Pilot-scale study on ozone-enhanced catalytic oxidation of waste gas emissions from the pulp and paper industry

Authors

  • Endalkachew Sahle-Demessie,

    Corresponding author
    1. U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, Cincinnati, OH 45268
    • U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, Cincinnati, OH 45268
    Search for more papers by this author
  • Catherine B. Almquist,

    1. Paper Science and Chemical Engineering Department, Miami University, Oxford, OH 45056
    Search for more papers by this author
  • Venu Gopal Devulapelli

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, Cincinnati, OH 45268
    Search for more papers by this author

  • This article is a US Government work and, as such, is in the public domain in the United States of America.

Abstract

Field studies were conducted at Domtar's Kraft pulp and paper mill (Hawesville, KY) for 2 weeks to investigate the treatment of high volume, low concentration (HVLC) waste gas streams using ozone-enhanced catalytic oxidation technology. The contaminants in the HVLC waste gas from the pulping area of the mill contained mainly methanol (360 ± 80 ppm) and total reduced sulfur compounds [dimethyl sulfide (DMS) (4400 ± 670 ppm) and dimethyl disulfide (150 ± 25 ppm). The catalysts used in the field studies included ViO2/TiO2 and CuO/MoO3/ã-Al2O3. The results of the field study showed ozone-to-sulfur ratios >2, space velocities <1000 h−1, and reaction temperatures ≥250°C are required to achieve >90% destruction of DMS in the HVLC waste gas stream. Ozone-enhanced catalytic oxidation has key environmental advantages over incineration, including mild operating temperatures and thus, lower energy costs and lower NOx formation. Although the technology appeared to be feasible at the laboratory scale, the field study data revealed that several obstacles must be overcome prior to this technology being implemented at large scale. © 2010 American Institute of Chemical Engineers Environ Prog, 2011

INTRODUCTION

The pulp and paper industry is one of the largest industry sectors in the United States in terms of resource usage and total discharges to the environment. However, over the past three decades the industry has steadily reduced its waste generation and discharge through improved processes and pollution prevention measures. For example, energy integration and process optimization techniques in the pulp and paper industry have been reported in the literature [1–5]. The pulp and paper industry has a long history of recycling its waste and its customers' waste. It is the largest recycler in the United States.

Existing and anticipated air pollution laws have forced industries to reduce air emissions. The integrated multimedia regulation, the “Cluster rule,” for the pulp and paper industry has the goal of reducing toxic pollutant releases to the air and water by almost 60% from 1998 levels by requiring mills to implement Maximum Achievable Control Technologies. The Cluster rule also provides individual mills with incentives to adopt Advanced Pollution Control Technologies that will further reduce pollutant discharges beyond current discharge limits of volatile organic compounds (VOCs) by nearly 50% [6–9]. Currently, the paper industry collects noncondensable waste gas streams from the pulp mill through extensive ductwork from various sources and burns them in a thermal oxidizer or recovery boiler. Although effective on many of the pollutants of concern, there are significant costs associated with this method of air pollution control, including capital and maintenance costs for the extensive ductwork, the costs of energy to move the air emissions from their sources to the incinerator, and the costs of fuel to maintain sufficiently high temperatures in the incinerator to oxidize the organic and total reduced sulfur (TRS) emissions.

Researchers have attempted to develop technologies to reduce emissions from the pulp and paper industry [10–15]. Burgess et al. [10] used a V2O5/TiO2 catalysts and oxygen at temperatures >350°C to convert the methanol in aqueous waste streams to formaldehyde [10], a product that could subsequently be used in resins. The aqueous waste streams used in the study contained approximately 50% methanol and <5 wt % TRS compounds.

The goal of this study was to demonstrate the use of ozone-enhanced catalytic oxidation for the treatment of gaseous emissions from a Kraft mill. Specifically, a mobile pilot-scale system was used to oxidize volatile organic and TRS compounds in a slip stream from a high volume, low concentration (HVLC) noncondensable waste gas stream from the pulping area of the mill. Previously, research on ozone-enhanced catalytic oxidation has been conducted in the laboratory, and pollutants [methanol and dimethyl sulfide (DMS)] that are typically present in Kraft pulp mill noncondensable gas streams were successfully oxidized [11–15]. As a follow-on to the bench-scale work, field tests were conducted to assess the reliability and performance of the technology with an actual waste gas stream.

