Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
D Workshop Presentations A DEVELOPING GENERATION OF OBSERVATION AND MODELING STRATEGIES James G. Anderson Harvard University In the next few decades, several problems of considerable societal interest will emerge with a central theme of the balance between societal objectives and scientific curiosity-driven research, both of which are very important for the fu- ture. Three problems of key importance in the area of observations and modeling are: · Migration of nitrate, sulfate, heavy metals, and organic soot from urban or regional areas into the broader environment and associated public health issues; Forecasting climate change and testing the forecast in a way that is ac- ceptable to a much broader range of individuals involved both in science and in public policy; and . . Ultraviolet dosage, which is in many ways a statement of what it is that society really cares about, particularly in connection with the ozone question. The way in which we attack these problems the observations, the model- ing, and the way public understanding evolves are crucial issues. For example, Charles Kolb's presentation provides a beautiful introduction to the issue of undersampling and the requirement for significant advances in core technology that must underpin the scientific case leading to effective public policy. 66
JAMES G. ANDERSON 67 The issue of carbon sources and sinks is strategically tied back to the ques- tion of nitrate, sulfate, heavy metal, and organic soot emission. If we can attack one problem and solve it, we will be attacking both. The question of how nitrate will affect human health is closely tied to carbon sources and sinks, which in turn are linked to climate. Consequently, answering the carbon-nitrate-sulfate question requires very high spatial resolution of fluxes, isotopes, and reactive intermediates. There are particular problems with region-specific studies (Box 11. For ex- ample, the Indian subcontinent as it links into the tropical region during the monsoon season is as different from the other seasons as any two regions on Earth. In addition, the way in which these systems couple from the regional to the global scale are extremely important for the issue of prediction. Describing a system and understanding it well enough to predict really separate the strategies of observations and modeling. In addition to seasonal signatures, we want to understand vertical fluxes and urban regional source sinks driven by human activity. We understand chemical transformations very well as illustrated by the Los Angeles basin. We know ex- actly what reactions are taking place, but the lack of specificity on the location and strength of sources and the way those couple into the chemical transforma- tion serves to prevent a link between science and public policy on that question.
68 APPENDIX D The strategy for analyzing NOX across the globe by comparing satellite ob- servations, models, and in situ aircraft observations hinges on the fact that the NO to NO2 ratio goes up dramatically with increasing height in the troposphere. For example, NO2 measurements from the European Space Agency's Global Ozone Monitoring Experiment (GOME) satellite provide a way of analyzing NO2 in the boundary layer and just above the boundary layer. Therefore it possible to test the differences between the GOME satellite measurements and global chemical mod- els. Those differences tell us where we have to attack the problem with a more sophisticated set of measurements. In particular, the large differences between the modeled and satellite-observed concentration fields establish the priorities for aircraft field campaigns that are the only means by which we can adjudicate these differences. Now consider the third question that of UV dosage. Here we have an emerging marriage between the atmospheric dynamics, chemistry, and medical communities because malignant melanoma, as we'll see in a moment, is a huge and increasing health issue. In fact, skin cancers are the only cancers worldwide that are increasing in a statistically important way in the face of developing medi- cal methods. So it's this union between the dynamics and the loss of ozone at midlatitudes (Figure 1) that provides the fundamental information that most of the ozone loss is taking place in the very lower part of the stratosphere. This leads to several questions: · Which mechanisms are responsible for the continuing erosion of ozone over mid latitudes of the Northern Hemisphere? · Will rapid loss of ozone over the arctic in late winter worsen? Are these large losses coupled to midlatitudes? · How will the catalytic loss of ozone respond to changes in boundary con- ditions on water and temperature forced by increasing CO2, CH4, and so forth? How we respond in the tuture requires an understanding of the mechanisms that control the long-term erosion of ozone at midlatitudes. We know why ozone is destroyed over the Antarctic and over the arctic, but there is also a seasonally dependent midlatitude ozone loss. The months of March, April, and May define a key period when schools let out, final exams are over, and the younger popula- tion gets a large episodic dose of ultraviolet radiation. This March-April-May period shows the largest long-term erosion over the past 20 years, approaching 10% ozone loss per decade. However, the strategy one takes depends, to an ex- tent, on whether this is a societal or a scientific issue. Let's take the position that we don't know the mechanisms that control ozone erosion over midlatitudes (which at this juncture is true) and assume that the erosion of ozone will continue as it has in the past decades. Then we're going to look at the coupling with the large ozone losses in the arctic triggered by conver- sion of inorganic chlorine to free-radical form on ice particles or cold liquid aero-
JAMES G. ANDERSON 69 FIGURE 1 Midlatitude ozone loss. sots each winter in the polar vortex. We come to the conclusion, in such an analy- sis, that it's the dynamical structure of the atmosphere that underlies the funda- mental cause of this midlatitude ozone loss, not chemistry directly. From the societal perspective, the raw numbers are large and important for basal cell carcinoma: 800,000 cases per year. Until we understand the mechanism of midlatitude ozone loss, a simple extrapolation may be inaccurate, but it pro- vides an important reference for discussion (Figure 2~. By simple extrapolation there would be an increase from 800,000 cases of basal cell carcinoma in the United States annually to nearly 1.9 million by 2060. The logarithmic depen- dence of the cross section results in a 2% increase in UV at these optical depths for a 1% decrease in ozone. The biological amplification factor emerges from the medical community, and it's a change in human morbidity. For malignant mela- noma, the numbers are very much smaller, but the fractional death rate, as op- posed to morbidity, is much greater. Consider the pattern of ozone over the Northern Hemisphere, and ask if this simply results from large ozone concentrations over the arctic migrating back into midlatitudes. That would correspond to chlorine- and bromine-catalyzed
70 APPENDIX D FIGURE 2 Hypothetical trend for basal cell carcinoma if the trends in midlatitude ozone erosion continue unchanged. ozone destruction over the winter arctic merging back into midlatitudes. It's a completely reasonable, simple, understandable hypothesis, and it emerges from the fact that in the early 1970s we had this dome of ozone that represented the sequestration of ozone moving from the low latitudes into the high latitudes. As the ozone moves northward and downward, it's sequestered in this dome, and that's the way the world has worked for millions and millions of years. Until the late l990s, as the ozone layer began to thin in many winters, it was emulating the Antarctic in a dramatic way. Does this low-ozone air created in the winter vortex flow back into midlatitudes, causing the observed minimum in the March-April- May period? An analysis of data from the last National Aeronautics and Space Administration (NASA) arctic mission indicates clearly that this does not occur. There are indeed large ozone losses in the vortex, but all of the large-scale flow is from the tropics northward and downward. Also, because we know the seasonal phase of CO2 and water over the tropical tropopause, we know that there is no communication backward from the polar regions to midlatitudes in these key months of ozone loss. Therefore, it isn't simply ozone-depleted arctic air moving back into midlatitudes. Does long-term ozone erosion result from chemical loss of ozone at midlatitudes in the lower stratosphere? Susan Solomon made the very reasonable suggestion that the penetration of cirrus clouds and cold aerosols into the lower stratosphere initiates the conversion of inorganic chlorine to free-radical form,
JAMES G. ANDERSON 71 and she carried out a number of modeling studies that support this.) However, evidence garnered from hundreds of crossings of the tropopause by the ER-2 aircraft which provides very high-resolution simultaneous measurements of tropopause position, water vapor and temperature, percentage of observations with ice saturation, and C1O concentrations demonstrates that cirrus clouds and cold aerosols capable of providing the heterogeneous site for inorganic chlorine to free-radical conversion do not exist. The absence of observed C1O is the most compelling point. It would take about a 50-200 parts per trillion (ppt) of C1O to drive the ozone loss that's observed, but the experimental measurements showed less than 2-4 ppt of C1O, all the way up to 4 km above the tropopause. Consequently, we do not believe that in situ loss of ozone is responsible for the long term trend in midlatitude ozone erosion, even though that would be the most reasonable explanation. This brings us back to the fundamental unsolved question of the coupling between the tropics and high latitude. As the temperature of the ocean surface warms in response to increasing CO2 forcing, how does it affect the boundary condition on the entrance of water vapor into the stratosphere? The tropical tropopause constitutes a valve that desiccates the stratosphere and strictly controls water vapor entering the stratosphere from the troposphere. Large-scale ascent operating above the tropical tropopause delivers this mix- ing ratio of water vapor into the high latitudes. For temperatures running from 192 to 200 K and a given water vapor curve of 6 parts per million (ppm), the trigger point for formation of high C1O is about 195 K. We have verified this experimentally using ER-2 observations with trajectory calculations demonstrat- ing the temperature history of air parcels moving within the vortex. Consequently, the amount of water vapor, as it increases in the stratosphere in response to increasing temperatures at the tropical tropopause, would instigate a shift in the threshold temperature required to instigate C1O formation to tem- peratures above 195 K, thereby aiding the dramatic loss of ozone in the arctic winter vortex. At the same time, the increase in water vapor in the vortex induces radiative cooling that drops the temperature of the wintertime lower stratosphere at high latitudes, thus exacerbating the destruction of ozone by chlorine radicals. The water vapor is the crucial quantity, and the structure of the tropics turns out to be the centerpiece for understanding all of these issues linking climate and trends in UV dosage. What is needed is to couple the boundary layer in the tropics all the way up through the tropical transition layer into the stratosphere (Figure 3~. The geo- graphic coverage is crucial. The eastern tropical Pacific over Central America is dramatically different from the western tropical Pacific. It is the structure of con- iSolomon, S.; Borrmann, S.; Garcia, R.R.; Portmann, R.; Thomason, L.; Poole, L.R.; Winker, D.; McCormick, M.P. J. Geophys. Res. 1997, 102(Dl7), 21411-21429.
