectra
(ca. Section 3) confirmed that the chloroalkanes were weak absorbers of the UV available in the
LS-PDU, while the hydrocarbons were found to be non-absorbers.
42
-------
TABLE 4.2
Summary of LS-PDU Exposure Conditions1 For
Samples Analyzed as Mixtures
Initial Cone.* Radiant Int.
BTEX
Benzene
Toluene
Ethyl Benzene
m-Xylene
Gasoline3
TCE/DCBz/Water-1
TCE
DCBz
Water
TCE/DCB?/Water-2
TCE
DCBz
Water
TCE/DCBz/Water-3
TCE
DCBz
Water
Benzene/Hydrogen peroxide
Benzene (w/F^Oi)
Hydrogen peroxide
Water
Benzene/Hydrogen peroxide-control
Benzene (dry)
Hydrogen peroxide
Water
i 1 .4 ppm
9,62
8.32
8,37
15,4
300
5.19
92.3
5.29
5.19
92.3
5.29
5.19
0
6.68
14.4
63.5
6.68
0
0
17.6 W/cm2
17.6
17,6
17.6
17,6
17.6
17,6
17.6
17.6
17.6
17.6
17.6
17.6
17.6
17,6
17.6
17.6
17.6
17.6
17.6
samples were run with a mean exposure time of 10 sec in dry air.
2Expressed at the initial conditions of 300°C and a mean atmospheric pressure of 745 torr.
3Taken as having a mean molecular weight of 140 g/g-mol and a liquid density of 0.75 g/ml
-------
The thermal and photothermal data obtained on chloroform are summarized in Table 4.3
and Figure 4.1. These data illustrate that, as expected, this example compound does not respond
favorably to the photothermal process using xenon arc illumination. These results are consistent
with the weak rate of photon absorption by this compound. Indeed, calculations indicate that
even if the photothermal quantum yield were unity, the result would be similar to that observed.
Indirect evidence from the behavior of carbon tetrachloride, which was seen as a product from
chloroform (ca. Section 5), suggests that this compound would likewise be unresponsive to the
photothermal process using xenon arc radiation.
Table 4.3
Summary of LS-PDU Results For Chloroform
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
o>
.£
._
._
E
CD
DC
OJ
,_
Temperature
300°C
400
500
550
600
fr(0)
100%
100
59.1
1.89
0.115
fr(17,6)
100%
97.4
60,6
1,95
0.00735
100-
80-
60-
40-
20-
•• Thermal
,,©.,.. photothermal
r
200
300
400 500 600
Temperature,
700
800
°C
Figure 4,1, Summary of thermal and photothermal data for chloroform exposed to 0 and 17.6
W/cm2 of xenon arc radiation for 10 sec in air.
44
-------
Given the lack of response from chloroform (as a parent compound) and carbon
tetrachloride (as a product from chloroform), thermal and photothermal tests with the alkanes
were not conducted. These results suggest the photothermal process would not be generally
effective in destroying weakly, or non-absorbing compounds. Possible mechanisms for
addressing these compounds may include using radiation sources which emit shorter
wavelengths than the xenon arc lamp used here, or utilizing UV light combined with photo-
initiators such as hydrogen peroxide or ozone.
4.2 CHLORINATED ALKENES
Certainly an important class of compounds often found in Superfund sites are the
chlorinated alkenes. Indeed, perhaps the most pervasive contaminant observed in these sites is
TCE. As discussed in Section 3, the alkenes are of interest photochemically because of the
spectroscopic properties resulting from the rc electron structure of the double bond and the
potentially high photo-reactivity of these compounds. These characteristics are further enhanced
by the presence of heteroatoms within the target molecule and the spectroscopic data presented
in Section 3 confirmed that these molecules are reasonably efficient absorbers of UV radiation.
The LS-PDU results for TCE and PCE are summarized in Tables 4.4 and 4.5, and in
Figures 4.2 and 4.3, respectively. The photothermal quantum yields are summarized in Figure
4.4. These data clearly show that the chloroalkenes do indeed efficiently decompose through a
photothermal process. Specifically, both compounds showed significant conversion at 300°C, the
lowest temperature at which the LS-PDU could be operated with the xenon arc illumination
system. At this temperature 12.2% of the TCE was destroyed and 36.6% of the PCE. In both
examples the onset of thermal decomposition began at about 500°C where the photothermal
process was destroying 57.3% of the TCE and 67.9% of the PCE. These examples illustrate that
the photothermal process can destroy chloroalkenes at temperatures where no thermal
destruction occurs.
45
-------
Table 4,4
Summary of LS-PDU Results For Trichloroethylene
Exposed 10 sec in Air to 0 and 18.1 W/cm2 Xenon Arc Radiation
Temperature
300°C
400
500
550
600
650
700
fr(0)
100%
100
100
79,4
27.0
1.13
0.0697
fr(18.1)
87,8%
70.1
42.7
18,7
6.90
0,238
Table 4.5
Summary of LS-PDU Results For Tetrachloroethylene
Exposed 10 sec in Air to 0 and 17,6 W/cm2 Xenon Arc Radiation
o>
c
"E
'<8
05
cc
Temperature
300°C
400
500
600
fr(0)
100%
100
97.1
79.9
fr(17.6)
63.4%
47.8
32.1
19.0
100-
80-
60-
40-
20-
200
300
700
800
400 500 600
Temperalure,°C
Figure 4,2. Summary of thermal and photothermal data for trichloroethylene exposed to 0 and
18.1 W/cm2 of xenon arc radiation for 10 sec in air,
46
-------
O9
'E
*c8
E
03
cc
100-
80-
60-
40-
20-
-•—• Thermal
-e— Photothermal
T
III
200
300
700
800
400 500 600
Temperature, °C
Figure 4.3. Summary of thermal and photothermal data for tetrachloroethylene exposed to 0
and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
0.6-
E
1
as
3
O
0.5H
0.4
0,3
0.2
0.1-!
-e— Trichioroethylene
-B— Tetrachloroethylene
200 300 400 500 600
Temperature, °C
700
800
Figure 4,4. Summary of the photothermal quantum yields for trichloroethylene and
tetrachloroethylene.
47
-------
The photothermal quantum yields summarized in Figure 4.4 illustrate that the quantum
yield (and hence the relative rate of excited state destruction) increases throughout the
temperature range studied. This demonstrates that the rate of reaction from the excited state is
being thermally accelerated as expected. Interestingly, even though the LS-PDU data clearly
shows a greater photothermal conversion for PCE than TCE, it has a lower quantum yield. For
example, at 600°C 27.0% of the TCE remained thermally versus 6.90% photothermally giving a
ratio of thermal-to-photothermal of 3.91 and a quantum yield of 0.409 as compared to 79.9% of
the PCE remaining thermally versus 19.0% photothermally giving a ratio of 4.21 but a quantum
yield of only 0.158. This suggests that the competing process of deactivation of the excited state
may be more efficient for PCE than for TCE. This is consistent with spectroscopic theory in that
as the allowedness of the SQ->SI transition increases, so does the allowedness of the reverse
process. [6]
One characteristic of the data shown here, and in other data sets in this Section, that
should be discussed is the apparent convergence of the thermal and photothermal data at high
temperatures. This apparent behavior is in part a consequence of how the data is presented.
Specially, at high temperatures both data sets are approaching a fraction remaining of zero,
however close examination of the data shows the photothermal data always remains below the
thermal. For example, in the case of TCE, 1.13% of the parent compound remains following a
thermal exposure at 650°C, while only 0.238% remains photothermally. The four fold better
performance of the photothermal process is significant, but appears small on the Figure.
Furthermore, these results merely serve to illustrate that the exposure conditions used in the
laboratory may be inappropriate for the actual processing of this material. The important feature
in these data is they demonstrate the photothermal process is capable of destroying these
materials and the extent of difference between the thermal and photothermal data is sufficient to
obtain the fundamental data needed to design a prototype photothermal system.
4.3 AROMATICS AND ARENES
To quickly assess the effectiveness of the photothermal process on aromatics and arenes
the thermal and photothermal destruction of benzene, toluene, ethyl benzene, and m-xylene were
obtained in a mixture referred to as BTEX. Although some interaction between the mixture
48
-------
components is likely, it is felt this will provide the most time effective means of determining if a
direct photothermal process is appropriate for these types of compounds.
The LS-PDU data for the thermal and photothermal decomposition of the individual
components of BTEX are summarized in Tables 4.6 through 4.9 and Figures 4.5 through 4.8.
The photothermal quantum yields are summarized in Figure 4.9. Comparing these data shows
that benzene was the most thermally stable component of the mixture followed by toluene, ethyl
benzene, and m-xylene. Furthermore, ethyl benzene and m-xylene showed nearly identical
decomposition behavior. The increase in photothermal destruction efficiency follows a similar
pattern. Specifically, benzene was the least affected by the xenon arc radiation, followed by
toluene and ethyl benzene and m-xylene with the latter two showing the largest (and similar)
photothermal effect. These data show that the aromatic compounds as a class are fair candidates
for photothermal processing, though caution would have to be taken with benzene due to it's
relatively high photothermal stability. Certainly a treatability study would be in order prior to
processing a waste with a significant concentration of benzene. Furthermore, as suggested above,
using an illumination system which delivers radiation deeper in the UV than the xenon arc
system used here should improve the performance of a photothermal system.
Table 4.6
Summary of LS-PDU Results For the Benzene Component of BTEX
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr( 17.6)
300°C 100% 99.7%
400 100 99.5
500 98.3 96.0
550 98.3 91.1
600 83.7 64.9
650 30.4 24.3
700 0.998 0.998
49
-------
Table 4.7
Summary of LS-PDU Results For the Toluene Component of BTEX
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr(17.6)
SOOT 100% 96.6%
400 100 97.0
500 93.9 88.0
550 94.1 74.3
600 53.5 24.8
650 4.95 2.80
700
Table 4.8
Summary of LS-PDU Results For the Ethyl Benzene Component of BTEX
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr( 17.6)
300°C 100% 91.7%
400 100 90.8
500 89.3 77.9
550 85.7 53.2
600 32.2 9.12
650 1.09 0.452
700
Table 4.9
Summary of LS-PDU Results For the m-Xylene Component of BTEX
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr( 17.6)
300°C 100% 90.8%
400 100 89.3
500 88.5 76.2
550 83.8 49.7
600 29.1 7.51
650 0.838 0.361
700
50
-------
ra
c
"c.
"to
03
cc
JB
o
100-
80-
60-
40-
20-
0
-»— Thermal
•o---- Phototherma!
"""I"""""! 'I
200 300 400 500 600 700 800
Temperature, °C
Figure 4.5. Summary of thermal and photothermal data for the benzene component of BTEX
exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
o>
c
*c
"S
05
ec
J£
o
5
100-
80-
60-
40-
20-
-*— Thermal
-o-.-- Photothermal
r1 i i | -r"ri i i i > ft I1 « ' * ' i "'• ' * ' 1 ' ' '
200 300 400 500 600 700 800
Temperature, °C
Figure 4,6. Summary of thermal and photothermal data for the toluene component of BTEX
exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
51
-------
o>
c
"c
"55
E
o
CC
Jfi
O
100-
80-
60-
40-
20-
*-
o-
Thermal
-o---- Photothermal
200
800
' I'"' ' l ' I ' *^ -T 1 1 1 I I I I • » • I • •
300 400 500 600 700
Temperature, °C
Figure 4.7, Summary of thermal and photothermal data for the ethyl benzene component of
BTEX exposed to 0 and 17,6 W/cm2 of xenon arc radiation for 10 sec in air.
O5
c
"c
.-
100-
80-
20-,
•I"'""!"""-1!1"""! ....... "{"i ........ 1 ......... i
200 300
i ......... I 'I I \ ........... i"
400 500 600
Temperature, °C
700
800
Figure 4.8, Summary of thermal and photothermal data for the m-xylene component of BTEX
exposed to 0 and 17,6 W/crn2 of xenon arc radiation for 10 sec in air.
52
-------
Interestingly, the quantum yields for the BTEX components shown in Figure 4.9 are
nearly constant at temperatures below SOOT, followed by a rapid rise to a maximum at
approximately 600°C. This clearly suggests an optimal temperature for processing these types of
materials of 600°C to maximize the destruction via photothermal pathways. It is also interesting
to note that while the photothermal conversion of ethyl benzene and m-xylene were nearly
identical (ca. Figures 4.7 and 4.8), the quantum yield for m-xylene is significantly lower than for
ethyl benzene. This result is a consequence of m-xylene having a higher photon absorption rate
constant (ca. Table 3.2), but not a higher rate of conversion than ethyl benzene. As in the case of
TCE and PCE, this result points to a difference in the competing processes of excited state
deactivation.