Summary of Bench-Scale Studies

Bench-scale studies were conducted prior to this field study using synthetic feed streams of individual and mixed pollutants (mainly methanol and DMS) that are found in the HVLC waste gas stream from the pulp mill blow tank. They were conducted to assess the influence of process parameters on the performance of ozone-enhanced catalytic oxidation technology. In addition, the bench-scale tests were run to investigate reaction kinetics and to screen potential catalysts. A summary of the bench-scale results is provided below as a basis for comparison to field-scale results.

  • Ozone-enhanced catalytic oxidation of methanol (15,000 ppm) was conducted at mild temperatures of 150–250°C using a V2O5/TiO2 catalyst. A 98+ % conversion with ozone-to-methanol ratio of 1.2 and gas hourly space velocity (GHSV) of 60,000 h−1 was achieved. The main degradation products of methanol were CO and CO2 with small quantities of methyl formate [11, 12].

  • The combination of gas phase and surface reactions on the V/TiO2 catalysts reduced the amount of ozone required for high conversions. In the absence of ozone the catalysts showed very low activity at temperatures lower than 250°C [11–13].

  • The oxidation of DMS (500 ppm) on 10% V2O5/TiO2 catalyst gave high conversion to SO2. Small amounts of partial oxidation products, such as dimethyl disulfide (DMDS), dimethyl sulfoxide (DMSO), and dimethyl sulfone (DMSO2) were also formed [13, 14].

  • Ozone was also tested as an oxidant with other catalysts including Cu, Mo, Cr, and Mn oxides, and mixed metal oxides supported on γ-alumina. The best catalytic activities were obtained over 10% V2O5/TiO2 catalyst and 10 wt % CuO–10 wt % MoO3 supported on γ-Al2O3 with 100% DMS conversion and high selectivity (≈96%) towards complete oxidation products, such as CO2 and SO2 [14].

  • The effect of moisture on the oxidation of methanol, methane thiol and DMS was studied using V2O5/TiO2 as a catalyst. Different feed compositions and gas hourly space velocities were used in the studies, which were conducted at 250°C and an O3-to-substrate mole ratio of 2.1. Conversions of methanol and TRS compounds were high and the main products were CO2, CO, and SO2 with small amounts of partial oxidation products. Increasing the relative humidity of the gas stream from zero to 18,000 ppm reduced the conversion of methanol, where as the conversion of TRS compounds was not affected, although the selectivity for partial oxidation products of TRS compounds increased. High moisture levels have inhibition effects on the reactions because of competitive adsorption of methanol and water on to the same sites [15].

  • Little catalyst deactivation or change in the product distribution was observed during the reaction time of 5 h in the presence of moisture [15].

  • The influence of gas hourly space velocity showed that product distribution can be affected with change in the contact time, which suggested that the reaction may be influenced by mass transfer limitations [11–15].

EXPERIMENTAL METHODS AND PARAMETERS

Project Site

The field project site was Domtar's Kraft mill in Hawesville, Kentucky. This mill is an integrated, with both pulping and paper machine operations located at the site. The mill has an annual paper production capacity of 634,000 tons and an annual pulp production capacity of 455,000 tones bleached Kraft hardwood pulp [16].

The waste gas line that was used for this study was a HVLC stream located at the top of the blow tank (approximately 200 feet above ground), which separates the pulp from the black liquor and volatile gases. The temperature of the HVLC stream was approximately 65°C, and it was under a slight vacuum pressure. Pictures of the process line from which the slip stream was taken are shown in Figures 1a and 1b. Some of the organic compounds in the waste gas line from the blow tank are by-products of the pulping process, formed in the digester as the cooking liquor reacts with the lignin in the wood chips. The HVLC waste gas stream is also saturated with water vapor. The pollutant concentrations in the waste gas stream are provided in Table 1. The project time period for the field study was May 19 through May 30, 2008.

Figure 1.