72 7 APPENDIX D FIGURE 3. Deep convection in the tropics. vector in the tropics, the control of water vapor in the middle-upper troposphere, and the formation of high-altitude cirrus in the region between 13- and 18-km altitude that must be understood before predictions of the impact of climate change can be made. There are important suggestions in the paleorecord. Consider the Eocene 50 million years ago. In central Wyoming, there were turtles, alligators, and palm trees, a combination of plant and animal life that extended into central northern Canada. It was an era defined by deep ocean temperatures, running 10 K above
JAMES G. ANDERSON 73 present and warm polar sea surface temperatures. The Northern Hemisphere con- tinental interiors were warm throughout the year. There was no glaciation. How this occurred is an important question that carries crucial messages for today- including how we attack the problem scientifically. The mechanisms responsible for Eocene climate have been suggested to be enhanced meridional heat fluxes. The ocean is always involved here, with reorga- nization of the atmospheric circulation, enhanced greenhouse warming due to high carbon dioxide, and reductions in global topography. Yet none of the models have been able to capture the gentle difference in temperature between the tropics and high latitudes. This was explained by Sloan and Pollard in 1998, when they introduced polar stratospheric clouds into the model.2 These are the site for het- erogeneous reaction, but they also are profoundly important for trapping infrared radiation. They introduced methane in our favorite reaction to produce the oxida- tion leading to the formation of water. They proposed that the methane came from swamps and wetlands. The Eocene went on uninterrupted for 10 million years, but the chemical lifetime of methane is about 7 years. Consequently, we don't believe that this is the explanation. We believe that polar stratospheric clouds are trapping high- latitude radiation in the infrared, but we believe the mechanism comes from the following: If CO2 enters the system or heat moves northward, the gradient be- tween the tropics and the poles begins to soften, leading to reduced excitation of gravity waves and planetary waves driving from the troposphere up into the strato- sphere. As the flux of upwelling gravity waves and planetary-scale waves is re- duced, so is the wave drag effect, which is the dominant pump that lifts material up in the tropics and pushes it down at high latitudes. If the effectiveness of that pump is reduced, the system relaxes back over the tropics so that the boundary condition on water vapor increases, allowing significantly more water to get into the system. We believe that this is the crucial climate state, and it is at the heart of our understanding of the current climate and also of UV dosage. This brings us back to analyzing the meridional cross section in three dimen- sions from the equator to the pole. A long-duration balloon that could remain in the lower stratosphere would allow us to sample vertically by lowering a pack- age. This is similar to stratospheric experiments from the 1980s, but the subtlety of the connection of the dynamics with the radiative and chemical properties demands an entirely different look. We know that we can scan 10 km back and forth, but now we have some tremendous help from fuel cells. The technology of fuel cells allows us to use solar energy for balloon flights lasting several months. Solar energy drives a very efficient propeller to position the balloon at whatever latitude we want to scan. This energy technology has produced major breakthroughs. Even though other approaches are important, we think that the subtlety of this connected system will 2Sloan, L.C.; Pollard, D. Geophys. Res. Lett. 1998, 25(18), 3517-3520.
74 APPENDIX D FIGURE 4 Atmospheric radiative forcing trends. emerge only out of observations of tracers, particles, and the velocity components associated with this when done in a very sensitive way from an immobile plat- form. We must still address the question of testing climate forecasts. Atmospheric CO2 levels are rapidly increasing, as documented in the Intergovernmental Panel on Climate Change (IPCC) report.3 A number of different scenarios have been Intergovernmental Panel on Climate Change (IPCC), Climate Change 2001: The Scientific Basis, Houghton, J.T.; Ding, Y.; Griggs, D.J.; Noguer, M.; van der Linden, P.J.; Xiaosu, D., Eds., Cam- bridge University Press: Cambridge, New York, 2001 (http://www.grida.no/climate/ipcc_tar/).
THOMAS W. ASMUS 75 developed, and it is clear that we're hugging the upper boundary on the release of CO2 that drives the forcing. We are projecting nearly 4 W/m3 by the middle of the twenty-first century (Figure 4~. This is emerging out of the natural variability, and it will be extremely important. The final point is the question of societal objectives. If you look at the ques- tion of climate forecast from societal objectives, what we need is an operational forecast that is tested and trusted. Yet this nation has no operational climate fore- cast. We have no test of the veracity of that forecast; and until we do, we cannot deliver what is needed to the public. The backbone of the climate forecast, of course, is the operational model that links the short-term E1 Nino scale to the longer term. The observing system is the key challenge for testing the veracity of calculations. Carbon sources and sinks have been discussed. I believe that upper ocean observations, climate data records at the surface, and benchmark observation that establish the long-term evolution of the climate in an absolute sense constitute the centerpiece of what must be done. DIESEL ENGINES FOR CLEAN CARS? Thomas W. Asmus DaimlerChrysler Corporation While the Diesel engine has been growing in popularity in the light-duty vehicle segments in many parts of the world, it has essentially stalled in the United States in these segments. In large part this results from the absence of financial incentives given the low levels of fuel taxation in the United States, plus an image problem based largely on experiences of two decades ago. Since that time, Diesel engine technology has improved in many ways, and this manifests itself as in- creased performance and fuel economy accompanied by reduced emissions, odor, and noise. In large part this has been made possible, not by any scientific break- throughs, but rather by machine design and manufacturing technology advance- ments particularly in fuel-injection system components. In the light-duty vehicle segments today the typical fuel economy expectation of a Diesel engine is 40% greater than that of the gasoline engine at equal vehicle performance on a fuel volume-normalized basis. Because Diesel fuel has roughly 15% more energy per gallon than gasoline, the benefit is roughly 25% on an energy-normalized basis. The U.S. regulatory mandates on criteria emissions have been within reach for gasoline and Diesel-powered vehicles to the present, but in 2007 the mandates will be beyond reach for the Diesel with any kind of sensible emissions abate- ment scheme. This, combined with a very weak business case based on low fuel taxation, effectively discourages U.S. industry investment in Diesel the only practical high-fuel-economy alternative to the more conventional and well-known
76 APPENDIX D gasoline engine technology. While the cost of a modern, high speed Diesel en- gine is substantially more than its gasoline counterpart, it is the most cost attrac- tive alternative to conventional gasoline. Evolution of the Diesel Engine Diesel engine combustion is a highly mixing-limited, stratified-charge pro- cess, and herein lies its formidable efficiency advantage over homogeneous- charge gasoline as well as its challenges with respect to NOx and particulate mat- ter (PM). Whereas the homogeneous-charge relies on inlet throttling for load control, the Diesel relies only on injected fuel quantity for this. Early attempts at managing this mixing-limited process often involved using an outside source of compressed air to assist the fuel atomization process. Later, much attention fo- cused on manipulating in-cylinder air flows to hasten the mixing process. Mean- while, machine design and manufacturing specialists found practical means to increase fuel-injection pressures and to better manage machining process for bet- ter control of fuel-injection precision. Ultimately there was less reliance on in- cylinder air flow manipulation to support the mixing process; hence higher en- gine speeds were enabled by faster combustion, while NOx and PM emissions were reduced. With this, the pre-combustion chamber Diesel became obsolete, and direct injection (open combustion chamber design) became the standard for Diesel engines of most all sizes and duty cycles. With this came reductions in combustion chamber surface area and also in-cylinder turbulence, both of which contribute to reduced heat losses. Typically a 15% increase in thermal efficiency accompanied this shift. High-pressure common-rail fuel systems are state of the art, they provide substantially improved combustion manipulation ability com- pared to all progenitors, and turbocharging has become standard on all small high-speed, automotive Diesel engines. (Turbocharging Diesel engines is highly beneficial to vehicle fuel efficiency since smaller engines with lower friction can be used. Any fuel efficiency benefit derived from turbocharging gasoline engines is somewhat more conditional.) Relative to conventional gasoline engines, these produce significantly higher torque density and near-competitive power density. At least a portion of Europeans' enthusiasm for Diesel power is their highly de- sirable drivability and performance characteristics. Engine-Out Emissions The fuel efficiency advantage of the Diesel over conventional gasoline is based primarily on the means of load control via injected fuel quantity while the air flow is essentially unrestricted (i.e., there is no inlet throttling); therefore the gas exchange (or pumping) loss is minimal. At light loads, therefore, the overall air-fuel ratio is sufficiently high that homogeneous-charge flame propagation is not possible. Hence, stratified charge, diffusion-limited combustion is the princi-