"2
.2
E
«*-*
1
0
0.70-
0.60-
0.50-
0.40-
0.30-
0.20-
0.10-
0.00-
_B_
- Benzene
- Toluene
Ethyl Benzene
- m-Xylene
i i
200
T"| i ..... 'i .............. i ..... ""
300 400 500 600
Temperature, °C
! ..... |"T""v ..... I-T-
700 800
Figure 4.9, Summary of the photothermal quantum yields for components of BTEX,
53
-------
4.4 CHLORINATED AROMATICS
The data described above illustrated that aromatic hydrocarbons can be destroyed by the
photothermal process. Furthermore, the data presented above for the chloroalkenes suggests that
molecules which contain relatively heavy heteroatoms may be particularly susceptible to the
photothermal process. These two types of compounds are brought together in the chlorinated
aromatics which are an important class of compounds which form the basic building blocks for
larger, more complex materials such as the highly toxic PNAs
The two example chlorinated aromatic compounds selected for this program where
MCBz and DCBz. The LS-PDU data for these materials are summarized in Tables 4.10 and
4.11, and Figures 4.10 and 4.11, respectively. The photothermal quantum yields are summarized
in Figure 4.12. As with the chloroalkenes the photothermal effect is clearly evident with a
significant level of photothermal destruction occurring at temperatures where no measurable
thermal destruction is taking place. In the case of MCBz the onset of thermal conversion occurs
at approximately 500°C whereas the photothermal process has destroyed 25.9% of the feed by
this point. Similarly, for the case of DCBz the onset of thermal destruction occurs at about
600°C at which point the photothermal process has destroyed 53.4% of the feed material.
Clearly the photothermal process is effective in destroying chlorinated aromatic materials.
Furthermore, the data suggest that the extent of conversion increases with chlorine content,
which is the opposite of the thermal response. Specifically, the thermal stability of a waste
typically increases with chlorine content, whereas the photothermal stability is seen to decrease.
This suggests the photothermal process will be particularly effective in destroying wastes with
high chlorine contents or contain other heavy components.
Table 4.10
Summary of LS-PDU Results For Monochlorobenzene
Exposed 10 sec in Air to 0 and 17.6 W/cm* Xenon Arc Radiation
Temperature fr( 17.6)
300°C 100% 85.3%
400 100 83.6
500 99.7 74.1
600 94.4 62.6
54
-------
Table 4.11
Summary of LS-PDU Results For o-Dichlorobenzene
Exposed 10 sec in Air to 0 and 18,1 W/cm2 Xenon Arc Radiation
Temperature
300°C
400
500
600
625
650
675
700
fr(0)
100%
100
100
100
87.8
67.8
33,5
0.406
fr(17.6)
71.3%
69,6
66,3
46.6
27.0
5,84
0,0397
100-
en
.£
"c
"ts
E
CD
cc
80-
60-
40-
20-
-»— - ThermaS
-e Photothermal
200 300 400 500 600
Temperature, °C
700
800
Figure 4.10, Summary of thermal and photothermal data for monochiorobenzene exposed to 0
and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
55
-------
O)
c
'c
"m
E
o>
cc.
o
10 (H
80
60-
40-
w
M
20-
^— Thermal
•©--•• Photothermal
Q
•i r" i t |""i'""'i r11 ••'! i i "i 'i | i » i "i"""-| i i T i""f > r'"i""i
200 300 400 500 600 700 800
Temperature, °C
Figure 4,11. Summary of thermal and photothermal data for o-dichlorobenzene exposed to 0
and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
0. SO-
CD
>
1
CO
O
0,40-
0.30-
0.20-
0.10-^
0.0-
-e— Monochlorobenzene
-B-— o-Dichiorobenzene
•t" i t r"|-i r'- r i""'"} "i i i T"| -i i i i y r t i i { r •r-r-t
200 300 400 500 600 700 800
Temperature, °C
Figure 4.12, Summary of the photothermal quantum yields for monochlorobenzene and o-
dichlorobenzene.
56
-------
The photothermal quantum yields summarized in Figure 4.12 show that for this example
the quantum yield of MCBz and DCBz are approximately equivalent at temperatures to 600°C
Furthermore, as in the case of the BTEX components, a rapid rise in quantum yields is observed
at higher temperatures, though in this case a maximum is not present. This suggests the:
photothermal conversion of these materials would benefit from operating at relatively high
temperatures.
4.5 CHLORINATED DIBENZO-P-DIOXINS
Although perhaps not as numerous as other types of sites, locations contaminated with
high molecular weight chlorinated materials such as polychlorinated biphenyls (PCBs),
dibenzofurans (PCDFs), and dibenzo-p-dioxins(PCDDs) are af concern because of their high
toxicity. Furthermore, the relatively high thermal stability and the requirement for exceptionally
efficient destruction of these types of compounds makes it difficult to destroy them by
conventional thermal means. Conceptually, these types of compounds should be readily
treatable via the photothermal process as they include the structural and compositional
characteristics that should render them very susceptible to the photothermal destruction.
Specifically, PCBs, PCDDs, PCDFs, and PNAs generally include a basic aromatic structure
which absorbs UV photons at relatively long wavelengths as well as numerous heteroatom
substitutions which, combine ta give a molecule that absorbs WV radiation very efficiently and
should be photo-active.
The results of thermal and photothermal tests with TCDD are summarized in Table 4.12
and Figure 4.13, with the quantum yield shown in Figure4,14, In this case the onset of thermal
decomposition was observed at 400°C at which point the photothermal process was destroying
98,69% of the feed. The extent of performance superiority of the photothermal process is
maintained throughout the temperature range studied with the level of improvement increasing
from approximately 72 fold at 4QO°C to over 1,200 fold (64.6% destroyed thermally versus
99.97 15% destroyed photothermally) at 600°C. This clearly shows the photothermal process is
well suited for destroying these types of wastes.
57
-------
Table 4.12
Summary of LS-PDU Results For 1,2,3,4-Tetrachlorodibenzo-p-dioxm
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature
30Q°C
400
500
600
650
700
fr(0)
100%
94.1
83,0
35,4
3.92
0,242
fr(17.6)
3.59%
1.31
0,394
0.0285
C
"5
._
E
®
oc
o
100-
80-
60-
40-
20-
Thermal
- - © - - Photothermaf
i
T
200
300
1 ' T '
400 500 600
Temperture, °C
700
800
Figure 4,13. Summary of thermal and photothermal data for 1,2,3,4-tetrachlorodibenzo-p-
dioxin exposed to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
58
-------
0.030-
0.025-
2 0.020-
I 0.015
° 0,010
0.005
0,000-
200
300
400 500 600
Temperature, °C
700
800
Figure 4.14, Summary of the phototherrnal quantum yields for 1,2,3,4-tetrachlorodibenzo-p-
dioxin.
The photothermal quantum yield data summarized in Figure 4.14 shows that over the
limited span for which data is available the quantum yield peaked at 400°C though it varied
over a narrow range. The interesting feature in this data is the low value of the quantum yield of
approximately 0.024. This shows that this material is relatively photochemically inactive.
However, this relative inactivity is offset by the strength of the absorption as shown by the
photon absorption rate constants listed in Table 3.2. Once again, this result is consistent with
spectroscopic theory which predicts shorter excited state lifetimes, and therefore less time for
decomposition to occur, for chemicals with strong absorption coefficients.
4.6 GASOLINE
The last sample tested as part of the initial evaluation process consisted of a commercial
grade automobile fuel (89 octane unleaded gasoline) to determine if the photothermal process
would be applicable to sites contaminated with this type of material. The complexity of this
sample is illustrated by the GC/FID trace shown in Figure 4.15. This data is from a non-
59
-------
destructive thermal exposure (300°C for 10 sec in air). Detailed inspection of this trace reveals
over 150 individual components. The selected peaks (identifications assigned by mass spectra)
show that the mixture is dominated by alkanes with some significant amounts of aromatic
compounds, most notably toluene and xylenes, at the high molecular weight end of the
chromatogram.
o
c
o
g-
©
EC
Q
u.
o
o
30-:
i
25-
20-=
1 5-^
w
10-:
5j
0.
—
C
1
\
) 5
2
,3 |
I
d
1
4
/
/
C
t
0
e
^
\
\
\
i
7
I
I
k
/
1
8
9
/
10
I i,,
jJIJLJUlW^
5 20 2
Retention Time, mtn
Figure 4.15, Example GC/FID chromatogram of gasoline exposed to non-destructive conditions
(300°C for 10 sec in air) illustrating the complex nature of this sample. Peaks
identified by GS/MS library searches include;
1) 2-Methyl-propane
2) 2-Methyl-butane
3) 2,3-Dimethyl-butane
4) 2-Methyl-pentane
5) Benzene
6) 2,3-Dimethyl-pentane
7) 2,2,4-Trimethyl-pentane
8) Toluene
9) o-Xylene
10) 1,2,4-Trimethyl-benz^ne
60
-------
The overall thermal and photothermal decomposition of the sample is summarized in
Table 4.13 and Figure 4.16. This data shows the weight percent remaining taken as the sum of
all the integrated GC/FID peak areas in the chromatograms from each run. This data suggests a
small amount of photothermal conversion at low temperatures (i.e., 300°C) that remains constant
until the onset of thermal decomposition at approximately 500°C. Above this temperature the
thermal and photothermal curves are essentially identical.
Table 4.13
Summary of LS-PDU Results For Gasoline
Exposed 10 sec in Air to 0 and!7,6 W/cm2 Xenon Arc Radiation
o>
c
CO
E
0)
cc
O)
0)
Temperature
300°C
500
600
700
fr(0)
100%
93,3
22,3
0.385
fr(17.6)
90,5%
89.0
21.5
0.341
100
i -r i r—i i i "i" i r i T"i
•""I"1""!1 i f fTT
200 300
i i i "i
700 800
400 500 600
Temperature, °C
Figure 4.16. Overall decomposition (sum of all integrated GC/FID peaks) for gasoline exposed
to 0 and 17.6 W/cm2 of xenon arc radiation for 10 sec in air.
An alternative analysis of the data from this complex sample is shown in Figure 4.17,
which summarizes the number of peaks observed in each ehromatographic trace. This data
61
-------
suggests an overall simplification of the sample as the photothermal chromatograms consistently
contained fewer peaks than the thermal. However, since the photothermal conversion data does
not show a significant improvement in reactor performance the peaks affected by the
photothermal exposure must be small. Again, the overall thermal and photothermal system
performance is roughly equivalent.
160-
to
Q
0,
o
IS
Q.
2
O)
a
to
o
o
O
i_
0)
JO
140-
120^
100-
80-
60
40-
20-
0
Q. ....
Thermal
,,.©— Photothermal
i ....... i "i .............. '»
200
300
i | i ..... i .......... r-r ...... {
400 500 600
Temperature, °C
'I'1""! I"
I !"•"» I""
700
800
Figure 4.17. Summary of the number of GC/FID peaks observed from the analysis of gasoline
exposed to 0 and 17.6 W/crn2 of xenon arc radiation for 10 sec in air.
Similar behavior is observed when individual components are examined, such as
methyl butane and toluene as shown in Figures 4.18 and 4.19, respectively. Specifically, little
photothermal effect is observed in the major aliphatic and aromatic components. This suggests
that the small differences in the overall reactor performance lies in the impact on minor
components.
62
-------
CB
C
*E
*«
E
CD
CE
5?
jg
O
2
100-
80-
60-
40-
20-
Thermal
_,. .0— Photothermal
r 1 T T-J-I-T-T-1 | I I 1 I | I I I I | I
200 300 400 500 600
Temperature, °C
700
800
Figure 4,18, Summary of the decomposition of the 2-methyl butane (identification assigned by
GC/MS spectral library) component of gasoline exposed to 0 and 17,6 W/cm2 of
xenon arc radiation for 10 sec in air.
D)
c
"
Q>
oc
J)
O
100-H
80-
60-
40-
20-
—»— Thermal
,,,0.... Photothermal
200
300
700
800
400 500 600
Temperature, °C
Figure 4.19. Summary of the decomposition of trie toluene (identification assigned by GC/MS
spectral library) component of gasoline exposed to 0 and 17.6 W/cm2 of xenon arc
radiation for 10 sec in air.