Photographs of the test site: (a) the tap on the mill's HVLC noncondensable gas line from which the process gas was taken for our study; and (b) the field test trailer and test equipment position relative to the blow tank at the mill. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Table 1. Waste gas composition in the exit stream of blow tank.
PollutantAverage concentration (ppm)
Methanol359 ± 80
Dimethyl sulfide4,384 ± 673
Methane thiolND
Dimethyl disulfide152 ± 25
Hydrogen sulfide350 ± 23
CO517 ± 45
SO292 ± 45

The Test System

A skid-mounted test system was constructed to treat up to 30 lpm of waste gas. A schematic of the test system is shown in Figure 2. A compressor was used to pull the waste gas from a process line at a Kraft mill and pressurize it to approximately 110 psig. A condensate trap was placed between the mill's process line and pilot system's compressor. The compressed gas was then sent through the test system at a controlled mass flow rate via mass flow controllers and rotameters. A second compressor was used to compress ambient air to approximately 10 bars. The compressed ambient air, then, was used to dilute the process gas during experimental trials and to purge the test system of the waste gas following each experimental trial. In the test system, resistance heaters were used to heat the gas up to 150°C prior to the reactor. Thermocouples were placed prior to, within, and following the heated zone to monitor the temperature of the gas in the test system. Temperature controllers were used to control the temperature in the heated zone during the experiments.

Figure 2.

Schematic of ozone-enhanced catalytic oxidation test system. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

A gas phase ozone generator equipped with corona discharge reactor (Ozone solutions, Hull, IA) was used for the generation of ozone. Ozone was produced using ionization of O2 surrounding a high-voltage conductor which is kept at high-potential gradient. High purity oxygen gas was passed through the ozone generator, which can produce 3–25 g/h of ozone. Ozone was mixed with the waste gas stream in the test system at the inlet to the reactor. The ozone generator was fed with up to 10 L/min pure oxygen, and it can generate up to 25 g/h of ozone in the oxygen stream. A tube furnace was provided to control the temperature of the reactor. The reactor was a 28 inch steel pipe with an inner diameter of approximately 2 inches. The catalyst was secured in the center of the reactor. A by-pass line was provided to assess the composition of the gas stream just prior to the reactor. A gas sensor (Testo Model 360) was used to monitor the inlet and effluent for oxygen, sulfur dioxide, carbon monoxide, hydrogen sulfide and carbon dioxide. A computer was used to track temperatures within the reactor during selected experimental trials. Sampling ports within the test system provided opportunity to sample the gas prior to the compressor, following the compressor, following the ozone injection port, and following the reactor.

Safety was strongly considered in the design of the test system. All streams from the test system were vented through an activated carbon filter to remove residual pollutants from the gas stream prior to release into the environment; the gases were released from the carbon filters approximately 3.6 m above the ground. Check valves were placed prior to the waste gas compressor, following each mass flow controller, following the ozone generator, and following the dilution air compressor. These check valves prevent backwards flow into the critical parts of the test system. In addition, a pressure relief valve was added to the test system just prior to the reactor inlet to prevent a build-up of pressure in the test system due to reactor or carbon bed plugging or operator error. Finally, all resistance heaters were wrapped in an insulation to protect both the heat tape and the operator from contacting hot surfaces.

Power to the test system was provided by the mill. The power to the test system entered through a fuse box on the test system, and the power was distributed in the box through appropriately sized fuses for each piece of equipment in the test system.

Experimental Design and Sample Collections

Multilevel factorial experimental design was used in testing the effectiveness of ozone-enhanced catalytic oxidation for treating waste gas effluents from the pulp and paper industry (Table 2). The design was run in two blocks and fully randomized. The experimental design factors include temperature, flow rate, flow composition, and the type of catalyst. The experimental approach used was to set the reaction parameters to their desired values and then characterize the gas stream prior to and following the ozone-enhanced catalytic oxidation reactor. The responses were conversions of feed pollutants and product selectivity.

Table 2. Experimental design for field tests.
Operating conditionSlip stream V1 (L/min)Diluent (air) V2 (L/min)Oxygen V3 (L/min)Total flowSpace velocity (h−1)Ozone (%)Furnace set temperature (°C)Catalyst
17121054544.30250Cu-Mo
214242010,9091.80250Cu-Mo
321363016,3621.20250Cu-Mo
47121054545.50200Cu-MO
514242010,9093.00200Cu-Mo
621363016,3622.00200Cu-Mo
77121054544.90250ViO2/TiO2
814242010,9092.90250ViO2/TiO2
921363016,3622.00250ViO2/TiO2
10100102054541.30250Cu-Mo
1113072054541.60250Cu-Mo
1260142054540.90250Cu-Mo

The experimental approach was to qualitatively characterize the waste gas from the Kraft mill process line and both qualitatively and quantitatively characterize the process gas in the pilot-scale test system before and after the reactor. Key analytes for the process gas were methanol, hydrogen sulfide, DMS, methyl mercaptan, and DMDS. Oxidation products were also assessed, including sulfur dioxide, carbon monoxide, carbon dioxide.