THOMAS W. ASMUS 77 pal heat release mechanism. Therefore, combustion and post-flame chemistry occur over a wide range of air-fuel ratios. The fraction of heat release that occurs in the range of stoichiometric yields very high temperatures and produces high levels of NOX. The portion of heat release that occurs in fuel-rich zones (equiva- lence ratio = 2 or greater; temperatures = 1400 K or higher) tends to produce PM or soot. Manipulation of local air-fuel ratios is the goal of fuel-injection and in- cylinder air-motion strategies. As described above, recent improvements in fuel-injection equipment have produced remarkable reductions in both of the principal emittants. This has been achieved mainly via higher injection pressures and precision control of the in- stantaneous rates of injection (with techniques described as rate shaping, split injections, pilot injection, post injection, etch. These advanced fuel-injection tech- niques enable a measure of control over the local mixing processes and thus the local air-fuel ratios. Trends toward increased levels of premixed burning tend to be beneficial in terms of reducing both of the principal emittants. These injection- mixing strategies, while very desirable, are limited by the need for timing preci- sion of the thermal autoignition process (i.e., timing of thermal autoignition is compromised when temporal proximity between injection and ignition is in- creased). While the emissions reductions resulting from these advancements are impressive, attainment of the so-called Tier 2, bin 5 legislated limits to be en- forced in 2007 is not possible by these means alone. At the far end of this spectrum is homogeneous-charge compression ignition (HCCI), which has been the subject of considerable research effort over the past several decades. This scheme strives to produce a truly (or nearly) homogeneous charge, which under light-load conditions, corresponds to air-fuel ratios too lean to support either flame propagation or diffusion-limited combustion. Hence, multisite thermal autoignition will occur, provided that a particular temperature threshold is attained. This process is capable of producing very low emissions, but timing control of the thermal autoignition process is severely compromised, since the injection event is no longer capable of precisely triggering heat release. Hence, this scheme is readily demonstrable on a laboratory scale but is generally incapable of sustained operation upon load changes due to perturbations in ther- mal equilibria. Diesel Engine Aftertreatmen Significant publicly and privately funded R&D activities have been in place in this area for the past decade, and while incremental improvements have been realized, they all fall short of what is necessary to achieve legislated requirements for 2007 with respect to both NOX and PM. Complicating matters are low exhaust temperatures at light loads characteristic of automotive-type duty cycles and the presence of significant quantities of oxygen in the exhaust stream. In addition, chemical reduction of NOX in a strongly oxidizing exhaust environment presents
78 APPENDIX D additional and formidable challenges. Obviously, chemical reduction of NOx re- quires that reductant chemicals be added to the exhaust in a manner compatible with the particular aftertreatment scheme. The so-called urea-SCR (selective cata- lytic reduction) scheme is perhaps the most effective means of reducing tailpipe NOx emissions. Aqueous urea is metered as precisely as possible and stoichio- metrically with respect to NOx into the exhaust just upstream of the SCR catalyst, the first stage of which hydrolyzes urea to ammonia. The final stage of the SCR system is an oxidation catalyst that oxidizes any excess of ammonia that may not have been consumed in the SCR catalyst itself. Despite functional attributes of the SCR approach, it comes with two formidable drawbacks: (1) a urea infra- structure would be needed, and (2) the U.S. Environmental Protection Agency (EPA) has virtually ruled out this option. The so-called NOx absorber (or trap) is generally seen as the most attractive option for the United States Although generally less effective than the urea SCR approach, this uses Diesel fuel as a source of chemical reductant. The catalyst is typically barium and platinum on an alumina substrate, where nitric oxide is cata- lytically oxidized to nitrogen dioxide and then reacts with the barium to form barium nitrate. When the barium becomes completely nitrated, the trap is regen- erated by the addition of Diesel fuel somewhere upstream of the catalyst. The twofold role of the secondary fuel addition is (1) to consume all excess oxygen and (2) to produce cracked products (H2, CO, and hydrocarbons in decreasing order of effectiveness) to effect chemical reduction of the trapped NOx. NOx absorbers are extremely sulfur sensitive and are rendered inactive when fuel- bound sulfur, which leaves the engine as SO2, becomes catalytically oxidized to SO3 whereupon it reacts "irreversibly" with barium to form the comparatively more stable barium sulfate. Desulfation of NOx absorbers requires temperatures sufficiently high as to degrade the catalyst. Fuel sulfur levels will directly impact desulfation frequency and hence catalyst life expectancy. This along with the denitration frequency will determine the fuel economy penalty associated with this form of NO aftertreatment. x PM traps are typically ceramic wall-flow filters that trap PM rather effec- tively. PM trap regeneration, on the other hand, has proven to be rather more challenging. Typically an additional source of heat must be added to effect the initiation of oxidative regeneration. Once regeneration is initiated, the exother- micity of the process can (and often does) overheat the filter, causing irreversible damage. Various means may be employed to reduce the required regeneration temperature. These basic aftertreatment elements have been demonstrated in a wide vari- ety of different configurations. All are very sensitive to duty-cycle diversity, fuel quality, packageability, and control approaches. High-mileage durability has not been thoroughly established for any of these systems. Some demonstrations have
THOMAS W. ASMUS 79 been reported for limited duty-cycle testing typically with "tailored," open-loop regeneration control. Overall Conclusions To date there has been no demonstration of system compatibility with the legislated emissions limits to be enforced under Tier 2, bin 5 (70 mg NOX per mile and 10 mg PM per mile) in 2007. (It is typically necessary to target 50% of these emissions limits to ensure high-mileage conformance.) In addition to the afore- mentioned technical obstacles, at current levels of fuel taxation in the United States and customer indifference to matters of fuel economy, private-sector in- vestment in Diesel system technologies is discouraged. In the automotive indus- try, lead times are such that without proven emissions capabilities as of this date, it is highly unlikely that this technology will be available in the United States by 2007 based on the world's most stringent emissions regulations. It is also note- worthy that Diesel technology is the most promising and cost-effective among the major fuel-economy enabling technologies. Opportunities Through the Chemical Sciences In the interest of mitigating some of the risks that could deprive the U.S. market of this high fuel-efficiency technology in light-duty segments, several areas in which the chemical sciences could play a role are listed below. While the chemical kinetics of the thermal autoignition process are relatively well understood, means of controlling the ignition timing in the engine cycle when operating in the HCCI mode are still elusive. Chemists and chemical engi- neers will need to help overcome this obstacle if HCCI is to be executable in automotive practice. The most practical means of aftertreatment control of NOX in the U.S. envi- ronment is the so-called NOX trap, and these are highly susceptible to sulfur poi- soning. Even with the reduced, legislated levels of sulfur in Diesel fuel, the long- term effects of sulfur on NOX trap performance is a major concern. Any means to reduce this sulfur sensitivity would contribute to the long-term success of Diesel technology. The health effects of Diesel PM are contentious, and there are strongly held opinions on both sides of the issue. What is clear is that with the aforementioned advances in Diesel fuel-injection technology, PM mass has indeed decreased, but the PM number has become relatively greater. More decisive scientific conclu- sions on PM toxicity in general could contribute to moving this debate toward the establishment of realistic positions on the issue.
80 APPENDIX D THE CO2 TECHNOLOGY PLATFORM Ruben G. Carbonell North Carolina State University The billions of pounds of organic and halogenated solvents used each year in chemical and materials manufacturing contribute significantly to the total dis- charge of volatile organic carbon (VOC) to the atmosphere and to the contam~na- tion of water and soil. In addition, drying operations utilize huge amounts of energy that is generated mostly by the burning of fossil fuels. As a result, there is great interest in finding novel alternative solvents that would result in cleaner, more efficient processes to enable a sustainable chemical and industrial manufac- tunng base.) High-pressure carbon dioxide is one of the leading alternative solvents for many applications because of its unique physical and chemical properties.2~3 It has a very accessible critical temperature and pressure (31°C and 73.8 bar) that enable its use in the liquid, supercntical, or gaseous state near room temperature. It is also highly compressible, so that its density and other physical properties can be varied over a wide range in order to control its solubility properties. Because it is generated as a by-product in the production of hydrogen, ammonia, and ethanol and because it is present in large quantities in many underground reservoirs throughout the world, it is relatively inexpensive compared to organic or haloge- nated solvents. So much carbon dioxide is currently generated by power plants and ends up in the environment, that even if all chemical and materials processes used CO2 instead of organic or aqueous solvents there would still be no need to generate CO2 from burning fossil fuels. However, the subsequent reduction in VOC and organic solvent emissions into the soil and aquifers would be quite significant. In the supercritical state, CO2 has a liquid-like density but a gas-like diffusivity, so mass-transfer and diffusion processes are greatly enhanced in this region. In its supercntical state it is also highly compressible, so that the density rises from approximately 0.5 g/mL at the critical temperature to about 0.9 g/mL at 10 °C. In that same range, the viscosity and surface tension remain at least an order of magnitude lower than those of water and most organic solvents. At these low temperatures the increased density enhances the solubility properties for or- ganic compounds while the vapor pressure drops significantly from the critical iTaylor, D. K.; Carbonell, R. G.; DeSimone, J.M. Annual Reviews of Energy and the Environment 2000, 25, 115-146. 2McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction, Principles and Practice; Butterworth-Heinemann, Oxford, UK; 1993; 2nd Edition. 3Hyatt, J. A., J. Org. Chem. 1984, 49, 5097-5101.