63
-------
In summary then, this data shows that the direct photothermal process would not be a
good candidate for treating fuel spill sites. However, using radiation sources other than xenon
arc (i.e., mercury vapor) or an indirect process using a powerful oxidizer like hydroxyl radicals
from the photo-dissociation of hydrogen peroxide may be worth pursuing.
4.7 MIXTURES OF TCE, DCBz, AND WATER VAPOR
In addition to the complex gasoline sample described above, data was obtained on a
simpler mixture consisting of TCE, DCBz, and water vapor. This mixture was selected to
simulate the principal threats from a site located in western Nevada for which we had actual field
data on the contaminant levels and identities and which was being considered for cleanup by soil
vapor extraction (SVE). A summary of the contaminants and relative concentrations are
summarized in Table 4.14. This information shows that the principal threats are from TCE and
dichlorobenzenes. Two test mixtures were formulated to simulate the relative concentrations of
TCE:
1) at the start of the SVE operation where it was assumed the components would be
in relatively high concentration, and
2) later where the concentration of TCE would be reduced relative to the DCBz as
DCBz's lower vapor pressure would result in it's slower removal.
A third mixture was formulated without water vapor to determine if this component had
any measurable effect on the system performance.
Mixture #1
As shown in Table 4.2 (which summarizes the initial concentrations of mixture
components), the first test mixture was composed of a TCE:Water:DCBz ratio of approximately
60:20:1 indicating the initial draw from the site would be mostly a TCE/water mixture with a
trace of DCBz. The test results for TCE and DCBz components are summarized in Tables 4.15
and 4.16, and Figures 4.20 and 4.21, respectively. The photothermal quantum yields are
summarized in Figure 4.22.
64
-------
Table 4, 14
Summary Of Contaminants At An Example Site In Western Nevada
C.QD.c...,Cm,g^cgl
Phenol 0,87
Chloroform 1,2
P.P-DDD 1,9
P.P-DDT 3.4
Bis(2-rthylhexyl)phthalate 4.9
2,4-DimethylphenoI 5.5
Ethyl Benzene 6,5
2-MethyInaphthalene 1 1
Naphthalene 1 1
Methylene Chloride 17
Chlorobenzene 21
Toluene 23
1 ,2,4-Trichlorobenzene 23
Xylenes 30
Trichloroethylene 280
Dichlorobenzenes 747
Table 4. 15
Summary of LS-PDU Results For the Trichloroethylene Component of Mix #1
Exposed 10 sec in Air to 0 and 17,6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr(17.6)
300°C 100% 82.0%
400 100 60.9
500 70.6 19.5
550 35.1 7,68
600 5,58 1.61
650 0,339 0,0962
700 0.0385
Table 4, 16
Summary of LS-PDU Results For the o-Dichlorobenzene Component of Mix #1
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr(17.6)
300°C 100% 63,1%
400 100 49.1
500 67,0 7.41
550 28.4 6.40
600 15,9 4.26
650 8.14 3,12
700 1.69 0.921
65
-------
OJ
c
*E
,_
QJ
jag
o
1001
BOH
60-
40-
20-
Thermal
,0— Photothermal
200 300 400 500 600 700 800
Temperature, °C
Figure 4,20. Summary of thermal and photothermal data for the trichloroethylene component of
a TCE;Water:DCBz (60:20:1) mixture exposed to 0 and 17,6 W/cm2 of xenon arc
radiation for 10 sec in air.
100-
D3
c
"E
._
®
DC
jg
O
—©..,, photothermal
40-
20-
200
700
800
T
400 500 600
Temperature, °C
Figure 4.21, Summary of themial and photothermal data for the o-dichlorobenzene component
of a TCE:Water;DCBz (60:20:1) mixture exposed to 0 and 17,6 W/cm2 of xenon
arc radiation for 10 sec in air.
66
-------
2
X
E
3
to
3
O
0,60-
0.5CH
0,40
0.30
0,20
o.ioH
0,00-
Triehloroethyiene, Mix #1
o-Dichlorobenzene, Mix #1
T r- T-pr-r T i j • i • i j • » * • j •
200 300 400 500 600
Temperature, °C
700
800
Figure 4,22, Summary of the photothermal quantum yields for the trichloroethylene and o-
dichlorobenzene components of a TCE:Water:DCBz (60:20:1) mixture.
Comparing the data in Figures 4.20 and 4.21 shows a significant improvement in reactor
performance operating in a photothermal versus thermal mode. Specifically, for both
components the onset of thermal decomposition occurs at approximately 400°C whereas the
photothermal process has destroyed 39.1% of the TCE and 50.9% of the DCBz. It is interesting
to note that the onset of thermal decomposition of the TCE and DCBz when tested as pure
compounds did not occur until approximately 500 and 600°C, respectively, indicating the
thermal stability of the mixture components were reduced compared to their stability when
tested as pure components. This type of behavior has been reported previously from studies of
thermal decomposition and has been attributed to interaction between the parent species and
reactive radicals formed from the decomposition of the mixture components. [10] Similar
behavior is seen in the photothermal data suggesting a similar mechanism is occurring in this
environment as well. Another interesting feature is seen in the DCBz data, where the apparent
stability of material increases following an initial rapid rate of decomposition. Reviewing prior
data with TCE tested as a pure compound (Section 4.2) revealed that this sample produces DCBz
as a PIC in very low yields (<1%). Therefore, what is being observed here in the mixture data is
67
-------
the initial destruction of the DCBz parent at low temperatures, followed by its appearance as a
PIC from the reaction of TCE. Although the absolute yield of DCBz from TCE is quite small, it
is significant compared to the amount of DCBz in the system. In any event, these data clearly
show that the photo-thermal process should be able to readily destroy this mixture under the
appropriate conditions.
Comparing the photothermal quantum yield data in Figure 4.22 with that for the pure
compound in Figures 4.4 (TCE) and 4.12 (DCBz) shows that an overall increase in the quantum
yield is observed illustrating the interaction of the mixture components increased the overall rate
of photothermal decomposition. It should be noted that the decline in the DCBz quantum yield is
a result of the production of this compound as a PIC from TCE which had the effect of reducing
the overall destruction rate on this component. Even with this effect, the quantum yield for
DCBz in the mixture exceeded that for DCBz as a pure compound.
Mixture #2
As shown in Table 4.2, the TCE:Water:DCBz ratio in the second test mixture was
approximately 1:20:1. The results from the tests with this sample are summarized in Tables 4.17
and 4.18 and Figures 4.23 and 4.24 for TCE and DCBz, respectively. The photothermal
quantum yields are summarized in Figure 4.25. Once again we find a comparatively early onset
of thermal decomposition at approximately 400°C where the photothermal process has destroyed
26.5% of the TCE and 40.8% of the DCBz. Although the extent of photothermal conversion is
not as great as in the first formulation, it is certainly significant and this mixture would still be
considered an excellent candidate for photothermal destruction.
Table 4.17
Summary of LS-PDU Results For the Trichloroethylene Component of Mix #2
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(17.6)
SOOT 100% 88.0%
400 100 73.5
500 95.4 56.0
600 68.8 32.7
650 44.0
700 3.92 0.648
68
-------
Table 4,18
Summary of LS-PDU Results For the o-Dichlorobenzene Component of Mix #2
Exposed 10 sec in Air to 0 and 17,6 W/cm2 Xenon Arc Radiation
Temperature
300°C
400
500
600
650
700
fr(0)
100%
100
90,2
77.2
48.9
5,11
fr(17,6)
69,2%
59.2
46.3
31,8
0.949
OS
.£
•—
*<5
E
CD
cc
05
-o
100-
80-
60-
40-
20-
- Thermal
., photothermal
Thermal, dry
Photothermal, dry
200
1 i '
300
T
400 500 600
Temperature, C
700
800
Figure 4.23, Summary of thermal and photothermal data for the trichloroethylene component of
a TCE:Water:DCBz (1:20:1) mixture exposed to 0 and 17,6 W/cm2 of xenon arc
radiation for 10 sec in air. Data for the same mixture run without water (1:0:1) is
also shown.
69
-------
O)
.£
"E
"5
E
03
cc
100-
80-
60-
40-
20-
.0.
4
A
Thermal
Photothermal
Thermal, dry
Photothermal, dry
200
1 i '
300
1 I • ' ' ' I" l ' ' * I '
400 500 600
Temperature, C
700
800
Figure 4,24. Summary of thermal and photothermal data for the o-dichlorobenzene component
of a TCE:Water:DCBz (1:20:1) mixture exposed to 0 and 17,6 W/cm2 of xenon arc
radiation for 10 sec in air. Data for the same mixture ran without water (1:0:1) is
also shown.
o>
.£
._
._
CD
CC
0.30-
0,25-
0.20-
0.15-
0.10-
0.05-^
-B-
Trichloroethylene, Mix #2
o-Dichiorobenzene, Mix #2
0.00 -"|""-i i i "i 'i "t" T"-r i 'i i' 'i ""i" i" |''"i' i t" "i" j :"i i-T T""|"'i r-'r-r
200 300 400 500 600 700 800
Temperature, C
Figure 4,25, Summary of the photothermal quantum yields for the trichioroethylene and o-
dichlorobenzene components of a TCE:Water;DCBz (1:20:1) mixture.
70
-------
In contrast to Mixture #1, the quantum yield data for this formulation summarized in
Figure 4.25 shows the quantum yield for DCBz is slightly higher in the mixture than the pure
compound (ca. Figure 4.12) and that for TCE is somewhat lower (ca. Figure 4.4). These tests
clearly illustrate that the behavior of mixtures can differ markedly from pure compound
performance. Our experience has been that overall mixtures tend to be less stable than pure
components, though this is not always the case. This emphasizes the need for feasibility studies
to be a part of any scale-up plan
Mixture #3
A brief set of tests were conducted with a formulation identical to Mixture #2 except the
water component was withheld giving a TCE:Water:DCBz ratio of approximately 1:0:1. It was
hypothetically considered that the water vapor might be participating in the photothermal
decomposition process in some indirect way such as acting as a radical scavenger or possibly
reacting with ozone (which may have been formed photochemically in the reactor) to form
hydroxyl radicals. The results from these tests for TCE and DCBz are summarized in Tables
4.19 and 4.20, and included in Figures 4.23 and 4.24, respectively. These data illustrate that the
water vapor had essentially no effect on the thermal or photothermal process.
Table 4.19
Summary of LS-PDU Results For the Trichloroethylene Component of Mix #3
Exposed 10 sec in AirtoO and 17.6 W/cm~ Xenon Arc Radiation
Temperature fr(0) fr(17.6)
300°C 100%
600 69,8 35.0%
Table 4.20
Summary of LS-PDU Results For the o-Dichlorobenzene Component of Mix #3
Exposed 10 sec in Air to 0 and 17.6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr(17.6)
300°C 100%
600 72,8 31.0%
71
-------
4.8 BENZENE AND HYDROGEN PEROXIDE
A brief series of tests were performed with a mixture of benzene, hydrogen peroxide, and
water vapor to determine if the photo-dissociation of hydrogen peroxide to hydroxyl radicals
could provide a mechanism of photothermally destroying compounds through an indirect
photothermal process. As shown in Table 4.2 the benzene:hydrogen peroxide:water ratio was
approximately 1:3:13. Furthermore, control tests were conducted with an identical sample of
benzene without hydrogen peroxide or water vapor.
The data for the hydrogen peroxide and control tests are summarized in Tables 4.21 and
4.22, respectively, and in Figure 4.26 with the quantum yield summarized in Figure 4.27. As
these data illustrate, essentially no effect was observed in either the thermal or photothermal
destruction of benzene. This was a rather surprising result in that the photochemical dissociation
of hydrogen peroxide is a common process for decontaminating dilute aqueous waste streams.
Thermal degradation of the hydrogen peroxide was considered unlikely as the control test was
nearly identical to the mixture test. An examination of the UV absorption spectrum perhaps
explains these results. As the spectrum in Figure 4.28 for hydrogen peroxide in water at 20°C
shows this sample is an exceptionally weak absorber in the wavelength region of xenon arc
emission (i.e., wavelengths greater than 230 nm). Indeed, it's nearly an order of magnitude
weaker than chloroform (ca. Figure 3.1). Therefore, even if the quantum yield for the photo-
dissociation of hydrogen peroxide were unity, the rate of reaction would be too slow to show
significant conversion of benzene on the time scale used here. This problem may be overcome
using other photo-initiators, radiation sources which reach further into the UV, or both.