The process waste gas line in the mill was sampled during the field study period (May 19 through May 30, 2008). Although the digester is a continuous process, there may be upsets or changes made during each day that would alter the process waste gas flow and composition. Gas stream and condensate samples were analyzed using gas chromatography with flame ionization detection (GC-FID) or a stack emission portable gas analyzer with real-time capability (Testo 360). Prior to flowing feed gas to the test system, the waste gas with diluents and oxygen (no ozone) was analyzed by directing the flow through the by-pass line. The GC-FID and the Testo 360 gas analyzer were used to analyze the key constituents in the waste gas. Once the analyses were within 5% of each other in three consecutive samples, the flow was directed through the reactor. The waste gas from the reactor was also analyzed using the GC-FID and the Testo gas analyzer. When the analyses were within 5% of each other in three consecutive samples, steady state was assumed.

During the test runs one sample of the feed stream and three samples of the reactor effluent gas stream were analyzed with the GC-FID. A three-way valve was used to direct flow from the outlet stream of the O3 generator to an ozone analyzer every hour to determine the ozone level in the feed. The HVLC gas stream, the reactor outlet stream, and the exit stream from the activated carbon beds were analyzed online with the Testo 360 gas analyzer.

Analysis Methods

The conversions of all volatile and semi-VOCs including waste gas components the reaction by-product compounds were calculated based on their disappearance from the inlet flow. Concentrations of all organic compounds were measured with an Agilent 6890 Series GC, equipped with a split/splitless injection port, Supelco-QTM-plot (30 m × 0.32 mm # 33577-04A) capillary column and FID detector. Samples were taken from the influent and effluent streams and introduced into the GC injection port by using a gas-tight syringe. Quantification of the all target analytes was done by performing the calibration of all the reactants and reaction products.

The Testo 360 gas analyzer was used to analyze the light gases, including oxygen, carbon monoxide, carbon dioxide, and sulfur dioxide. The Testo gas analyzer pulls 1 lpm flow into the unit for analysis during operation. The data from the gas sensor was read from the display screen on the instrument and recorded as appropriate in a field study notebook. The gas sensors were delivered with manufacturer's calibration for each gas. These calibrations were checked with ambient air for oxygen and with purchased calibration gases for carbon monoxide, carbon dioxide, and sulfur dioxide.

An ozone analyzer (Model-OLA, Ozone Services, Yanco Industries, Burton, BC, Canada) connected in the bypass loop from the outlet stream of the generator to the reactor was used for measuring ozone concentration in the stream.

RESULTS AND DISCUSSION

The composition of the process gas being delivered to the pilot-scaled test system over the 2-week field study is summarized in Figure 3. Noted is that the predominant emissions in the HVLC waste gas stream are DMS and methanol. However, the methanol was contained predominantly in the condensate formed as the HVLC waste gas stream cooled in the drop line from the top of the blow tank to ground level, where the pilot-scale system was located. The flow stream was cooled to about ambient temperature followed by a compression stage to 10 bars prior to feeding to the test system. The process gas stream was diluted with ambient air and/or oxygen, according to the experimental design shown in Table 2, prior to the reactor. The approximate water vapor concentration in the reactor feed stream was calculated to be 3000 ppm by assuming the stream exiting the compressor was saturated with water vapor at near ambient temperature (30°C) and at the pressure in the compressor (10 bars).

Figure 3.

Process gas composition flowing into the pilot-scale test system, following condensation and compression but prior to dilution. The balance of the process gas stream predominantly consists of nitrogen, oxygen, and water vapor. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

At the core of the process is the catalyst that utilizes ozone to oxidize organic pollutants at temperatures ≤250°C. The oxidation is a combination of gas phase and surface reaction where ozone adsorbs to the surface of the catalyst and forms oxygen adsorbed species such as ozonide, superoxides, and atomic oxygen.

Chemical Reactions

Ozone catalyzed oxidation of waste gas compounds methanol and TRS resulted in partial oxidation products of DMDS, DMSO, and DMSO2 and the formation of CO2, SO2, and water as complete oxidation products. Effluent stream composition of each component of the reactor outlet stream compared to the in the inlet stream (prior to the reactor) at a reactor temperature of 250°C and for two different catalyst is given in Figure 4. Noted is that the concentrations of the oxidation products, including SO2, CO, CO2, and DMSO2 increase as expected.

Figure 4.