RUBEN G. CARBONELL 81 pressure, thus decreasing the potential cost of processes utilizing liquid instead of supercritical CO2. When considered as a solvent for chemical reactions, carbon dioxide offers a major advantage because it is essentially inert over wide ranges of temperatures and pressures. This prevents CO2 from participating in any chemical reactions that can contaminate the product or terminate important elementary steps in a reaction that may control molecular structure or molecular weights. In addition, if used as a solvent or as a plasticizer, its high diffusivity and high volatility allow it to evaporate quickly and completely, eliminating all chances of contaminating the desired product. The high volatility also decreases the costs associated with solvent removal from a solid substrate when compared to water and organic sol- vents. Many of these properties of CO2 have been known for years,2 but aside from some small specialty applications such as the extraction of caffeine from coffee beans and the fractionation of some polymeric compounds, CO2-based processes have not made major inroads in industry. Over the last decade, interest in the use of CO2 as a solvent has seen a great resurgence as a result of the discovery of some unique solubility properties associated with CO2 that have enabled the syn- thesis of fluoropolymers in carbon dioxide as well as the rational design of sur- face-active materials that are soluble in CO2. Polymerization reactions are carried out industrially by a number of various methods, many of which involve organic solvents and water in the form of emul- sions. These solvents often participate in the reaction, leading to unwanted side reactions. The polymer must then be dried, resulting in the expenditure of large amounts of energy, often leaving the product with significant amounts of unreacted monomer, and residual solvent. There are tremendous advantages in using CO2 as a solvent in these types of reactions because the product ends up completely dry at the end of the process and any residual monomer could be extracted prior to product recovery. DeSimone and coworkers4 5 6 showed that liquid and supercritical CO2 could be used as a solvent for the solution polymer- ization of fluorooctyl acrylate and the precipitation polymerization of fluoroolefins such as tetrafluoroethylene with extremely high conversions, high molecular weights, narrow molecular weight distributions, and high purities. These discoveries led to the commercial development by DuPont of a novel, con- tinuous polymerization process for Teflon based on this technology.7 A pilot plant is currently operating in Bladen County, North Carolina, the first of what is likely to be a long-term investment by the company in CO2-based polymerization processes. 4DeSimone, J. M.; Guan, Z.; Elsbernd, C. S. Science 1992, 257, 945-947. 5Romack, T. J.; Combes, J. R.; DeSimone, J. M. Macromolecules 1995, 28, 1724-1726. 6Romack, T. J.; DeSimone, J. M.; Treat, T. A. Macromolecules 1995, 28, 8429-8431. 7McCoy, M. Chem. & Eng. News 1999, 77(4), 11-13.
82 APPENDIX D Because of the high solubility of fluorocarbons (and siloxanes) in CO2, it is not surprising that the first commercial direct synthesis of polymers in CO2 was achieved with fluorinated matenals. Siloxanes are also fairly soluble in carbon dioxide, but most other polymeric materials are not. CO2 is an excellent solvent for high-vapor-pressure matenals, but it is not a good solvent for high-molecular- weight, polar, or inorganic matenals. To enhance the capability of CO2 to dis- solve these types of matenals, it is necessary to use cosolvents or to design sur- factants that are soluble in CO2. The recognition that high-molecular weight fluoropolymers and siloxanes are soluble in CO2 enabled the design of such sur- factants via the covalent coupling of highly CO2-philic moieties such as poly(fluorooctyl acrylates) with oleophilic or hydrophilic moieties. DeSimone and others) ~ 9 i0 have demonstrated that these CO2-soluble surfactants are able to form micellar aggregates in CO2, and that these aggregates are able to enhance significantly the solubility of high molecular weight, oleophilic, and hydrophilic species. These CO2-soluble surfactants are also able to stabilize latex particles in CO2, and they can be used to form water-in-CO2 and CO2-in-water emulsions. It is interesting that the m~cellization process in compressed CO2 is reversible. i2 At low CO2 pressures, the surfactants dissolve in Cellar form. However, as the pressure is increased, the Moselle size (aggregation number) decreases until the surfactant solubility is so high that the m~celles break up. Reduction of the pres- sure leads to reaggregation and recovery of the Cellar structure. The solubility and functionality of these surfactants depend strongly on the ratio of CO2-philic to CO2-phobic groups, their chemical structure, and the morphology of the mol- ecule (block or random copolymers). As a result, there are many ways of control- ling system properties by careful control of the chemistry as well as the process conditions. One of the more natural applications for the use of these surfactants is in cleaning applications with CO2 as the solvent. At least one commercial liquid CO2-based dry-cleaning process) is currently in the marketplace, providing an outstanding alternative to perchloroethylene, in both cleaning quality and com- patibility with garments, as well as in environmental and human health consider- ations. The ability to synthesize CO2-soluble bifunctional materials based on fluoropolymers or siloxanes has enabled many other potential applications for these novel matenal.i One of the more exciting opportunities is in the area of hMaury, E. E.; Batten, H. J.; Killian, S. K.; Menceloglu, Y. Z.; Comes, J. R.; DeSimone, J. M. Am. Chem. Soc. Div. Polym. Chem. 1993, 34(2), 664-665. 9DeSimone, J. M.; Maury, E. E.; Menceloglu, Y. Z.; McClain and, J. B.; Romack, T. J. Science 1994, 265, 356-359. 10McClain, J. B.; Betts, D. E.; Canelas, D. A.; Samulski, E. T.; DeSimone, J. M. Science 1996, 274, 2049-2052. 1lFulton, J. L; Pfund, D. M.; Capel, M.; McClain, J. B.; Romack, T. J.; Maury, E. E.; Combes, J. R.; Samulski, E. T.; DeSimone, J. M; Capel, M. Langmuir 1996, ll, 4241-4249. 12Buhler, E.; DeSimone, J. M.; Rubinstein, M. Macromolecules 1998, 31, 7347-7355.
RUBEN G. CARBONELL 83 coatings, with particular emphasis on microelectronics fabrication. This is an area in which the unique properties of CO2 can lead to both process and environmental advantages. For decades, photolithography has utilized a variety of organic and aqueous solvents to deposit, develop, and strip photoresists on silicon wafers. Currently, commercial photolithographic processes utilize chemically amplified photoresists and photoacid generators (PAGs) that are deposited from organic solvents using a spin coating technique. The patterns are then developed using an aqueous alka- line solution of tetramethylammonium hydroxide. Once the resist patterns have been etched into the underlying substrate, the photoresist is removed using a sol- vent- or water-based stripping step. Although these wet processes have served the industry well for years, as feature sizes continue to be reduced there are compel- ling reasons to reevaluate the use of conventional solvents and aqueous solutions and to implement a new "dry" technology. To enable the manufacture of feature sizes with dimensions of 130 nm and less, the microelectronics industry uses photolithographic exposure tools operat- ing at very short wavelengths of light (365 nm, 248 nm, 193 nm, and soon 157 nary). Lithography at 157 nm is particularly challenging since only a few classes of polymers have the requisite transparency at this wavelength. Thus most con- ventional polymeric materials are ill suited as 157-nm resists because the radia- tion cannot penetrate through films of the required thickness. Although most poly- meric materials have a high optical absorbance at 157 nm, highly fluorinated polymers are relatively transparent, thus establishing that future generations of 157-nm resists will need to be highly fluorinated to be useful. Fortunately, these same glassy fluoropolymers are very soluble in liquid and supercritical CO2. We are attempting to design novel polymers and processes so that we can deposit the photoresist from liquid CO2, develop the image after exposure with supercritical CO2, and strip the post-etched material with supercritical CO2 at higher pressures. Preliminary results have been obtained with a random copoly- mer of 1,1-dihydroperfluorooctyl methacrylate (FOMA) and 2-tetrahydropyranyl methacrylate (THPMA). With less than 30 mol % of the THPMA monomer, this photoresist in its fully protected form was found to be soluble in compressed liquid CO2 at moderate conditions. In order to spin-coat films using liquid CO2, the photoresist had to be soluble in liquid CO2 at liquid-vapor equilibrium condi- tions where there is a meniscus that can allow the CO2 liquid phase to spread and wet the wafer and to evaporate in a controlled manner leaving a uniform film of the photoresist on the wafer substrate. To achieve the desired solubility, all spin coating was performed under subambient temperatures (6-10 °C) that increase the density and solvating strength of CO2 nearly 40% relative to room tempera- ture. This was adequate to dissolve the photoresist at the desired liquid-vapor equilibrium conditions. Two fluorinated analogs of conventional PAGs were also designed to provide suitable solubility in liquid CO2. Polymer solutions for spin casting were prepared by dissolving 20 wt % of
84 APPENDIX D PFOMA-r-THPMA in liquid CO2. This liquid CO2-photoresist-PAG solution was maintained in a high-pressure view cell equipped with sapphire windows so that complete dissolution could be confirmed visually. The spin coating was carried out in a specially designed and built high-pressure spin coating tool.~3 i4 The pressure in the spinning chamber was maintained at precise values ranging from 5 to 15 psi below the equilibrium vapor pressure of the CO2-photoresist solution. Control of this chamber pressure was a key parameter in the determination of film quality and thickness, as the pressure differential was what governed the evapora- tion rate of the liquid CO2. Precision control of the evaporation rate, spinning speed, and system temperature were required to ensure consistent, uniform coat- ings. Extensive work was conducted to determine the ideal range of rotational speeds, pressures, and concentrations to be used to produce uniform, lithographic quality films, resulting in films with 3% variation in film thickness over the entire wafer and a root mean square (rms) roughness of 0.4-0.5 nm. The photoresist and PAG formulations were exposed and imaged at 248 and 193 nm, resulting in highly encouraging feature sizes of 0.8 ,um. Even though this technology is not yet ready for full commercial implementation, it does point the way toward the development of a new lithographic process that is totally "dry" and compatible with existing cluster tools in the industry to couple to vacuum operations. All of this is accomplished with a single environmentally friendly solvent. In other related areas of microelectronics fabrication we have also succeeded in developing a high-pressure free meniscus coating apparatus capable of depos- iting monolayer to 300 A thin films of perfluoropolyethers with great uniformity and roughness characteristics.~4 We believe that the low viscosity of CO2, coupled with its excellent wetting properties, will enable whole new classes of thin-film coating operations that will at the same time be environmentally responsible. These are likely to be important, not just for microelectronics applications but also for biomedical and nanotechnology formulations. Even though there are still many technical and economic barriers to the total acceptance of these technolo- gies, we believe that environmental pressures as well as technical requirements for pure component systems with high uniformity will over time help "dry" CO2- based processes play an increasingly important role in industrial environments. 169. i3DeSimone, J. M.; Carbonell, R. G. U.S. Patent 6,001,418; 2001. i4Hoggan, E. N.; Novick, B. J.; Carbonell, R. G.; DeSimone, J. M. Semicon. Fabtech 2002,16,