Table 4.21
Summary of LS-PDU Results For Benzene in the Presence of Hydrogen Peroxide
Exposed 10 sec in Air to 0 and 17,6 W/cm2 Xenon Arc Radiation
Temperature fr(0) fr(17,6)
300°C 100% 99,7%
400 100
500 99,3 97,0
600 89,7 85.7
650 48.5 41.7
700 8,63 7.48
72
-------
Table 4,22
Summary of LS-PDU Results For Benzene in the Absence of Hydrogen Peroxide
Exposed 10 sec in Air to 0 and 17,6 W/cm2 Xenon Arc Radiation
Temperature
300°C
650
fr(0)
100%
48.4
fr(17.6)
41,5%
The quantum yield data for the benzene component of this mixture shown in Figure 4.27
illustrates the quantum yield for benzene is quite small, though it rapidly increases with
temperature. Given that the conversion of benzene in the control sample was nearly identical to
the mixture sample the results for the photothermal conversion and quantum yield reflect those
for pure benzene,
100-
OS
"E
._
0)
DC
o
5
80-
60-
40-
20-
-•—— Thermal
-e— Photothermal
A Thermal (w/o H202)
A Photothermal (w/o H2Q2)
200
300
700
800
400 500 600
Temperature, °C
Figure 4.26. Summary of thermal and photothermal data for the benzene component of a
Bz:H2O2:Water (1:3:13) mixture exposed to 0 and 17.6 W/cm2 of xenon are
radiation for 10 sec in air. Data for the same mixture run without hydrogen
peroxide or water (1:0:0) is also shown.
73
-------
33
0)
>
E
*~*
c
PQ
3
O
0,040-
0,035-3
0,030
0.025
0.020
0,015
0.010
0.005'
0.000
200
300
-T I I I I I | I I I I | I
400 500 600
Temperature, °C
700
800
Figure 4,27. Summary of the photothermal quantum yields for the benzene component of a
Bz:H2O2:Water (1:3:13) mixture.
10-
o
*
o
.2
o
*£*
X.
LU
k*
JS
o
8
4.
2-
i — i — i
ra"-i — i
200
240 280 320
Wavelength, nrn
j
360
Figure 4.28. Absorption spectrum for hydrogen peroxide in water (referenced against water) at
20°C illustrating that this compound is a very weak absorber of UV radiation.
74
-------
Table 4,23
Summary Of Photothermal Quantum Yields
For Samples Tested As Pure Compounds
Temp
3()0°C
400
500
550
600
650
675
700
TCE
0.109
0.217
0,380
0.551
0,409
0.423
—
....
PCE
0.089
0.123
0.136
—
0.158
—
—
...
MCBz
0,107
0,109
0,1,39
...
0.163
--_
....
„.
DCBz
0.121
0.125
0.101
0.160
0.180
0.330
0.426
TCDD
0.0219
0.0248
0.0212
—
...
...
—
...
Key to abbreviations; TCE Trichloroethyiene
PCE Tetrachloroethylene
MCBz Monochlorobenzene
DCBz o-Dichlorobenzene
TCDD 1,2,3,4-Tetrachlorodibenzo-p-dioxin
Table 4.24
Summary Of Photothermal Quantum Yields
For Samples Tested As Mixtures
Temp Bz
300°C 0.00482
400
500
550
600
650
Key to
0.00646
0,0245
0,0705
0.211
0.167
abbreviations;
Tol
0.0257
0.0200
0.0376
0.129
0.393
0.274
Bz
Tol
Eth Bz
Xyl
TCE #1
DCBz #1
TCE #2
DCBz #2
EthBz Xyl TCE#1 TCE #2
0.0652 0.0479 0.181 0.116
0.0635 0.0486 0.301 0.188
0.0785 0.0556 0.583 0,241
0.256 0.180 0.581
0.634 0.435 0.412 0.246
0.413 0.252
Bezene component of BTEX
Toluene component of BTEX
Ethyl benzene component of BTEX
m-Xylene component of BTEX
Trichloroethyiene in mixture #1
o-Dichlorobenzene in mixture #1
Trichloroethyiene in mixture #2
o-Dichlorobenzene in mixture #2
DCBz#l DCBz #2
0.156
0.232
0.514
0,322
0.263
0.125
0,171
0.156
0.177
75
-------
Table 4.25
Summary of Pseudo First-Order Thermal Oxidation Kinetic Parameters
Measured For The Test Compounds Used In This Project
Name
Chloroform
Carbon Tetrachloride1
Trichloroethylene
Tetrachloroethylene1
Monochloro benzene1
o-Dichlorobenzene
1,2,3,4-Tetrachlorodibenzo-p-dioxin
Benzene (BTEX)2
Ethyl Benzene (BTEX)2
Toluene (BTEX)2
m-Xylene (BTEX)2
Frequency Factor
4.30xl09sec"1
2,88xl05
2.40x108
2,57x106
8.32X104
S.SOxlO18
1.72X104
6.00x109
L61xl08
3.36xl08
1,49x10s
Activation Energy
38,4 kcal/mol
26,0
37.0
33,0
23.0
85,0
20.3
45,6
36,6
38.9
36.3
*As sited in literature reference 5,
2As measured in a mixture of benzene, toluene, ethyl benzene, and m-xylene,
4,9 SUMMARY REMARKS
Overall, the breadth of applicability of the photothermal process illustrated by the LS-
PDU results follows the same general order as the strength of UV absorption. If the pollutant
absorbs the available UV radiation, the photothermal liability generally increases with chlorine
substitution as well as the addition of other functional groups as demonstrated by the alkyl
benzenes. These trends are in accord with accepted spectroscopic theory.
76
-------
SECTION 5
PRODUCTS OF INCOMPLETE CONVERSION
An important aspect of any destruction process is its ability to convert the hazardous
material to mineral products of complete conversion such as carbon dioxide, water vapor, and
hydrogen chloride. Previous research on thermal destruction has shown that a common feature of
the thermal decomposition of organic compounds is the production of organic products of
incomplete conversion (PICs), or products of incomplete combustion as they are called in the
incineration literature. [10] Furthermore, it is frequently reported that conventional, low-
temperature, direct photochemical processes produce complex mixtures of organic by-products
in relatively high yields. [9] Therefore, it is important to demonstrate that the photothermal
process is capable of completely mineralizing (convert to mineral products) the organic
compounds of interest. In the course of obtaining the data discussed in Section 4, information
was also collected on the formation of PICs from both the thermal and photothermal tests.
Analysis of this data varied from general observations to detailed identification and
quantification of individual PICs. This information is summarized in the paragraphs that follow
grouped in a similar manner as the discussion in the previous sections.
5.1 ALKANES AND CHLORINATED ALKANES
As discussed above, the alkanes as a class are considered non-absorbing compounds,
while the chlorinated alkanes as very weakly absorbing within the spectral range of xenon arc
emission (i.e., wavelengths greater than 230nm). Photothermal destruction data for chloroform
suggests that the rate of photon absorption for these compounds is so small that even if they had
quantum yields of unity they would not respond well to the photothermal process using xenon
arc radiation. For this reason only chloroform was examined as an example of this class, and an
examination of the PICs from chloroform suggest a similar response for organic products.
77
-------
Since chloroform was found to be a relatively poor responder to the photothermal
process detailed quantification of products was not conducted. However, identifications were
assigned using the mass spectral data and the relative yields calculated using the GC/FID
response factor for chloroform. Furthermore, the relative yield for the product phosgene was
calculated using the GC/MS response factor for chloroform. These data are summarized in
Tables 5.1 and 5.2, and Figures 5.1 and 5.2, for the thermal and photothermal tests, respectively.
Table 5.1
Summary Of LS-PDU Data For Chloroform
Exposed For 10 sec In Air1
TempfC) CHCb COC12 CCLt C2C14 C2HCb C2C16
300
400
500
550
600
100
100
59.1
1.89
0,115
.
.
18.8
32,1
3.52
.
.
3,34
10.6
17.0
.
-
6.30
31.0
50.1
-
.
1.71
0.190
.
.
_
21.3
26.0
0.198
1 Values are weight % remaining taken as CHCb. COCh data is based on GC/MS data, all others
are GC/FID; f Compound Response]/[CHCl3 ResponseJoxlOO%
Table 5.2
Summary Of LS-PDU Data For Chloroform Exposed
To 17.6 W/cm2 Xenon Arc Radiation For 10 sec In Air1
Temp(°C) CHC13 COC12 CCLt C2C14 C2HC15 C2C16
300
400
500
550
600
100
97.4
60.6
1.95
0.00735
.
,
15.6
37.0
5.45
_
,
3,09
8.82
18,9
.
,
3.40
21.8
20.9
,
0.0864
1.96
0.150
_
.
0.437
18.7
41.4
0.258
1 Values are weight % remaining taken as CHCb. COCh data is based on GC/MS data, all others
are GC/FID; [Compound Response]/[CHCb Response]oxlOO%
78
-------
o>
c
"c
"5
®
CC
x:
OJ
100-
80-
60-
40-
20-
„»_ Chloroform
-*— Phosgene
-*— Carbon Tetrachtorid©
~*— Tetrachloroethylene
-*«—— Pentachloroethane
-+•— Hexachioroethane
T—i—i—i—j—i—r—i—r-|—i—i—i—r
200 300 400 500
Temperature,
600
700
°C
Figure 5.1. Summary of LS-PDU data far chloroform and its major PICs exposed for 10 sec in
air. GC/FID response factors for all PICs except phosgene were taken as
chloroform. Similarly, the GC/MS response factor for phosgene was taken as
chloroform. Note the relative yields of tetrachloroethylene and carbon tetrachloride
continue to increase at high temperature.
Reviewing the data presented in Tables 5.1 and 5.2, and Figures 5.1 and 5.2, shows
similar behavior for most of the PICs from chloroform. Specifically, the PICs appear at the onset
of decomposition (500°C thermally, and 400°C photothermally), reach a maximum yield at
approximately 550°C, then quickly disappearing at higher temperatures. However, two PICs
show a much different behavior. Specifically, in the thermal PICs tetrachloroethylene and
carbon tetrachloride initially appear at much lower yields than the other PICs and continue to
increase in yield even at high temperatures. This suggests these may in fact be PICs resulting, at
least in part, from the decomposition of the other PICs. Interestingly, the yield of
tetrachloroethylene is clearly suppressed at high temperature (i.e., 600°C), whereas the
photothermal yield of carbon tetrachloride is not. This suggests that PICs may follow the same
79
-------
trend observed in the conversion of the waste feed. Specifically, PICs which efficiently absorb
UV radiation should be readily destroyed by the photothermal process while non-absorbing PICs
may be more resistive to the process. Therefore, it would be unlikely to produce highly toxic,
high molecular weight PICs (such as TCDD) from a photothermal system as these materials
have been shown to be efficiently destroyed by the photothermal process.
c
"c
"w
E
03
cc
JC
D)
100-
80-
60-
40-
20-
Chloroform
Phosgene
Carbon Tetrachloride
Tetrachloroethylene
Pentachioroethane
Hexachloroethane
T—I fT
200 300 400 500
Temperature, °C
600
700
Figure 5.2, Summary of LS-PDU data for chloroform and its major PICs exposed to 17.6
W/cm2 of xenon arc radiation for 10 sec in air, GC/FID response factors for all
PICs except phosgene were taken as chloroform. Similarly, the GC/MS response
factor for phosgene was taken as chloroform. Note the relative yield of
tetrachloroethyiene, which absorbs UV radiation, decreases at high temperature,
while carbon tetrachloride, a weak UV absorber, continues to increase.
5.2 CHLORINATED ALKENES
In contrast to chloroform, which produced several major PICs, only two significant PICs
were formed from the thermal and photothermal decomposition of trichloroethylene.
Specifically, GC/FID analysis showed carbon tetrachloride as a major organic product, along
80
-------
with phosgene as detected with the LS-PDU's GC/MS channel. In this case the yield of carbon
tetrachloride was calibrated with an analytical standard. An attempt was made to calibrate the
phosgene GC/MS response, but we were unable to establish a reproducible reference at the low
concentrations required. These problems have been attributed to the high reactivity of this
material with trace water present in the sample handling equipment. Note that the TCE data was
taken prior to the chloroform data reported above for which an alternative semi-quantitative
method for phosgene was used.