Concentrations of pollutants in the reactor outlet stream relative to its concentration in the inlet stream at a reaction temperature of 250°C and feed flow rate of 10 L/min for reaction either on CuO-MoO3-γ-Al2O3 or V2O5/TiO2.

The reaction mechanism for the generation of active species from ozone in the absence and in the presence of water is shown below. Nascent oxygen species is produced in presence of third body (i.e., *= catalyst surface), which is the active species for the catalytic oxidation of methanol and TRS compounds with ozone in absence of water [17, 18]. The moisture effect originates from the direct reaction with H2O in the vapor phase to produce [BOND]OH species. Such water-derived species can activate O2 or O3 molecules or modify the electronic state of the vanadium that is exposed at the surface without changing the reaction mechanism. Generation of active species at the surface of a catalyst:

equation image(1)
equation image(2)
equation image(3)

Main reactions:

equation image(4)
equation image(5)
equation image(6)
equation image(7)
equation image(8)
equation image(9)
equation image(10)

Catalyst Development

Two inexpensive metal oxide catalysts, CuO-MoO3/γ-Al2O3 and V2O5/TiO2, were used for this study, although more catalyst and catalyst supports were screened for the bench-scale tests [11]. The use of these catalysts for treating waste gases, and gases containing TRS compounds with oxygen as an oxidant at 200–400°C have been reported before [15–23]. Copper-based catalysts with promoters have shown high activities at low temperatures and resistance to sulfur fouling [18–23]. The activity and stability of copper-based catalysts, such as CuO/γ-Al2O3 can be promoted with Mo for the oxidative decomposition of (CH3)2S2 [21].

Differences in the effectiveness of the catalysts are shown in Figure 4. The V2O5/TiO2 catalyst yielded much more CO but less SO2 than did the CuO/MoO3 catalyst. This data agreed with the bench-scale study where the best catalytic activity was obtained over 10 wt % CuO–10 wt % MoO3/γ-Al2O3 for TRS compounds, which afforded complete conversion of DMS to CO, CO2 and SO2 at 100°C and mole ratio O3-to-DMS of 1.8. Both catalysts appear to be effective in oxidizing H2S. However, neither catalyst oxidized methanol nor TRS compounds completely under the field test conditions. This result differs significantly to our bench-scale studies, which showed excellent conversion (nearly 100%) of both methanol and DMS at 250°C [11–13]. Some of the reasons for this discrepancy between reactor performances at the bench-scale and those at the larger pilot-scale system are the following: the pilot scale may be lacking good mixing of the reactants and ozone; nonideal plug flow; the catalyst area was not used effectively; and nonisothermal flow conditions in the larger reactor. This result highlights the need for additional experimental and reactor modeling studies for the transition between laboratory-scale studies and large-scale implementation.

The carbon and sulfur balances of the reactor inlet and outlet stream were accounted only at ∼60% and 40%, respectively. Some of the possible for reasons are suggested for the low sulfur and carbon balances: (1) DMSO, the partial oxidation product of DMS, adheres to the walls of the test system and did not elude from the test system as a gaseous by-product, which was observed in the laboratory-scale tests as well [11, 12], (2) the sulfur in the gas stream may react with the copper in the CuO/MoO3 catalyst, which is supported by the fact that blue precipitate, which was potentially copper sulfate, was observed in the effluent line of the pilot-scale test system, and (3) the H2S sensor and CO2 data on the gas analyzer were suspected of being inaccurate. The CO2 data is calculated rather than measured data by the Testo gas analyzer, and it appeared to respond very slowly to changes in CO2 in the stream being analyzed.

Process Variables

Figure 5 shows the effect of space velocity and reaction temperature on DMS conversion over the CuO/MoO3 catalyst in the pilot-scale test system. The DMS conversion decreased with increasing space velocity and decreasing reaction temperature as was observed in the bench-scale tests. As space velocity increases, the contact time between the process gas and the catalyst decreases. Therefore, the observed trend was expected. Likewise, reaction rates increase with temperature. Therefore, the observed decrease in DMS conversion as reaction temperature decreased was also an expected trend.

Figure 5.