UMA CHOWDHRY 85 SMART, SUSTAINABLE GROWTH Uma Chowdhry The DuPont Company This paper describes DuPont's approach to addressing the dual challenge of achieving business growth through new products that create a high standard of living while also protecting the environment for future generations. We have made environmentally sustainable growth (Box 1) through technical innovation our primary challenge, and this ideology serves as a great motivator for our people and our businesses. DuPont's bold and uncompromising commitment to sustainable growth with the intent of integrating economic, environmental, and social factors forms the foundation for our future. This simple statement represents an enormous chal- lenge. It requires that we focus our efforts and also track our environmental im- pact. At DuPont we are measuring the way we "create shareholder value while decreasing our environmental footprint along our value chains." The term "foot- print" includes raw materials, energy, emissions and waste, as well as injuries, illnesses, and environmental incidents. We are integrating this concept into all business decisions, into all local actions at the community level, and into helping lawmakers make fact-based decisions in enacting new pieces of legislation. This is a tall order, but if we are to protect the environment for future generations, we must take on this challenge. Our core values (Box 2), which have stood the test of time, provide us with the organizational culture to take on this enormous challenge of sustainable growth. With strong leadership, we believe we can stay the course. Since the days of our company's founders, we have had a mindset of "zero injuries" to our people. We are extending that same mindset to environmental excellence and
86 APPENDIX D have set demanding stretch goals for ourselves to remain at the forefront, leading the challenge for the chemical industry, and influencing our customers and com- petitors to adopt the same goals of zero injuries to people or to the environment. As DuPont enters its third century and we reflect on the chemical industry over the last 200 years (Box 3), it is clear that we have contributed very signifi- cantly to improved standards of living and to material prosperity for people around the world through better housing, apparel, transportation, and food. The industry has created both economic and societal value over the last two centuries. The commitment to continue this pursuit now has to be accompanied by a commit- ment to protect our environment. The chemical industry's challenge lies in cleaning up contaminated water, reducing toxic air emissions, and reducing energy consumption and waste that leads to soil contamination (Box 4~. Today the most pressing issues for water contamination are heavy metals, nondegradable biologically active substances, and persistent bioaccumulative toxins. Emissions from the chemical industry have contributed to ozone-depleting materials and greenhouse gases, causing global climate change. Our land is contaminated in certain areas with toxic compounds,
UMA CHOWDHRY 87 heavy metals, and biological waste. All of these harmful consequences of the material prosperity we have enjoyed are compromising the health of present and future generations. Looking broadly at the spectrum of industries that generate 95% of water, air, or soil emissions, we see that they range from chemicals, metals and mining, food, and paper to petroleum, utilities, and equipment manufacturers. The chemi- cal industry ranks first, second, and fifth in the generation of water contaminants (Figure 1), air tonics (Figure 2), and soil contaminants (Figure 3~. It is therefore incumbent upon the chemical industry to take a strong stand on reducing the environmental footprint that it has generated and could potentially create in the future. We must progress from an attitude of just complying with federal regulations to one where we earn the public's trust and move on to sustainable development. This corresponds to a journey: from compliance, to earning the public's trust, to sustainable development. A quick review of the results over the past decade at DuPont demonstrates our commitment to sustainability (Box 5~. While production volume grew 35%, we kept energy consumption flat through innovation in our manufacturing pro- cesses. Impressive reductions were made in air carcinogens (86%) and tonics (73%), in greenhouse gas emissions (63%), and in hazardous waste generated (40%) as well as reduction in deep well disposal (82%~. Conservatively these environmental improvements have also led to over $1 billion in savings for DuPont. What is good for the environment can also be good for business. DuPont is transforming itself as it enters its third century. Using a three- pronged strategy of integrated science, knowledge intensity and "Six Sigma" pro- ductivity (see below), we are focused on sustainable growth (Box 6~.
88 60 Million sof pounds 40 20 O - APPENDIX D 50 36 23 19 Treatment, Paper Storage, Disposal & Solvent Recovery Facilities Chemicals Metals Food Mining Petroleum Electric Electrical Generating Equip. Facilities FIGURE 1 Chemical water emissions by industry. These industries contribute 95% of the total emissions (1999 data). Source: DuPont Consulting Solutions analysis of EPA Toxics Release Inventory, 1999. 300 250 Miiilon s of 200 - pounds 150 - 100 - 50 - 288 o- L 186 107 92 92 60 57 48 ~\~x ~~ ~; ikb ~ ~ ~c: FIGURE 2 Chemical air emissions by industry. These industries contribute 95% of the total emissions (1999 data). Source: DuPont Consulting Solutions analysis of EPA Toxics Release Inventory, 1999.
UMA CHOWDHRY 4,000 ~ 3,935 250 'A 200 150 100 50 o 89 258 220 49 6 4 1 Metals Electric Primary Treatment, Chemicals Paper Coal Food Stone/ Electrical Generatin Metals Storage, Mining Clay/ Equip. 9 Disposal & Glass Facilities Solvent Recovery Facilities FIGURE 3 Chemical soil contamination by industry. These industries contribute 99.6% of the total emissions (1999 data). Source: DuPont Consulting Solutions analysis of EPA Toxics Release Inventory, 1999.
9o APPENDIX D DuPont's heritage is its ability to create value with science and technology. In the twentieth century, we combined chemistry and engineering to create new materials such as nylon, Teflon, Lycra, and Kevlar that are household words to- day and have brought the world fashion, comfort, and protection. We are trans- forming ourselves by building on our heritage of innovation to begin integrating biology and chemistry into creating value. We are investing in building our capa- bility in biotechnology because we see enormous potential in using this platform of new technologies to ensure sustainable development. Another pathway to sustainability is to generate more business value but with fewer pounds of material. We have a metric called shareholder value added per pound of product produced (SVA/lb). SVA is the value created above the cost of capital. By selling high-value services coupled with our products, we can en- hance progress toward high SVA/lb. When coupled with other financial metrics, S VA/lb provides an indicator of future sustainability for different growth strate- g~es. The focus on asset productivity at DuPont is relentless. Improving yields, uptime, and throughput helps us delay capital expenses and reduce raw material usage and waste. We have adopted the Six Sigma methodology that uses statisti- cally significant, data-based understanding to reduce defects in our manufactur- ing processes. This methodology, coupled with process innovation, has led to lower raw material usage, lower energy consumption, lower emissions, and lower waste with annual pretax savings of over $1 billion over the past three years. Thus, through technological innovation, we have made significant progress to- ward creating economic and societal value while reducing our environmental foot- print (Box 7~. Several examples will illustrate the use of technical innovation to provide "greener" products and processes. These are elimination of CC14 waste during phosgene production; production of Tyvek envelopes from waste; Smart automo-
UMA CHOWDHRY 91 five finishes; biobased routes to chemicals, polymers, and fibers; and a biomass ~- reilnery. The first example deals with reduction of carbon tetrachlonde formation in the production of a hazardous chemical, phosgene (Figure 4~. Phosgene serves as an intermediate for some of our specialty fibers and agricultural intermediates. The State of New Jersey required us to reduce carcinogenic CC14, which forms as a by-product when chlorine and carbon monoxide are heated at high pressure. Our goal was to reduce CC14 formation from 500 to 100 ppm. Our approach was Required by State Regulations Laboratory Results Actual after 1 year CCI4 Reduction With DuPont Commercial Catalyst (>250,000x) ~ ~ ~ ~ ~ ~ ~ ~ /' A? / ~ 1 1 1 1 1 1 1 1 1 1 0 10 20 30 40 50 60 70 80 90 100 % CCI4 Reduction FIGURE 4 Reduction of CC14 through improved catalyst
92 APPENDIX D MnO2 ~ NCHICN NC If NH2 B23 11 ADN alginate capsule 5-CVAM Chemical · 25% conversion · 20% yield · organic solvent; heat Biobased · 97% conversion · 93% yield · aqueous solvent; mild · reduced catalyst cost by $0.20/lb of 5-CVAM reduced catalyst waste by 200X FIGURE 5 Chemical vs. big-based route to an agricultural intermediate. to find superior catalysts that, in fact, reduced CC14 formation from 500 to less than 10 ppm, exceeding regulatory requirements. We are licensing this technol- ogy to other phosgene manufacturers to ensure that this development is available not only for DuPont but for other manufacturers as well. Another example shows the ingenuity of our technical community in being able to use recycled water and milk jugs made of high-density polyethylene into Tyvek envelopes that are lighter than conventional Tyvek used for housewrap during construction (Figure 5~. The added benefit to the airlines of lower-weight cargo is the lower use of fuel and lower emission of greenhouse gases. Polyethyl- ene-based Tyvek is a non-woven material used to wrap houses under construction to improve insulation. This use of Tyvek reduces energy consumption by tenfold each year for the homeowner. In our automotive coatings business, we lowered waste by developing a coat- ing with very high-solids content. The advanced analytical tools available today allowed us to measure and control the distribution of low molecular weight mate- rial in the coating. This, combined with hybrid cross-linking, provided a thinner coating with superior durability and scratch resistance coupled with smoothness and higher gloss. Not only did we reduce waste at our manufacturing site by decreasing the use of solvents, but we were able to help our customer Daimler- Chrysler reduce its volatile organic compounds by 1.5 pounds per vehicle, haz- ardous air pollutants by 0.45 lbs per vehicle, odor emissions by 86%, and total raw material usage by 20%. Sustainable development should allow all parties along a producer's value chain to benefit from product and process improve- ments. Switching to examples of big-based solutions (Figure 5), we demonstrated the use of an enzyme a renewable resource in the production of 5-
UMA CHOWDHRY cyanovaleram~de, an agricultural intermediate 93 i1 important for citrus growers and some vegetable farmers. The results using a big-based catalyst vs. a conventional oxide catalyst show dramatic improvements. The chemical process operated at 25% conversion and 20% yield and used an organic solvent in a typical chemical reaction. The use of a specific enzyme allows 97% conversion and 98% yield and employs an aqueous solvent under mild conditions. The big-based operation also allows reduction of catalyst cost by $0.25 per pound and results in twentyfold waste reduction. This impressive example confirms that it is possible to use re- newable resources to create economic value while creating societal value and reducing our dependence on petroleum-based feedstocks. We embarked on a large program five years ago to demonstrate that we can use our technical innovation power, coupled with a partner's capability in engi- neenng enzymes, to make bulk chemicals and fibers cost effectively. Our goal was to produce a specialty fiber we call "Sorona" (Figure 6), which incorporates the attractive properties of nylon, Dacron and Lycra, resulting in superior soft- ness, vibrant color, UV and chlorine resistance, and stain resistance coupled with stretch and recovery. Market test development shows that consumers find this combination of functionalities very attractive. Chemical routes to polypropylene terephthalate-based fibers, which we have branded Sorona, use hazardous chemicals such as ethylene oxide and carbon monoxide and are subject to the environmental problems of a typical chemical process (Figure 7~. We undertook the enormous challenge of producing 1,3- propanediol (3G) from glucose in one step as shown in Figure 8. First attempts resulted in very low yield. Working with Genencor, our team of scientists and engineers has genetically modified an enzyme to gain a 120-fold improvement in yield over five years (Figure 9~. This "tour de force" of modify- ing an E. cold host to achieve high conversion and selectivity has demonstrated to us and to the world what biocatalysis can do for the chemical industry. We have scaled up the process and are currently operating a pilot plant. The impressive results we have demonstrated with bio-3G prompted the De- partment of Energy (DOE) to award a multim~llion dollar grant over four years to DuPont, Diversa, and the National Renewable Energy Laboratory (NREL) to use the biomass from our 3G process as fuel. Our target is to demonstrate the use of renewable resources (corn sugar) and simultaneous production of fuel from the biomass a "biorefinery" as an economically viable concept (Box 8~. 3GT (Sorona) HO`C,C`C,OH ,3-Propanedioi (3G) + ,,C¢C~ Terephthaiatic Acid OH it ,C~ ,C~ ,C~ Polypropylene terephthalate (3GT) FIGURE 6 Sorona (polypropylene terephthalate), an advanced polymer-fiber.