The mole % yield of carbon tetrachloride (normalized by the original amount of
trichloroethylene) from the thermal and photothennal tests are summarized in Tables 5.3 and
t
5.4, and Figures 5.3 and 5,4, respectively. These data illustrate that the photothermal process
produces this PIC in a yield comparable to the thermal process (18.9% versus 15.0%), This
result is consistent with the results of the tests with chloroform described above. Specifically, the
direct photolysis process may not be effective in suppressing PICs which do not efficiently
absorb UV radiation. However, the data presented here shows that the photothermal process is
capable of mineralizing this product at high temperatures, though this is likely due to thermal
reactions.
Table 5.3
Summary Of LS-PDU Data For Trichloroethylene
Exposed For 10 sec In Air1
Temperature°C Trichloroethylene Carbon Tetrachloride
300 100
400 100
500 100
550 79.4 3.15
600 27.0 15.0
650 1.13 10.6
700 0.238 0.276
!Data are mole % remaining; fCompound]/[Trichloroethylene}oxlOO%
81
-------
Table 5,4
Summary Of LS-PDU Data For Trichloroethylene Exposed
To 18.1 W/cm2 Xenon Arc Radiation For 10 sec In Air1
Temperature^C
319
406
506
554
605
655
Trichloroethylene
87,8
70,1
42.7
18,7
6.90
0.238
Carbon Tetrachloride
1.77
10,6
17,8
18.9
8.05
lData are mole % remaining; [CompoundMTrichloroethylene]oxlOO%
OS
.£
"c
'(6
E
0)
a:
O
5
100-
80-
60-
40-
20-
-»— Trichloroethylene
HI— Carbon Tetrachloride
0 — ""•"•• ?""""' IT "-"I |""i i i "i "i"""'1"
200 300
I"-! i i-'ft""i | r " i i i
400 500 600 700 800
Temperature,0C
Figure 5.3, Summary of LS-PDU data for trichloroethylene and its major PIC exposed for 10
sec in air.
82
-------
100-
O)
c
"tz
"w
E
03
cc
s°
0s
J)
O
Tnchloroethylene
Carbon Tetracrtloride
0
"T'T"1"'T"T T"""1 -J'-r'T"""! r '{ I" "I—I1 "I "I"
200 300 400 500 600
Temperature,°C
700
800
Figure 5.4, Summary of LS-PDU data for trichloroethylene and its major PIC exposed to 18,1
W/cm2 of xenon arc radiation for 10 sec in air.
With respect to phosgene, although quantitative data is not available, qualitatively the
results are similar to that for this PIC from chloroform. Specifically, the GC/MS data shows that
phosgene is destroyed at high temperatures in a manner similar to the carbon tetrachloride.
It should be noted that PIC data was not obtained for tetrachloroethylene as this
compound was run over a limited temperature range primarily to measure the photothermal
quantum yield.
5.3 AROMATICS AND ARENES
During the period that the LS-PDU tests were being conducted on BTEX, the data
system associated with the LS-PDU's GC/MS channel was not operating properly, so mass
spectral information on PICs for this mixture are not available. However, general observations
can be made by examining the GC/FID data from these tests. Example GC/FID traces from
thermal exposures (300,600, and 700°C), and photothermal exposures (300,600, and 700°C)
are shown in Figures 5.5 and 5.6, respectively.
83
-------
I
o
i
20-
BTEX, 0 W/cm1, 300*C, 10», Air
1
\
BTEX, 0 Won*. 600'C, 10*. Mr
BTEX, 0 W/cm* 700"C, f 0s, >Wr
1
\
2
\
1
3
\
2
/
3
I
1 * * * *
4
/''/
u .
S 10 15
Ratanfen Time, roin
20
Figure 5.5. Example GC/FID chromatograms from BTCX exposed to 300, 600, and 700°C for
10 sec in ak showing that benzene is the most stable component. 1) Benzene, 2)
Toluene, 3) Ethyl Benzene, 4} m-Xylene, 5} PIC PI, 6) PIC P2,7) PIC P3,8) PIC
P4.
84
-------
I
a
I
e
100-
80-
60-
40-
20-
0-
BTEX 17.6 W/cm1, 300'C, 1C*, Air
\
\
V
>
'
100-
80-
60-
40-
20-
BTEX, 17.6 Wan*, 600-C, 10*. Air
4 i e 7
3\(
\
100- BTEX 17,6 W/cm*, 700°C, 10s, Air
80-
60-
40-
20-
\ 1 » "•"«•"' * 1 '"' ' ' ' t
5 10 . 15
Ret^itioo Tme, min
20
Figure 5.6, Example GC/FID chromatograms from BTEX exposed to 300, 600, and 700°C and
17.6 W/cm2 for 10 sec in air showing that benzene is the most stable component,
1) Benzene, 2) Toluene, 3) Ethyl Benzene, 4) m-Xylene, 5) PIC PI, 6) PIC P2» 7)
PIC P3, 8) PIC P4,
85
-------
These data illustrate that four relatively major (maximum yields on the order of 10%)
organic products were formed. Furthermore, although specific identifications cannot be assigned
to these PICs, labeled PI through P4 in Figures 5,5 and 5,6, the chromatographic techniques
used here (temperature-programmed using a non-polar liquid phase) suggests the molecular
weight of these products increases in the order of PI to P4, and that all are higher in molecular
weight than the parent BTEX components. The general production and disappearance behavior
of these PICs, taking their GC/FID response as benzene, are shown in Figures 5.7 and 5,8 for the
thermal and photothermal tests, respectively. Comparing these figures shows behavior consistent
with the discussion above. Specifically, the maximum PIC yields from the photothermal process
are lower than for the thermal process, and the effect increases with molecular weight (i.e., the
higher molecular weight PICs are destroyed or suppressed more efficiently than the lower
molecular weight products). Furthermore, Figures 5.7 and 5.8 illustrate that the parent benzene
component is far more stable than any of the organic PICs and should therefore serve as a
suitable performance indicator for this mixture.
100-
o»
c
"c
"5
I
cc
§s
*-»
JC
S5
I
80-
200
400 500 600
Temperature, °C
700
800
Figure 5.7. Summary of LS-PDU data for the benzene component of BTEX and the major
PICs when BTEX is exposed for 10 sec in air. The data illustrates benzene is far
more stable than any of the organic products. GC/FID response factors for all PICs
were taken as benzene.
86
-------
O)
c
"c
"TO
E
0)
DC
O)
I
100-
80-
i » i I t t i i y i i i 'ry i i i i i f
200 300 400 500 600
Temperature, °C
r"""t""""i"
700 800
Figure 5,8, Summary of LS-PDU data for the benzene component of BTEX and the major
PICs when BTEX Is exposed to 17,6 W/cm2 for 10 sec in air. The data illustrates
benzene is far more stable than any of the organic products. GC/FID response
factors for all PICs were taken as benzene.
5.4 CHLORINATED AROMATICS
The organic PIC data for o-dichlorobenzene is summarized in Tables 5.5 and 5.6, and in
Figure 5.9 and 5.10 for the thermal and photothermal tests, respectively. These data illustrate
that only two PICs, monochlorobenzene and 2-chlorophenol, where produced from this
compound. Both of the observed products reached relatively small maximum yields under both
thermal and photothermal conditions. Specifically, monochlorobenzene reached a maximum
yield of approximately 1.05% (photothermal) and 1.07% (thermal), while 2-chlorophenol
achieved a yield of 1.26% (photothermal) and 0.48% (thermal). Both products were efficiently
destroyed at temperatures above 675°C.
As in the case of tetrachloroethylene, PIC data was not obtained for monochlorobenzene
as this compound was run over a limited temperature range primarily to measure the
photothermal quantum yield.
87
-------
Table 5.5
Summary Of LS-PDU Data For o-Dichlorobenzene
Exposed For 10 sec In Air1
Temperature, °C o-Dichlorobenzene Chlorobenzene 2-ChIorophenol
300 100
400 100
500 100 0.0236 0.0509
600 100 0.116 0.167
625 87.8 0.3396 0.25 1
650 , 67.8 0.734 0,482
675 33.5 1.07 0.314
700 0.406 0.0142
'Data are mole % remaining; [Compound]/[Q-DichlQrobenzeneJoxlQO%
Table 5.6
Summary Of LS-PDU Data For 0-DicbJorobenzene Exposed
To 18.1 W/cm2 Xenon Arc Radiation For 10 sec In Air1
Temperature, °C o-Dichlorobenzene Chlorobenzene 2-Chlorophenol
300 71.3 — 0.513
400 69.6 0.124 0.781
500 66.3 0.468 1.15
600 46.6 0.925 1.26
650 27.0 1.05 0.469
675 5.84 0.301 0.0480
700 0.0397
*Data are mole % remaining; [Compound]/[o-Dichlorobenzene]0xlQO%
88
-------
0}
c
"c
"«
E
®
DC
jg
O
2
100-
80-
60-
40-
20-
-•—• o-Dichlorobenzene
-*— Chlorobenzene
~*— 2-Chlorophenoi
..... i ........... i .......... i ...... -i""} i i
200 300
i i i
......... f "•« « '«"""""'"
400 500 600
Temperature, °C
700
800
Figure 5,9, Summary of LS-PDU data for o-dichlorobenzene and its major PICs exposed for 10
sec in air showing the relatively small yield of the organic products.
o>
c
*E
"eg
E
-------
5.5 CHLORINATED DIBENZO-P-DIOXINS
The LS-PDU tests on TCDD wererelatlvely brief and consequently detailed PIC analysis
was not conducted. However, examination af the GC/F1D data daes show a general perspective
on the production of PICs from this compound. The LS-PDU GC/FID traces from a thermal
exposure at 300°C (100% remaining), 600°C (35,4% remaining), and a photothermal exposure
at 600°C (0,0285% remaining) are summarized in Figure 5,11, The 600°C thermal data shows
the production of numerous PICs which often accompany the thermal decomposition of organic
compounds. This is an interesting example in that it is one of the few cases where the LS-PDU
reactor can be operating in a dominantly photothennal mode (i.e., where there is very high
photothermal conversion, but limited thermal conversion) and the photothermal trace clearly
shows that not only is the parent TCDD destroyed under these conditions, but nearly all of the
associated products as well. Recall that the LS-PDU's cold trap is operated at - 160°C, so that all
but the lightest organic products are collected and analyzed. This emphasizes that the
photothermal process differs significantly from conventional direct photochemical processes
(i.e., near-ambient temperature processes that act directly on the contaminants without the use of
sensitizers or catalysts) in that photothermal decomposition reactions lead to the complete
mineralization of the waste feed. [11, 12, 13, 14]
5.6 MIXTURES OF TCE, DCBz, AND WATER VAPOR
Mixffl
As shown in Table 4.2 the initial concentrations of TCE and DCBz in the LS-PDU
reactor were 300 and 51.9 ppm (58:1), respectively. Given that the concentration of TCE is 58
times greater than DCBz it is not surprising that the decomposition chemistry is dominated by
TCE. This is clearly illustrated in Tables 5.7 and 5.8, and Figures 5.12 and 5.13, which
summarizes the thermal and photothermal data, respectively, taking the initial concentration of
TCE as the basis for comparison. Indeed, on this scale the DCBz is shown as a minor
component that quickly drops below the level of interest (99% conversion) following a
photothermal exposure.
90
-------
300'C thermal (no eonv»r»lon)
TCDD
600"C, therm*)
Organic PICt
.. ... I ,...
TCDD
SQQ'C, 17,6 W/cm* x*non ire radiation
§ 10 15 20 25 30
Chromatographlc Retention Time, min
Figure 5.11. Example GC/FID chromatograms from TCDD exposed to 300 thermal, 600
thermal, and 600°C photothermal (17,6 W/cm2) for 10 sec in air showing that the
photothermal process destroys the complex mixture of PICs as well as the parent
compound.