Effect of space velocity and reaction temperature on methanol and DMS conversion over CuO/MoO3-γ-Al2O3 catalyst at 200 and 250°C. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The effects of ozone-to-sulfur ratio in the feed stream on DMS conversion and selectivity to SO2 are shown in Figure 6 and for CO in Figure 7. As observed in the bench-scale tests increased ozone/sulfur ratio results in greater DMS conversion and greater selectivity toward SO2. The DMS conversion increased to exceed 90% only when the ozone-to-sulfur mole ratio was about two, which is a lot higher amount than needed in the bench-scale tests. At higher ozone concentrations the DMS conversion was relatively independent of the ozone-to-sulfur ratio. The selectivity to SO2, however, is rather independent of the ratio until the ozone-to-sulfur ratio of two or more, the selectivity to SO2 increases significantly. The selectivity to CO also increases with ozone-to-sulfur ratio >2. This result was rather unexpected, since it would be more intuitive if the CO oxidized to CO2 at higher ozone/sulfur ratios. It is possible, however, that as the ozone/sulfur ratio increased, the compounds that adhered to the surfaces of the test system were oxidized, thus increasing the observed apparent formation of CO with ozone/sulfur ratio. In addition, the measurements of CO2 concentrations were inconsistent and did not meet quality control checks, it is expected that the selectivity to CO2 would also increase with increasing ozone-to-sulfur ratios.

Figure 6.

Effect of ozone/sulfur ratio on DMS conversion and selectivity to SO2. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 7.

Effect of ozone/sulfur ratio on DMS conversion and selectivity to CO. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Comparing Laboratory Studies with Pilot-Scale Tests

The laboratory studies indicated that combined use of ozone and catalysts were essential for the complete oxidation of methanol and TRS compounds such as DMS to CO2 and SO2 (>99%) [10–14]. Laboratory scale reactors were valuable in estimating kinetic parameters, determining effects of process variables and for catalyst development. However, experimentation on a large pilot-scale was necessary to elucidate the operational problems and difficulties, such as the effects of maintaining isothermal conditions, improper mixing, clogging of feed and outlet ports, solids accumulation, etc. Based upon bench-scale studies, operational issues were not anticipated. However, operational issues arose during the field study, and they are listed below:

  • The axial temperature profile in the packed bed reactor showed a wide temperature range from the top of the reactor to the bottom, where the higher temperatures were at the bottom of the reactor. The temperature-time-history of the different axial points in the reactor for three flow rates is shown in Figure 8. The results show that it took more than an hour to reach steady-state conditions. In the absence of long preheating times of the feed stream, a large variation of temperatures were observed at different axial positions and the system was slow to reach steady-state conditions when the set furnace temperature was 200°C. Noted is that the operating temperature on average 50°C less than the specified reactor temperature. The high flow rates 10–30 L/min allowed high Reynolds numbers (19,000–5700) and short contact times of 1.8–5 s. The packed bed temperature was governed by the inlet stream temperature and heat transfer from the wall and packed bed to the flow was not as fast as heat convection. Heat from the exothermic reaction was minor as the concentrations of contaminants were low. The nonisothermal reactor condition could have affected the performance of the catalyst significantly and may account for the discrepancy from the bench-scale test results.

  • Blue precipitate formed in the reactor and deposited on the lines of the test system. The color of the precipitate (bright blue) suggests that it is copper sulfate.

  • The ozone generator was damaged prior to the field study, and the replacement ozone generated did not have the desired capacity that the former one had. Therefore, our field study was conducted at lower ozone-to-sulfur ratios than desired.

  • The sulfur and carbon balances were low, and the likely reason for the low carbon and sulfur balances is the presence of DMSO, a partial oxidation product of DMS, which adheres to the test system surfaces.

Figure 8.

Axial temperature distribution within the packed bed reactor during the test runs. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

CONCLUSIONS

The pilot-scale field study conducted at the Domtar Pulp Mill using a slip stream from a HVLC waste gas stream at an integrated Kraft pulp and paper mill allowed us to successfully assess the feasibility of ozone-enhanced catalytic oxidation for the destruction of organic and TRS constituents at their source. Although the technology appeared to be feasible at the laboratory scale, several obstacles, as noted above, must be overcome prior to this technology being implemented in large scale. Insufficient destruction efficiencies for the organics and TRS compounds were observed in the field study for this technology to be considered as a replacement for incineration. The presence of value-added materials from partial oxidation products was not observed. Ozone/sulfur ratios >2, low space velocities <1000 h–1, and reaction temperatures >200°C are required to achieve >90% DMS destruction.

Acknowledgements

Authors are grateful for the support of Dennis Waldroup, Domtar Paper Company and the Hawesville operation management, Dan Hart from Miami University for technical help, and Mark Kemper and Albert Foster from US EPA for their assistance in the field work.

Ancillary