94 Shell: Degussa: ~ + CO + H2 MACHO + H2O + H2 FIGURE 7 Traditional chemical routes to 3G. In Nature: Two microorganisms convert sugar to 3G stepwise. me- HOOH HOOH For an Industrial Process: A single . . · . · . mlCrOOrganlSm IS GeSlreO FIGURE 8 Nature combined with metabolic engineering to produce 3G. 180 - 160 - 140 - 120- 100- a~ ~ 80- (D 60- co 40 - 20 - O ~ APPENDIX D OH OH / - - 1994 1995 1996 1997 1998 1999 2000 FIGURE 9 Bio-3G titer history.
UMA CHOWDHRY 95 DuPont' s commitment to sustainable development begins with a strong lead- ership commitment to stay the course through economic cycles, to set stretch goals for each decade and to drive innovation. Specific 2010 goals for energy are to · derive 25% (up from 10% today) of corporate revenue from renewable resources; · source 10% of our energy needs from renewable resources; · reduce greenhouse gases by 65% vs. 1990; and . keep total energy use flat vs. 1990. Our commitment to use our powerful technology base coupled with comple- mentary skills from universities, government labs, and/or other companies to en- sure sustainable development is steadfast. Further examples of technical programs to illustrate this commitment include fuel cells, sensors, and catalysts for control- ling auto emissions, waterborne coatings, supercritical solvents, and recyclable polymers. Major environmental trends that we see for land, air, water, and transporta- tion of environmentally hazardous materials are shown in Box 9. These trends require that we get ahead of these issues and lead the chemical industry in the reduction of toxic metal (e.g., Sb, Sn, As) compounds, greenhouse gases, mer- cury emissions, and sulfur from gasoline and diesel, and find ways to control and sequester CO2. Reduction of arsenic, as well as nitrates and ammonia, in drinking water is necessary. It is also imperative in these days of terrorism that we reduce transportation and storage of hazardous materials and continue our drive to de- velop inherently safer processes. DuPont has prospered in the last two centuries. As we enter our third century (Figure 10), we are committed to use the power of biology coupled with our
96 APPENDIX D Maturity Maturity Growth Birth Birth 1 802 1 900 2000 2100 FIGURE 10 The transition to smart, sustainable growth.
BARRY DELLINGER 97 traditional strengths in chemistry and engineering to what we hope will become the "century of sustainability." The ultimate goal of sustainability has to be con- servation of global ecosystems for the future of our planet and of humanity. We cannot settle for anything less. THE ORIGIN AND NATURE OF TOXIC COMBUSTION BY-PRODUCTS Barry Dellinger Louisiana State University Introduction Although combustion and thermal processes are necessary to provide for the essential needs of our existence, they are intrinsically "dirty" and emit a variety of air pollutants. Some of these pollutants are well known, well understood, and subject to significant control. However, combustion is a complex process that results in formation of many pollutants that are not well characterized as to their nature or origin. As a respon- sible society, it is incumbent upon us to examine these issues, determine their importance, and endeavor to eventually resolve and address each of them. Combustion-related air pollution can be classified in many ways. Table 1 presents one such classification that may assist in identifying and prioritizing research needs. We actually have a pretty good idea of the identity of most of the emissions from combustion on a mass basis; on the order of 99.9% of emissions from a reasonably controlled combustion source are carbon dioxide, carbon monoxide, and simple hydrocarbons such as methane and ethane. However, the remaining fraction is a complex myriad of pollutants that is not fully characterized and can contain toxic species. TABLE 1 Categorization of Combustion-Generated Air Pollutants. Category Examples Smog precursors Acute tonics Toxic air pollutants Endocrine disrupting chemicals Halocarbons Fine particles Persistent radicals CO NOx, volatile organic compounds (VOCs) Butadiene, Polycyclic aromatic hydrocarbons (PAHs) Dioxins, oxy-PAHs Chlorinated hydrocarbons (CHCs), brominated hydrocarbons (BHCs) Metals and organic constituents Semiquinones
98 TABLE 2 Chemical Reactions Zones in Combustion Systems. APPENDIX D Reaction Decomposition Formation Zone Conditions Mechanisms Mechanisms 1 Pre-flame T = 200-1000°C Molecular eliminations, Molecular eliminations, tr << 1 s bond fission, bimolecular complex radical-molecule [O2]~50% EA radical attack pathways, recombination- association reactions 2 Flame T = 1000-1800°C Bimolecular radical attack, Complex radical- tr ~ O.Ols bond fission, molecular molecule pathways, [O2] ~ 50% EA eliminations molecular eliminations, recombination- as sociation reactions 3 High-temperature T = 600-1100°C Molecular eliminations, thermal tr= 1-lOs bond fission, bimolecular [O2] = 50-100% EA radical attack association reactions complex radical-molecule pathways, molecular elimination 4 Gas quench T = 80-600°C Molecular eliminations, Recombination- tr ~ lOs bond fission association reactions [O2] = 3-9% EA 5 Surface catalysis T = 1000-1800°C Surface-catalyzed Surface-catalyzed tr = lOs to 10 min decomposition synthesis [O2] = 3-9% EA An analytical chemist would be interested primarily in characterizing these emissions by developing new analytical techniques and continuous monitoring apparatus. A combustion scientist might be most interested in identifying the origin and mechanism of their formation. A combustion engineer might focus on developing methods for their mitigation through design of control technology or combustion modification to prevent their formation. Table 2 presents a combustion system from the viewpoint of a combustion scientist. It identifies reaction zones, the conditions that exist within these zones, and classifications of reactions that can occur under these conditions. The vast majority of these pollutant-forming pathways involve free radicals. It is generally assumed that these radicals are formed in the high-temperature flame zone of combustion systems. However, reactions occurring in the post- flame, thermal zone (Zone 3) and the gas-quench and surface-catalysis zones (Zones 4 and 5), may also form radicals responsible for pollutant formation. In some cases, the radicals may be stable and act as pollutants themselves. Small reactive radicals (e.g., HO, H. O ~ have lifetimes of less than a micro-
BARRY DELLINGER 99 second. Organic radicals are less reactive and may have lifetimes of several mi- croseconds. Resonance-stabilized organic radicals, such as cyclopentadienyl and propargyl, can be even more stable and less reactive, with lifetimes in the milli- second range. Catalytic cycles similar to those well studied by the tropospheric chemistry community can result in measurable steady-state concentrations of radi- cals that exist for several seconds after the combustion event that initiated their formation. Finally, recent evidence suggests that semiquinone-type radicals con- tained in some types of particles may persist indefinitely. Combustion Chemistry Research Opportunities While some sources and mechanisms of combustion-generated air pollution have been the subject of considerable study, other sources are poorly character- ized and not very well understood. Examples are: 1. flares and plumes, 2. soot and PAH formation by resonance-stabilized radicals, 3. endocrine disrupting chemicals (EDCs) and oxy-PAH, 4. gas phase reactions of halogenated hydrocarbons, 5. surface-mediated pollutant formation, and 6. particulate-stabilized free radicals. Flares and Plumes Industrial flares and plumes represent a potentially significant source of air pollution that are poorly characterized and controlled. Conditions are ideal for post-flame thermal reactions and photolytic reactions at elevated temperatures (photothermal reactions). The elevated temperatures in flares and plumes (~50 to 600 °C outside of the visible flame) can result in accelerated rates of formation of oxy-PAH and nitro-PAH. At higher temperatures within the flame zone of com- bustors, oxy-PAH and nitro-PAH are likely to be destroyed, but under the rela- tively mild conditions of flares and plumes, the rates of formation can be acceler- ated without their subsequent destruction. The elevated temperatures and exposure to solar radiation can result in fast photothermal reactions that lead to the forma- tion of both combustion-type pollutants and photochemical pollutants. Research has shown that at elevated temperatures, the rate of absorption of solar radiation and photochemical quantum yield can increase up to tenfold. Soot and PAH Formation by Resonance Stabilized Radicals For many years, the reactions of small organic radicals, containing even num- bers of carbons, such as vinyl, ethynyl, and butadienyl, have dominated the theory of molecular growth to form soot and PAH. However, it has been recognized recently that odd-carbon radicals such as propargyl and cyclopentadienyl are sta-
100 APPENDIX D bilized and play a significant, possibly dominant role. The properties and elemen- tary reactions of these resonance-stabilized species are poorly characterized from the chemical viewpoint. They can undergo a variety of isomerization, recombina- tion, and addition reactions for which rates have not been determined, and their impacts on PAH formation pathways have not been assessed. EDCs and Oxy-PAH It is now known that endocrine disrupting chemicals are emitted from com- bustion sources. Interest has focused on the emissions of polychlorinated dibenzo- p-dioxins and polychlorinated dibenzofurans (PCDDIF), which are also known carcinogens. However, oxy-PAH, epoxides, and other oxygenated species are known EDCs. These can also be emitted from combustion sources, although they are not well characterized. They are semipolar compounds that are difficult to analyze. Thus, improved methods of analysis are needed in conjunction with bio- logical testing to determine the nature and quantity of EDC emissions from com- bustion sources. Gas-Phase Reactions of Halogenated Hydrocarbons Chlorinated and brominated materials are burned or thermally treated in a variety of combustion sources including hazardous and municipal waste incinera- tors, industrial processes, backyard trash burning, and accidental fires. Chlori- nated materials are used in a wide range of applications and brominated com- pounds are fire retardants used in many devices including electronic circuits. Although there has been some research on the reactions of CHCs and BHCs in the past 20 years, too little is known about their reactions considering the magni- tude of the environmental impact. Elementary reaction studies of gas-phase reac- tions of Cat and C2, CHCs, and BHCs are needed to understand their most funda- mental reaction properties. Reactions of the chlorinated and brominated benzenes and phenols are important intermediate steps in the formation of PCDDIF. Recent kinetic models indicate that the gas-phase reactions may be quite important and elementary gas-phase reaction studies have been overlooked by researchers. Surface-Mediated Pollutant Formation Research has shown that the presence of catalytic surfaces and particles in- creases the yields and rates of formation of PCDDIF in combustion systems over the reaction temperature range of 200-600 °C. Transition metals such as copper can increase the rate of chlorination, molecular growth, and aromatic condensa- tion reactions to form PCDDIF. Also, reactive species can attack a carbon matrix to chlorinate and fragment the carbon lattice-forming PCDDIF as well as other chlorinated hydrocarbons. Although research to date has focused on surface-me-