91
-------
Table 5.7
Summary Of LS-PDU Data For TCE/DCBz/Water Mixture #1
Exposed For 10 sec In Air1
Temperature,°C Trichloroethylene o-Dichlrobenzene Carbon Tetrachloride
300 100 " 1.83
400 100 1.83
500 70.6 1.23 0.447
550 35.1 0.520 2.58
600 5.58 0.291 4.24
650 0.339 0,149 2.40
700 0.0385 0,0309 0.212
!Data are mole % remaining; [Compound]/[Trichloroethylene]0xlQO%
Table 5.8
Summary Of LS-PDU Data For TCE/DCBz/Water Mixture #1
Exposed To 17.6 W/cm2 Xenon Arc Radiation For 10 sec In Air1
Temperature,°C Trichloroethylene o-Dichlrobenzene Carbon Tetrachloride
300 82.0 1.15 0.156
400 60.9 0,899 0.714
500 19.5 0.136 2.85
550 7.68 0.117 4.09
600 1.61 0.0780 3.90
650 0.0962 0.0571 2.12
700 — 0.0169 0.0933
*Data are mole % remaining; [Compound]/[Trichloroethy!ene]oxlOO%
As shown in Tables 5.7 and 5.8, and Figures 5.12 and 5.13, carbon tetrachloride was the
only organic product produced in significant yield. Specifically, the thermal and photothermal
yields both reached approximately 4% between 550 and 600°C. This contrasts sharply with the
data reported earlier for TCE decomposed as a pure compound where the maximum yields of
carbon tetrachloride were from approximately 15% (thermal) and 19% (photothermal). This
reduction in yield may result from the additional hydrogen sources (i.e., o-dichlorobenzene and
water) providing a stable "sink" for the chlorine providing a more favorable pathway for the
formation of hydrogen chloride versus carbon tetrachloride. These data also show that the yield
of carbon tetrachloride is only slightly affected by the photothermal exposure. Once again, this is
consistent with observations discussed above on weakly absorbing compounds.
92
-------
As in the earlier work with TCE as a pure compound, phosgene was observed in the
GS/MS traces, but was not quantified.
Mix #2
Recall that a second formulation of the TCE/DCBz/Water mixture was tested in which
the relative molar concentration of TCE and DCBz were approximately the same (1: 1).
Interestingly, when the data from this mixture was analyzed with respect to PICs, no major
organic PICs were found. This is consistent with the results comparing the PIC yields from TCE
tested as a pure compound versus Mixture #1. Specifically, the yield of the major organic
product carbon tetrachloride may be reduced by the presence of a hydrogen source, in this case
DCBz. This illustrates that the composition of a mixture can influence the relative stability of the
parent compounds and the yield of PICs. This suggests that treatability studies be conducted in
conjunction with the first field trials to make a determination of the potential for PIC formation.
100
80-
o>
c
"E
"5
E
CD
QC
M
0
Trichloroethylene
o-Dichlrobenzene
Carbon Tetrachloride
20-
0-
200 300 400 500 600
Temperature, °C
700
800
Figure 5,12, Summary of LS-PDU data for the trichloroethylene and o-dichlorobenzene
components of a TCE/DCBz/water mixture exposed for 10 sec in air showing the
relative yield of carbon tetrachloride, the only major organic product from this
mixture.
93
-------
100-
80-
o>
c
"E
•-
E
o>
DC
5?
J*
o
Trichloroelnylene
o-DicMrobenzene
Carbon Tetrachloride
200 300 400 500 600
Temperature, °C
700
800
Figure 5,13. Summary of LS-PDU data for the trichioroethylene and o-dichlorobenzene
components of a TCE/DCBz/water mixture exposed to 17.6 W/cm2 of xenon arc
radiation for 10 sec in air taking the initial concentration of TCE as the basis for
comparison showing the relative yield of carbon tetrachloride, the only major
organic product from this mixture.
5.7 SUMMARY
The discussion above shows that unlike conventional, low-temperature photochemical
detoxification, the photothermal process is capable of destroying organic PICs which may be
formed. Furthermore, the process is particularly effective in destroying PICs which absorb near-
UV radiation. Also the data for TCDD, in which the LS-PDU destroyed the test compound very
efficiently at a relatively low temperature, illustrates that the photothermal process is capable of
destroying PICs when the reactor is operating in a predominately photothermal mode. Clearly,
the photothermal process does not proceed through simple, stepwise dehalogenation as is often
the case in conventional, low-temperature photochemical reactions.
One major product observed in these tests which is of particular interest is phosgene,
COC12- Phosgene has been reported as a major product from low-temperature decomposition
processes, and understanding its production and developing methods for its destruction are
94
-------
important because of its relatively high toxicity. One possible reason for the high yields of
phosgene in these, and similar tests, is that at the relative low temperatures used here, the major
mineral carbon product is carbon monoxide rather than carbon dioxide. Furthermore, the
decomposition of compounds with large amounts of chlorine relative to hydrogen is molecular
chlorine rather than hydrogen chloride. Therefore, it is likely the phosgene is being produced by
the reaction of carbon monoxide with molecular chorine as;
CO+C12—>COCl2 (5.1)
This may explain why phosgene has not been observed from the decomposition of
compounds like chloro- and dichlorobenzene where the major chlorine product is hydrogen
chloride rather than chlorine. Therefore, it is likely that phosgene production may not be a
problem in actual waste streams as long as there is an ample hydrogen source available. In
circumstances where a waste contains a relatively large amount of compounds containing
chlorine atoms and an insufficient source of compounds containing hydrogen atoms it may be
possible to efficiently destroy the phosgene by reacting it with added water as;
COC12 + H2O -> CO2 + HC1 (5.2)
In a full-scale system this could easily be achieved by the water which is naturally
present in most process streams or injecting water into the feed stream. Since the mixture tests
suggests that there is certainly no adverse effect of the presence of water vapor, the suppression
of phosgene in a full-scale system should not present a significant problem.
In summary, as with the applicability to the types of waste feed, the only PICs which will
present a significant challenge to the photothermal process are those which do not absorb UV
radiation effectively. This suggests that radiation sources with the shortest possible UV radiation
be used. Furthermore, the mixture test results illustrate that the mixture components can
significantly alter PIC yields. This suggests that feasibility studies be conducted as a part of any
scale-up activities to determine the probability of PIC formation.
95
-------
SECTION 6
BASIC DESIGN FOR THE PROTOTYPE PDU
With the data presented in the preceding sections there is now a sufficient body of
knowledge available to embark on the design of a prototype PDU. This discussion begins with a
review of the potential radiation sources that may be used in a large-scale PDU and the selection
of a specific illumination system for the prototype unit. The geometry of the reactor vessel
which encloses the illumination system to form the basic reactor unit is discussed as well as how
the configuration of this unit affects its performance. Finally, the overall configuration is
examined to arrive at a specific proposed design. From this design the capital and operating costs
are estimated.
6.1 LAMP SELECTION
Reviewing the commercial literature on industrial illumination systems shows that there
are several types of lamps that are possible candidates for the prototype PDU. These include
lamps such as mercury, xenon, mercury-xenon, low, medium, and high pressure, short arc and
linear, continuous and flash lamps. Investigation of these lamps included reviewing the technical
literature from American Ultraviolet Co., Hanovia, ILC Technology, Ushio Inc., and VBI
Technologies and discussions with various technical representatives. The important
characteristics considered included spectral distribution and relative UV output, total output,
geometry, and service life. The primary candidates considered were high-energy short-arc lamps
(recall this type of lamp is used in the LS-PDU), flash lamps, and linear medium-pressure
mercury arc lamps. All three types of lamps offered relatively high output in the range of 10+
kW, but there were increasing concerns about service life of the short arc and flash lamps in the
higher power ranges and there were concerns about these types of lamps being able to operate in
a high temperature environment for long periods of time. Furthermore, most production short arc
96
-------
lamps are relatively inefficient generators of UV radiation The flash lamps do much to
overcome this problem, but suffer from relatively short service life. In contrast, mercury arc
lamps offer relatively long service life (often greater than 2,000 hours), efficiently produce UV
radiation, and should be able to operate at relatively high temperatures. Taking all these factors
into consideration medium pressure mercury arc lamps were selected for the design of the
prototype PDU.
The emission spectrum for a Havonia medium pressure mercury lamp delivering a
nominal 1 W/cm2 is shown in Figure 6, LThe emission spectrum for a comparable xenon arc
lamp is also shown. This illustrates the principal differences between the xenon arc lamp used on
theLS-PDU and the mercury arc lamps being considered for the prototype PDU. In the former
anly about 2.5% of the total output is in the UV at wavelengths shorter than 300 nm and it is
spread out as a continuous emission. In contrast, medium-pressure mercury lamps emit
approximately 15% of their energy at these short wavelengths and is concentrated in several
discrete emission bands. The significant improvement in the photon absorption rate constants of
selected organic compounds using mercury versus xenon arc radiation is illustrated by the values
given in Tables 6.1 and 6.2, respectively, for 1 W/cm2 radiation from each source. The values in
these tables may be used to calculate the photon absorption rate constant at other radiant
intensities by using an appropriate scaling factor. Comparing these data shows an average 10
fold increase in photon absorption rate constant illustrating that as a radiation source mercury arc
is clearly superior to high pressure xenon arc.
Hanovia produces mercury arc lamps for conventional photochemical and UV curing
applications in a variety of sizes. Their largest lamps are rated at 200 and 300 W/in and can
deliver total radiant energy on the order of 15 kW with nominal arc lengths of up to 75 and 50
in., respectively. Furthermore, the envelope temperature of these lamps is typically on the order
of 800°C illustrating that they should be capable of operating in the high temperature
environment of the PDU and their geometry should permit a relatively even illumination of the
PDU interior.
97
-------
E
o
V)
c
0.050-
0.040-
0.030-
0,020-
0.010-
0.0
Mercury
Xenon
I 1 I I 1 1 I IIIIIIiI«ilIII|IIII|ll T ( I* II
220 230 240 250 260 270 280 290 300
Wavelength, nm
Figure 6.1. Comparison of the radiant intensity as a function of wavelength for 1.0 W/cm2 of
high pressure xenon and medium pressure mercury arc radiation illustrating that the
radiation from the latter is concentrated in several high intensity bands in the UV,
Table 6.1
Photon Absorption Rate Constants Using 1 W/cm2 Medium Pressure Mercury Arc Illumination
Temp
IOQ°C
200
300
400
500
600
Clfrm
9.94e-5
3.63C-4
9.63C-4
2,91e~3
Ctet
S.37e-4
1.58e-3
3.57e-3
5.39e-3
8.49e-3
l.51e-2
TCE
3.21e-2
6,66e-2
9.44e-2
U8e-l
!,76e-l
2.37e-l
PCE
2.84e-l
3.91e-i
4,55e-l
5.58e-l
6.12e-l
Bz
3.04e-2
3.57e-2
4.24e-2
4,95e-2
5.49e-2
5.43e-2
MCBz
6.12e-2
7.t3e-2
7.26e-2
8.79e-2
I.OSe-1
DCBz
9.11e-2
I.Wc-i
1.23e-l
1.80e-I
2.25e-l
Tol
6.67e-2
7.22e-2
7,66e-2
8,20e-2
7.89e-2
EthBz
6.61e-2
7.38e-2
7.94e-2
8.55e-2
i.lOe-1
Xyl
8.33e-2
8.64e-2
9.23e-2
l.06e-l
I.14e-l
TCDD
7.85eO
7,80eO
l.lOel
98
-------
Table 6.2
Photon Absorption Rate Constants Using 1 W/cm2 Xenon Arc Illumination
Temp
100°C
200
300
400
500
600
Key to
Cifrm
6,67e-6
2,26e-5
5,30e-5
2,36e-4
Ctct TCE PCE Bz
4.93e-5 1.95e-3 2,05e-3
9.00e-5 4.13e-3 l.76e-2 2,78e-3
I.99e-4 5.89e-3 2,75e-2 3.37e-3
3,89e-4 8,81e-3 3,23e-2 4.49e-3
6.59e-4 1.19e-2 4.36e-2 5.10e-3
I.18e-3 I.62e-2 4,88e-2 6.23e-3
MCBz
6.65e-3
8.02e-3
8.80e-3
1.15C-2
1.35e-2
DCBz
l.41e-2
1.59e-2
1.64e-2
2.30e-2
2,69e-2
Toi
6.36e-3
7.20e~3
8.I3e-3
9.50e~3
1.04e-2
EthBz
6.13e-3
7.26e-3
8.35e-3
9.16e-3
1.36e-2
Xyl
9.45e-3
l.lOe-2
1.21e-2
L44e-2
1.70e-2
TCDD
8.62e-J
9.78e-l
1.44eO
abbreviations: Clfrm Chloroform
Ctet Carbontetrachloride
TCE Trichloroelhylene
PCE Tetrachloroethylene
Bz * Benzene
MCBz Monochlorobenzene
DCBz o-Dichlorobenzene
Tol Toluene
EthBz Ethyl Benzene
Xyl m-Xylene
TCDD 1,2,3,4-Tetrachlorodibenzo-p-d ioxin
6.2 BASIC REACTOR VESSEL DESIGN
With the illumination system selected, the design of the vessel which will house the
radiation source can be designed. To begin it is informative to examine the radiation distribution
around the lamp. To make this calculation a relatively simple model was developed to describe
the radiation distribution surrounding a linear radiation source. The starting point for this model
examines the radiation distribution about a point source. This point source model is then
extended to a linear source by considering the latter as a line of point sources.