BARRY DELLINGER 101 dieted PCDD/F and chlorinated hydrocarbon formation, this same research sug- gests that many types of pollutants can be formed by similar processes. Post- combustion cool-zone formation of pollutants may explain an important combus- tion dilemma, which is how seemingly thermally fragile compounds can be emitted from combustion systems. Particle-Stabilized Free Radicals Recent research has shown that combustion sources can generate radicals that are stabilized by associated with particulate matter. This same particulate matter becomes a component of airborne PM2.5. (fine particulate matter (smaller than 2.5 microns in diameter). PM2.5 is known to initiate lung cancer and car- diopulmonary disease; however, the mechanism has not been identified. DNA and cellular assay results indicate that combustion and PM2.5 can cause radical induced damage to DNA. Based on electron paramagnetic resonance (EPR) stud- ies, the responsible species appear to be semiquinone-type radicals. These studies reveal that radicals, heretofore thought to be too unstable to survive in the atmo- sphere, can be stabilized by association with particles and initiate biological damage. Conclusions and Recommendations Research on the environmental aspects of combustion is inherently multidisciplinary. Fundamental research is needed within specific areas; instru- ment development is needed to facilitate this research; and interdisciplinary col- laborations are needed to evaluate the health impacts of combustion-generated pollution. Research Recommendations · Research on photothermal and thermal reactions in flares and plumes in- cluding photothermal chemistry and spectroscopy as well as destruction and for- mation of toxic air pollutants · Elementary reaction kinetic studies of resonance stabilized radicals and how they impact formation of PAH and PAH radicals · Mechanistic studies of the partial oxidation of PAHs by thermal, pho- tolytic, and photothermal pathways . Elementary reaction kinetic studies of CHCs and BHCs with specific em- phasis on the reactions of chlorinated phenol and other dioxin precursors; chemi- cally activated displacement reactions; C1, Br, H. O. and HO reactions; and ah initio molecular orbital calculations Research on surface-catalyzed pollutant formation including transition .
102 APPENDIX D metal catalyzed dioxin formation, surface-catalyzed formation of CHC and toxic air pollutants, and catalytic destruction Efforts to characterize fine particles including speciation of toxic metals and organometallic surface binding · Research on persistent radicals in the environment including characteriza- tion of their structure, mechanisms of formation, mechanisms of stabilization, and pathways of biological redox cycling . Instrument Development Needs · Methods for study of fast surface reactions · Surface analysis techniques for organometallic binding and radical char- acterization . Dependable methods for studying elementary gas-phase reactions of or- ganic radicals at elevated temperature . systems Methods for studying the spectroscopic properties of high-temperature · Techniques for metal speciation Recommendations for Support of Research · Development of m~croarrays for rapid screening of biological end points of complex mixtures Risk assessment methods for multiple pollutants from multiple sources Biochemical reactions of environmentally persistent free radicals COMPUTATION AND ENVIRONMENTAL SCIENCE David A. Dixon Pacific Northwest National Laboratory Environmental chemical science deals with issues of scale as much as any area of chemistry and the issues of scaling in space and time dominate environ- mental science. The goal of environmental science is to understand the current state of the environment based on our knowledge of the past and to use this infor- mation to be able to predict the future state. For example, given current practices for manufacturing, what will be their long-term environmental impact? Given potential environmental remediation strategies, what will these lead to? One does not want to use a remediation strategy that will have unforeseen consequences and introduce new environmental issues. No one wants to repeat the mistakes of the past, for example, the wide release of chlorofluorocarbons (CFCs) into the atmosphere that led to stratospheric ozone depletion. Although, we are interested
DAVID A. DIXON 103 in the results at large spatial and temporal scales, detailed insight into behavior at the molecular scale is key to understanding (1) how humans have impacted the environment, (2) how to remediate anthropogenic impacts on the environment, and (3) how to minimize future anthropogenic impacts on the environment. Computing has revolutionized the way that we live and the way that we practice science. The entire research enterprise has been undergoing a revolution over the past two decades as it exploits the advances that are occurring in com- puter hardware and software and in new mathematical and theoretical approaches. This revolution is based on the utilization of high-performance computers (now massively parallel) to solve the complex equations that describe natural phenom- ena (e.g., the Schrodinger equation for electronic motion in molecules; Newton's equations of motion for the classical motion of hundreds of thousands of particles such as those in a protein). Modeling and simulation is now considered to be the third branch of science, bridging experiment and analytical theory. The role of simulation in the modern scientific and technical endeavor cannot be underesti- mated, and the use of effective modeling and simulation plays a critical role in modern scientific advances. There are a number of important roles that modeling and simulation play in the scientific enterprise. First, modeling, theory and simu- lation can enhance our understanding of known systems. Second, they can pro- vide qualitative and quantitative insights into experimental work and guide the choice of which experimental system to study or enable the design of new sys- tems. This is most useful if the simulation has been benchmarked on well-estab- lished systems to validate the approach. Third and finally, simulations can pro- vide quantitative results to replace experiments that are too difficult, dangerous, or expensive and can extend limited experimental data into new domains of pa- rameter space. For example, accurate thermochemical and kinetic calculations for the design of nuclear waste processing facilities and green chemical processes or for predicting tropospheric oxidation processes relevant to aerosol formation are needed due to missing experimental data. In addition, simulation allows one to explore temporal and/or spatial domains that are not accessible by present experimental methods. For example, it is now possible to explore different chemi- cal reaction pathways not directly accessible by experiment to learn why they are not favorable or to find missing steps in a mechanism. High accuracy from a simulation is important. A factor of 2 to 4 in catalyst efficiency may determine whether a chemical process is economically feasible or not, and a factor of 4 in a rate constant at room temperature (25 °C) corresponds to a change in the activation energy on the order of just less than 1 kcal/mol. For a 50:50 starting mixture of two components, a change in the free energy, /\G, of less than 1.5 kcal/mol leads to a change in the equilibrium constant by a factor of 10, leading to a 90:10 mixture at 25 °C. The requirement for such accuracy means that we must be able to predict thermodynamic quantities such as bond dissocia- tion energies (De or Doff) and heats of formation HAHN to better than 1 kcal/mol and activation energies to within a few tenths of a kilocalorie per mole a daunt-
104 APPENDIX D ing computational task. Rapid advances in hardware, algorithm development, theory, and software are enabling computational scientists to attack larger and more complex problems with higher-accuracy and higher-fidelity models. Based on advances in computational science over the past two decades, we often know how to dramatically improve the quality of the simulation given sufficient com- puting resources. If we are to gain the maximum impact from simulations, one must aim for the highest possible accuracy in the simulations given the available resources, and one must continue to develop methods that can take advantage of the significantly increased computational resources to be available in the future. This latter point is critical due to the rapid evolution of computer hardware, driven mostly by the consumer industry. It has recently been shown that computational chemistry methods can pro- vide the accuracy required to reliably solve complex environmental problems but accuracy significantly increases the computational demands. Examples of how computational chemistry is being used to impact environmental science include the following: · Accurate properties prediction for radionuclides, including actinides and lanthanides, to understand their migration in the vadose zone (e.g., the Hanford site), and their chemical behavior in waste tanks (e.g., Hanford and Savannah River) such chemical reactivity information is needed for detailed subsurface and groundwater reactive transport models. · Reliable prediction of thermodynamic and kinetic properties for chemical processes (e.g., reactions of chlorinated hydrocarbons on surfaces and in aqueous systems, atmospheric oxidation of organic precursors to ozone and aerosols) as well as for designing green chemical manufacturing processes. Molecular-level studies of chemistry in solution and at interfaces, includ- ing mineral interfaces (e.g., the behavior of metal ions in aqueous solution and on metal oxide or clay surfaces for vadose zone, tank, and groundwater remediation and catalysis) a detailed understanding of redox (electron transfer chemistry) is broadly needed; studies of the interactions of biological molecules with surfaces for bioremediation are also needed and being pursued. · Reliable prediction of spectroscopic properties to aid in the interpretation of experiments for determining speciation in the environment (e.g., surface, sub- surface, groundwater) and in tanks, as well as for chemical process control. · Reliable prediction of chemical processes for carbon management includ- ing aerosol formation (organic oxidations, inorganic NH3-H2SO4-H2O chemistry, and nucleation processes) and sequestration and capture of CO2 (e.g., in geologic formations and in the ocean). · Fundamental molecular processes relevant to cell signaling pathways for biological remediation and risk assessment including protein-protein interactions and enzymatic reactions. .