Examining the radiant intensity from a single point in a weakly absorbing media it can
be shown that;
I = P/(4rcd2) (6.1)
where I is the radiant intensity (W/cm2) at distance d (cm) from a point source radiating with a
power of P (W). For a linear source consisting of n point sources, Equation 6.1 becomes;
(6,2)
99
-------
where the summation is carried out over the length of the source. Therefore, the total radiation at
any point in space near a linear source is taken as the sum of the intensity contributions from the
individual points along the source. If each point source is considered to have the same power,
Equation 6,2 simplifies to;
I =
(6.3)
where Pe is the power (W) from a single element in the source. Equation 6.3 is the general model
for calculating the radian! intensity as any point in space around a linear source (Equation 6,3 may
actually be used for nearly any source as Song as the appropriate function for dj is used). Applying
this simple model to a 15 kW, 200 W/in lamp gives the results summarized in Figure 6.2, This
Figure illustrates that the radiant intensity drops off relatively quickly as the distance from the
lamp increases. This would suggest a reactor vessel of narrow dimensions to keep the reactants as
close to the lamp as possible. Furthermore, the radiant intensity diminishes quickly past the ends
of the lamp, suggesting the length of the reactor be limited to approximately the length of the arc,
2.00'
-------
The model for the intensity at a single point can be extended to an average intensity if the
vessel is well mixed. Specifically;
Iavg -
(6.4)
where Iavg is the average intensity (W/cm2), I; is the intensity at a single point (W/cm.2), n is the
total number of points, and the summation is carried out throughout the volume of interest. This
calculation was carried out for a cylindrical vessel with a fixed length of 250 cm enclosing a single
200 W/in lamp with a nominal arc length of 200 cm located on the vessel axis. The radiant
intensity was calculated at points spaced 2.5 cm apart throughout the volume of the vessel. Points
within 5 cm of the arc were excluded from the average to take into account the size of the tube
which will enclose the lamp. The results, summarized in Figure 6.3, show that the mean radiant
intensity decreases as the vessel radius increases. This is consistent with a simple cylindrical model
where the wall (with surface area given by 7i(diameter)(length)) is evenly illuminated, losses out
the ends are neglected, and the cylinder is of fixed length. For the perspective of designing a
reactor with a high radiant intensity this again suggests a narrow reactor vessel.
1.2-
o
o
«j
._
«
tr
CO
1.0-
0.8
0.6
0.4
0.2-
Mean Radiance « 25,8(Vessel Radius)*103
R= 0.9984
l i ' I J i » « '
20 40 60
Vessel Radius, cm
' '
80
100
Figure 6.3, Mean radiant intensity as a function of vessel radius for a PDU chamber measuring
250 cm long enclosing a single 200 W/in, medium pressure mercury arc lamp with
a nominal 200 cm arc showing that the mean radiant intensity decreases as the
chamber radius increases,
101
-------
Examining the plug flow reactor performance model given in Equation 1 .1 shows that
the radiant intensity (embodied in the photon absorption rate constant, kab) is one of two
parameters directly affected by the vessel geometry, the second being the mean residence time, t.
Specifically, the mean residence time may be expressed as;
t = V/F (6.5)
where t is the mean residence time (sec), V is the volume of the reactor vessel (m3), and F is the
flow rate through the vessel (mVsec). In the case of a cylindrical reactor vessei;
t = 7iR2l/F (6.6)
where R is the radius of the vessel (m) and 1 is the length (m). This shows that the mean
residence time is proportional to the square of the vessel radius if the vessel length is constant,
Taking into consideration that the mean radiance is proportional to the reciprocal of vessel
radius, the product of rate of absorption and'mean residence time (ca. Equation 1,1) should
increase linearly with radius. Hence, the overall reactor performance should increase
exponentially with vessel radius and even though the radiant intensitydecreases with the radius
of the vessel, the overall performance model suggests a vessel of large diameter,
One consequence of a relatively large reactor is that it is possible to include more than
ane lamp within the vessel, The impact of multiple lamps on the radiant intensity can be
calculated by extending the simple model for a single lamp (Equation 6.3) as:
I = ZZ(l/di2)Pe/4jt (6.3)
where the summations are carries out for each location i and each lamp j. This calculation was
carried out for a cylindrical vessel with a diameter of 120 cm and a length of 250 cm enclosing
lamps rated at 200 W/in with a nominal arc length of 200 cm located on a 30 cm radius from the
vessel axis, The results, summarized in Figure 5.4, show that the mean radiant intensity
increases approximately linearly with the number of lamps. This suggests that a single chamber
may include several lamps if necessary and the increase in radiant intensity will be
approximately proportional to the number of lamps.
102
-------
o
0)
o
J5
T5
-------
0)
o
2
T>
EC
c
01
2
1,6-
1.4-
1,2
1.0-
0.8-
0.6-
0.4-
0.2-
i|iiii|iiii|iii-r-I--i r~T—r-|—i
10 20 30 40 50
Lamp Distance From Vessel Centerline, cm
60
Figure 6.5, Mean radiant intensity as a function of the radius on which the lamps are mounted
for a PDU chamber measuring 250 cm long and 120 cm in diameter enclosing four
200 W/in, medium pressure mercury arc lamps with a nominal 200 cm arc showing
that the mean radiant intensity decreases as the lamps are mounted closer to the
vessel walls which are assumed to be non-reflective.
In summary, there are several factors to be considered in designing the prototype PDU.
Specifically, the type and size of lamps, the size and geometry of the reactor vessel, and the
number and placement of the lamps within the vessel and each of these elements can be altered
to meet a required specification. In general, however, the discussion above suggests using linear
mercury arc lamps housed parallel to the axis of a cylindrical vessel which is only marginally
longer than the arc length and of relative large diameter. A conceptual schematic for such a
vessel is shown in Figure 6.6. Now that a basic reactor system has been established, the
laboratory data can be used to predict the performance of this, and similar systems.
104
-------
250 ass
jfjji
to on
Figure 6.6.
Basic design for a prototype PDU chamber based on the laboratory tests with the
LS-PDU and assuming the use of six 200 W/in. medium pressure mercury arc
lamps with a nominal arc length of 200 cm. Measuring 200 cm in diameter and 250
cm long, the vessel would have an internal volume of approximately 270 ft3 and a
mean radiant intensity of 1.23 W/cm2.
105
-------
6.3 PREDICTED PDU REACTOR PERFORMANCE
A diagram of a basic PDU reactor vessel is given in Figure 6.6. This Figure shows a
PDU reactor vessel housing six medium pressure mercury arc lamps with a nominal arc length
of 2 m. The vessel is 250 cm long and 200 cm in diameter giving an internal volume of
approximately 8 m3, or 270 ft3. With the lamps mounted on a 50 cm radius and delivering 200
W/in, the mean radiant intensity would be 1.23 W/cm2. In the paragraphs that follow the
predicted performance of a PDU based on this design is discussed.
Recall that the performance model presented in Equation 1 .1 is applicable to a plug flow
reactor (PFR). However, it is difficult to envision the flow through the reactor shown in Figure
6.6 behaving in a plug flow manner. Therefore, it is inappropriate to use a plug flow model.
Indeed, since this model gives the highest theoretical conversion it would predict overly
optimistic performance and lead to a design which would give disappointing results. In the
extreme case the reactor performance may approach that of a completely stirred tank reactor
(CSTR) where the performance is described by;
(6.6)
In practice it is likely that the actual flow through the reactor will be intermediate
between the extremes of PFR and CSTR. Fortunately, the intermediate regime can be readily
described using established models such as the tanks-in-series (TIS) model;
fr=U^(kvd + kab}?/N]}-N (6,7)
where N is the equivalent number of "tanks" (CSTR elements) needed to correctly describe the
residence time distribution of the reactor system. The advantage of this model is that by
appropriately selecting the value for N, any residence time distributions from plug flow (N of
infinity) to completely mixed flow (N of unity) can be modeled. [15] Using the pseudo first order
rate parameters and photothermal quantum yields from the tests with the LS-PDU, the
absorption spectra from the HTAS, and the emission spectrum from the lamp manufacture,
Equations 1.7 (for the calculation of kab) and 6.7 can be used to predict the overall reactor
performance.
106
-------
Taking TCE as an example, the PDU performance was estimated using the reactor design
shown in Figure 6.6, Based on our experience with reactors of similar geometry, the residence
time distribution through the PDU vessel was taken as equivalent to two "tanks" (i.e., setting N
equal to 2 in Equation 6.7), The mode! results are summarized in Figure 6,7 as the time required
to achieve 99% conversion at 300,400, 500, and 600°C as a function of the number of PDU
chambers operating in series. This Figure clearly illustrates that the system performance
increases rapidly with increasing temperature and by operating several chambers in series. With
respect to the latter, these results show that as the number of chambers increases the overall
system performance rapidly approaches that of a PFR, Indeed, from the perspective of time
required to reach a specified level of conversion, there is only marginal improvement above 4
chambers, suggesting from a performance standpoint a PDU system should consist of at least 4
chambers. From a capacity standpoint the PDU can be easily scaled by adding additional
chambers in series, or sets of chambers in parallel
The model results presented in Figure 6.7 can also be expressed in terms of system
capacity expressed as the volumetric flow rate of gas through the system. Specifically;
C = (VN7t99)(Tref/TpDu) (6.8)
where C is the system capacity (cfm), V is the volume of each PDU chamber (270 ft3), N' is the
number of chambers, 199 is the time required to reach 99% destruction (min), Tref is the
temperature of the reference state (°K), and TPDU is the mean temperature of the PDU (°K).
Equation 6,8 was used to predict the overall system capacity for a PDU based on the
chamber illustrated in Figure 6.6 processing materials selected from the laboratory tests. The
reference temperature was taken as 293°K (20°C). The results for TCE and PCE are shown in
Figure 6.8 and 6.9, the components of BTEX in Figures 6.10, 6.11,6,12, and 6.13, MCBz and
DCBz in Figures 6,14 and 6.15, and for TCDD in Figure 6.16. These model results illustrate
that the overall system capacity increases nearly linearly with the number of chambers.
Therefore, from a capacity standpoint the size of the PDU may be scaled by simply adding the
required number of chambers to the system.
107
-------
8
to
c
.2
"55
<5
c
o
O
o>
0)
.2
o
1,000
800-
600-
400
200-
300°C
40QSC
500°C
6000C
T
4 6
Number of Chambers
8
10
Figure 6,7, Predicted time to achieve 99% destruction (t99>of trichloroethylene using a series of
PDU chambers as illustrated in Figure 6.6 operated in series assuming each
chamber may be modeled as 2 CSTRs in series. This illustrates that the overall
system efficiently approaches that of an ideal plug flow reactor as the number of
chambers increases and that the rate of improvement rapidly slows with more than
4 chambers. This suggests that a PDU system should be comprised of at least 4
chambers, or a system of chambers with a mean residence time equivalent to 8
CSTRs in series.
The model results summarized in Figures 6.8 through 6.16 illustrate that operating
temperature has a significant impact on the overall system capacity. In the case of the
chloroalkenes (ca. Figures 6.8 and 6.9) a steady increase in throughput is predicted as
temperature increases. In all of the aromatic materials a somewhat limited effect of temperature
is seen at temperatures below approximately 500°C. with a significant increase at higher
temperatures. This suggests that if the system capacity is inadequate at temperatures below
500°C the cost of achieving temperatures greater than 500°C may be offset by the significant
increase in the system capacity.
108
-------
E
"5
"o
a
co
Q
1,000
800-
600-
400
200-
0
4 6
Number of Chambers
Figure 6.8. The predicted processing capacity of a PDU achieving 99% destruction of
trichloroethylene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases rapidly with temperature, and nearly linearly with the number of
chambers.
1,000
800-
o
w
>>
•—
CL
a
O
600-
400-
200
300°C
4CX)°C
500°C
600°C
0246
Number of Chambers
Figure 6.9. The predicted processing capacity of a PDU achieving 99% destruction of
tetrachloroethylene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly with temperature, and nearly linearly with the number of
chambers.
109
-------
o
CO
o
RJ
Q.
«
O
500
400
3QQ-H
200-
100-
4 6
Number of Chambers
8
Figure 6,10, The predicted processing capacity of a PDU achieving 99% destruction of benzene
in BTEX as a function of mean operating temperature and number of chambers
using the basic design illustrated in Figure 6.6 showing the capacity increases
significantly at temperatures above 500°C, and nearly linearly with the number of
chambers,
500
E
"5
o
8.
to
O
400-
300-
200
100-
4 6
Number of Chambers
8
10
Figure 6.11. The predicted processing capacity of a PDU achieving 99% destruction of toluene
in BTEX as a function of mean operating temperature and number of chambers
using the basic design illustrated in Figure 6,6 showing the capacity increases
significantly at temperatures above 50Q°C, and nearly linearly with the number of
chambers.
110
-------
o
to
"o
s.
co
O
500-
400-
300-
200-
100-
300°C
400°C
. 500°C
600°C
0246
Number of Chambers
Figure 6.12. The predicted processing capacity of a PDU achieving 99% destruction of ethyl
benzene in BTEX as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6,6 showing the capacity
increases significantly at temperatures above SOO°C, and nearly linearly with the
number of chambers.
500-
E
*o
o
8.
co
O
400-
300-
200-
100-
HV 3CX3°C
HI 400°C
^ 500°C
^ 600°C
0246
Number of Chambers
Figure 6,13, The predicted processing capacity of a PDU achieving 99% destruction of m-
xylene in BTEX as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above 500°C, and nearly linearly with the
number of chambers,
111
-------
1
"o
C8
Q.
CC
O
500
400-
300-
200-
100-
4 6
Number of Chambers
8
Figure 6,14, The predicted processing capacity of a PDU achieving 99% destruction of
monochlorobenzene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above 400°C, and nearly linearly with the
number of chambers.
500
400-
o
o
5
Q.
W
O
300
200-
100-
3QCPC
400°C
500°C
600°C
0
I ' ' ' r
4 6
Number of Chambers
10
Figure 6.15. The predicted processing capacity of a PDU achieving 99% destruction of o-
dichlorobenzene as a function of mean operating temperature and number of
chambers using the basic design illustrated in Figure 6.6 showing the capacity
increases significantly at temperatures above 500°C, and nearly linearly with the
number of chambers,
112
-------
5,000
4,000-
o 3,000-
5^
1
§• 2,000-
o
1,000-
4 6
Number of Chambers
Figure 6,16. The predicted processing capacity of a PDU achieving 99% destruction of 1,2,3,4-
tetrachlorodibenzo-p-dioxin as a function of mean operating temperature and
number of chambers using the basic design illustrated in Figure 6.6 showing the
relatively high system capacity for this compound and that the processing rate
increases nearly linearly with the number of chambers.
• 3oo°C
400°C
« 500°C
*—-600°C
6.4 ESTIMATED COST OF A PROTOTYPE PDU SYSTEM
The overall capital and operating costs for the PDU chamber illustrated in Figure 6.6
were calculated as summarized in Table 6.3. In this table the costs for the shell and insulating
firebrick where taken as similar to that reported for a hazardous waste incinerator afterburner
and corrected to 1995 costs.[16,17] The cost for the lamps, lamp wells, and lamp ballasts were
from the manufacturer's literature (Hanovia, 1994). These estimates suggest the majority of the
capital costs will be in fabricating the reactor shell, followed by the lamp wells, ballasts, and the
estimated costs of the system support structure and equipment.
With respect to the amortized costs (using a simple linear depreciation model), Table 6.3
suggests the most expensive component will be the lamps and the lamp wells. The 2,500 hours
used for the lamp life were based on the manufacture's estimate assuming 5 hours of operation
for every lamp start. Discussions with the manufacturer indicated that since the lamps will likely
see continuous service in the PDU significantly longer service life is possible which should
reduce overall cost of the lamps, and hence the operating costs of the PDU. The second highest
113
-------
amortized cost is the lamp wells which are expected to degrade from attack from dust and water
vapor. The cost estimates assumes the lamp wells are replaced with new wells for every 10,000
hours of operation. If the wells can be replaced with refurbished units, this cost could be
reduced.
Table 6.3
Estimated Costs For A PDU Chamber Fitted With
Six 200 W/in Medium Pressure Mercury Lamps
Item
Carbon Steel Shell
Firebrick Insulation,
Lamp Wells
Lamp Ballasts
Support Structure
Sub Total
Lamps
Sub Total
Electrical Service7
Grand Total
Hourly Cost8
$24,300 !
$1,370 2
$8,400
$8,500
$8,500 5
$51,070
$3,000
$54,070
Expected Life
20 years
5 years
2 years3
5 years4
20 years
Annual Cost
$1,215
$274
$4,200
$2,125
$425
6 months6
$8,239
$6,000
$14,239
$22,500
$36,739
$7.34
1 Assuming $86/ft2.
2Assuming $L44/ft2-in.
3Assuming 10,000 hour life and 5,000 hours of operation per year.
4Assuming 20,000 hour life and 5,000 hours of operation per year.
5 Assuming support structure and equipment is 20% of the chamber cost less support,
^Assuming 2,500 hour life and 5,000 hours of operation per year.
'Assuming $0.05/kW-hr and 5,000 hours of operation per year.
^Assuming 5,000 hours of operation per year.
114
-------
With respect to consumable materials, only electricity for the lamps is considered here.
As Table 6.3 shows electricity is the largest single contributor to the overall cost of the PDU.
The relatively high electric energy requirement comes from the fact that only about 15% of the
electrical energy is converted to useful UV radiation. If other lamps are made available, such as
low pressure mercury or xenon excimer, this cost could be considerably reduced. This cost is in
part offset by the thermal contribution to the system from the lamps. Specifically, in a well
insulated vessel the 90 kW supplied to the chamber by the six 15 kW lamps is sufficient to heat
approximately 600 cftn of air saturated with water vapor (as from a soil vapor extraction unit)
from 15°C to 500°C. Therefore, depending on the specific site requirements, the heat from the
lamps should reduce the size of a preheater, or eliminate it entirely.
In summary, the estimate summarized in Table 6.3 suggest the overall operating cost for
a PDU based on the design presented in Figure 6.6 should be less than $ 10/hr-chamber including
the depreciation of the equipment, replacement lamps, and required electrical service.
115
-------
SECTION 7
CONCLUSIONS
Based on the results of the laboratory study of the photothermal detoxification process
and the subsequent modeling of the performance of a large-scale PDU, the following
conclusions may be made.
* Hazardous organic coYnpounds whose molecular structure includes alkene or
aromatic structures are likely to absorb near-UV radiation which is necessary for
the photothermal detoxification process,
• The oear-UV absorption spectrum of organic compounds shifts to longer
wavelengths and increases in overall intensity with increasing temperature leading
to an overall increase in efficiency of the photothermal process at high
temperatures.
• Organic compounds which efficiently absorb near-UV radiation (i.e., chlorinated
alkenes, chlorinated aromatics, chlorinated dibenzo-p-dioxins, etc.) are relatively
easily destroyed by the photothermal process.
« Molecules which only weakly absorb near-UV radiation (i.e., alkanes and
chloroalkanes) are not good candidates for a direct photothermal process using
medium-pressure mercury lamps,
• The photothermal process is capable of destroying most organic PICs though
caution should be taken to monitor for the emission of low molecular weight
PICs as part of any scale-up activities.
« Of the high intensity lamps currently available, linear, medium pressure mercury
lamps are the most suitable for a large-scale PDU for the treatment of chlorinated
116
-------
alkenes, chlorinated aromatics. and dioxins due to the high UV output, long
service life, and geometry of these types of lamps.
A large-scale PDU should include at least four cylindrical reactor chambers
operating in series, enclosing lamps mounted near the chamber centerline, and at
a relatively high temperature (i.e., 500-600°C).
The capacity of the PDU system can be adjusted through selection of appropriate
operating conditions (i.e., number of lamps, operating temperature, etc.),
operating chambers in series to increase efficiency and capacity, or sets of
chambers in parallel.
The overall capital cost for a large-scale PDU (270 ft3 chamber enclosing six 200
W/in. lamps) should be on the order of $50,000 per chamber and an operating
cost (including the amortized cost of the capital expenses) of less than $ 10/hr.
117
-------
SECTION 8
REFERENCES
1. J. L. Graham and B. Bellinger, "Solar Thermal/Photolytic Destruction of
Hazardous Organic Wastes," Energy, 12, No. 3/4, pp. 303-370, 1987.
2. J.L. Graham, B. Bellinger, D. Klosterman, G. Glatzmaier, and G. Nix, "Disposal of
Toxic Wastes Using Concentrated Solar Radiation," In Emerging Technologies in
Hazardous Waste Management II, American Chemical Society, Washington, DC,
Chapter 6, pp. 83-109, April 1991.
3. R. Hulstrom, R. Bird, and C. Riordan, "Spectral Solar It-radiance Data Sets For
Selected Terrestrial Conditions," Solar Cells, 15 (1985). pp. 365-391.
4. S. L. Murov, "Handbook of Photochemistry," Marcel Dekker, NY, 1973.
5. B. Dellinger, "Determination of the Thermal Stability of Selected Organic
Compounds," Hazardous Waste, 1, No. 2, pp. 137-157, 1984.
6. S. P. McGlynn, T. Azumi, M. Kinoshita, "Molecular Spectroscopy of the Triplet
State", Prentice-Hall, Inc., Englewood Cliffs, NJ, 1969.
7. A. Levy, "The Accuracy of the Bubble Meter Method For Gas Flow
Measurements," J. Sci. Instrum., Vol 41, pp. 449-453, 1964.
8. W. G. Herkstroeter, "Special Methods in Absorption Spectroscopy," in Creation
and Detection of the Excited State, Vol. 1, Part A, Edited by Angelo A. Lamola,
Marcel Dekker, Inc., New York, NY, 1971.
9. J. L. Graham, D. L. Hall, and B. Dellinger, "Laboratory Investigation of the
Thermal Degradation of a Mixture of Hazardous Organic Wastes-I," Environmental
Science & Technology, 20, pp. 703-710, 1986.
10. B. Dellinger, "Emission of Toxic Organic Compounds, a Discussion of the
Hazardous Waste Incineration Critical Review," APCA Journal, Vol. 37, No. 9,
September, 1987.
11. E. Lipczynska-Kochany, J. R. Bolton, "Flash Photolysis/HPLCApplications. 2.
Direct Photolysis vs Hydrogen Peroxide Mediated Photodegredation of 4-
Chlorophenol As Studied by a Flash Photolysis/HPLC Technique," ES&T, Vol.26,
No. 2, pp. 259-262.
12. J. Hawarl, A. Demeter, R. Samson, "Sensitized Photolysis of Polychlorobiphenyls
in Alkaline 2-Proponol Dechlorination of Arochlor 1254 in Soil Samples by Solar
Radiation," ES&T, Vol.26, No. 10, pp. 2022-2027.
118
-------
13. H. Muto, M. Shinada, Y. Takizawa, "Heterogeneous Photolysis of Polychlorinated
Dibenzo-p-dioxins on Fly Ash in Water-Acetonitrile Solution in Relation to the
Reaction with Ozone," ES&T, Vol. 25, No. 2, pp. 316-322.
14. Chemical Engineering, p. 17, April 20, 1981.
15. 0. Levenspiel, "The Chemical Reactor Omnibook", OSU Book Stores, Inc.,
Corvallis, OR, 1989.
16. G. A. Vogel, E. J. Martin, "Estimating Capital Costs of Equipment of Facility
Components," Chemical Engineering, Vol. 90, No. 24, pp. 87-90, 1984.
17. M. S. Peters, K. D. Timmerhaus, "Plant Design and Economics for Chemical
Engineers," 2nd Edition, McGraw-Hill Book Co., New York, NY, 1968.
119
------- |