DAVID A. DIXON . 105 Structural biology and functional genomics for assessing the health im- pact of environmental contaminants and for bioremediation. · Chemical processing, including tank waste processing and separation systems for tank wastes as well as sensor design, green chemical processing strategies for waste remediation, and chemi- cal and petroleum production to minimize waste streams and energy con- sumption, homogeneous and heterogeneous catalyst design (e.g., controlled oxi- dation of organics to produce intermediates for the chemical process industry or NOX or SOX emission reduction from combustion systems), and models of the behavior of waste storage systems (e.g., glasses for radio- nuclide storage in Waste Isolation Pilot Plant or Yucca Mountain) over long time periods (hundreds of thousands of years). . Developing a thorough understanding of combustion chemistry to reduce unwanted emissions (e.g., NOX abatement strategies for lean-burn engines) and to improve system performance a large number of chemical species and reactions are involved in the combustion of hydrocarbon fuels, and little is known about the highly reactive intermediates and many of the reactions. Computational Design of Catalysts: The Control of Chemical Transformation The U.S. petroleum, chemical, biochemical, and pharmaceutical industries are the world's largest producer of chemicals, ranging from "wonder" drugs to paints to cosmetics to plastics to new more efficient energy sources. The U.S. chemical industry represents 10% of all U.S. manufacturing, employing more than one million Americans. It also is one of the few industries that has possessed a favorable balance of trade. The petroleum and chemical industries contribute ~$500 billion to the gross national product of the United States These industries rely for their financial well-being on their ability to produce new products by using energy-efficient, low-cost, environmentally clean processes, with a mini- mal number of undesirable side products. Key ingredients in 90% of chemical manufacturing processes are catalysts. A catalyst's role is to make a chemical reaction that produces a desired product proceed much more efficiently than it otherwise would by changing the kinetics of the process. Catalysis and catalytic processes account for nearly 20% of the U.S. gross domestic product and nearly 20% of all industrial products. Chemical transformations in industry take a cheap feedstock (usually some type of hydrocarbon) and convert it into a higher-value product by rearranging the carbon atoms and adding functional groups to the compound. About 5 quads per year are used in the production of the top 50 chemi- cals in the United States and catalytic routes account for the production of 30 of these chemicals, consuming 3 quads. Improved catalysts can increase efficiency
106 APPENDIX D leading to reduced energy requirements, while increasing product selectivity and concomitantly decreasing wastes and emissions. A process yield improvement of only 10% would save 0.23 quad per year! In addition, production of the top 50 chemicals leads to almost 21 billion pounds of CO2 emitted to the atmosphere per year.) Improved catalysts can help reduce this carbon burden on the atmosphere. As new products become ever more sophisticated, the need to quickly develop new catalysts grows rapidly in importance. A fundamental understanding of chemical transformations is needed to enable scientists to address the grand chal- lenge of the precise control of molecular processes by using catalysts. Whereas Mother Nature is very effective at designing catalysts such as enzymes, we are decidedly less so. The most common approach to catalyst design used to be Edisonian. Try something; if it works, try to improve on the design by systematically changing the chemical nature of the catalyst; if it doesn't work, try something else. This approach is highly intensive, in terms of both time and ex- pense, and most of the time, it did not work well. In addition, catalysts developed using this approach often produced undesirable by-products, and the catalyst it- self may pose an environmental hazard. For many catalytic processes it is still unclear just how the catalyst works. A more desirable approach to catalyst design is to analyze at the molecular level exactly how catalysts function and to use this information to lead to the discovery of new systems and to optimize the design of others. Without this information, it is impossible to "tune" the catalyst to have the desired effect. For example, even the most sophisticated experimental techniques are unable to provide the details of the chemical reactions occurring at the surface of a heterogeneous catalyst or information about how to tune a homogeneous catalyst to gain a factor of 2 to 4 in performance. The coupling of theory and experiment provides the most profound in- sights into catalyst behavior enabling the design of new catalysts. A combined theory-experiment approach was used to design a new alloy catalyst for ammonia synthesis leading to the first new ammonia catalyst since Haber and Bosch's work in the early l900s that is better than iron. The development of a Lewis acidity scale based on high-level computations led directly to the design of a mixed metal catalyst for the liquid phase production of the hydrofluorocarbon refrigerant HFC-134a, the only such available liquid phase process. The computational design of practical catalysts for industrial and commer- cial applications requires the ability to predict, at the molecular level, the detailed behavior of large, complex molecules as well as solid-state materials. Although intermediate-level computations can often provide insight into how a catalyst works, the true computational design of practical catalysts for industrial and com- mercial applications will require the ability to predict accurate thermodynamic iTonkovich, A. L.Y.; Gerber, M. A. The Top 50 Commodity Chemicals: Impact of Catalytic Pro- cess Limitations on Energy, Environment and Economics, PNL-10684, Pacific Northwest National Laboratory, Richland WA, 1995, 99352.
DAVID A. DIXON 107 and kinetic results. There are enormous technical challenges in the computational design of catalysts. These include the number of different scales that must be considered, the size of the active domain, the need to treat heterogeneous struc- tures, the need to consider the effective environment in which the catalyst acts (gas phase, solid phase, solution phase, at interfaces), and the need to treat com- plex metal interactions as most catalysts involve transition or lanthanide metals. Another set of challenges is that catalysts must be designed for use in real envi- ronments such as chemical reactors. The goals for computational catalysis on next-generation computer architectures are: 50 teraflops (109 floating-point operations per second): accurate calcula- tions for realistic, isolated homogeneous catalyst model systems (<1.0 kcal/mol thermodynamics, <50% error in reaction rates). 250 teraflops: accurate calculations for realistic homogeneous catalyst model systems in solution and heterogeneous catalysts in vacuum (<1.0 kcal/mol ther- modynamics, <50% error in reaction rates). 1000 teraflops: accurate calculations for realistic homogeneous catalyst model systems in solution and heterogeneous catalysts in solution (<1.0 kcal/mol thermodynamics, <50% error in reaction rates). Success in these computational approaches will enable the long-sought goal in catalyst research of the design of efficient catalysts from first principles. Such new catalytic systems will revolutionize how we manufacture the chemicals that are such a part of everyday life, from efficient automotive fuels to polymers to cancer-fighting drugs, as well as the design of catalysts to minimize tropospheric environmental pollution from combustion energy systems. The ability to computationally design efficient new catalyst systems would lead to a revolution in the design of green chemical manufacturing processes that minimize (1) the usage of raw materials, (2) energy utilization, and (3) the environmental impact of the process and waste stream. Improved catalytic converters for combustion systems would enable the use of different fuel types as well as the development of lean-burn engines that minimize the impact of engine emissions such as NOR on the formation of tropospheric pollutants such as ozone and harmful aerosols. Fur- thermore, the development of lean-burn engines with low emissions would have a large impact on reducing CO2 emissions to the atmosphere because these en- gines burn the fuel much more efficiently. Computational Environmental Molecular Science for DOE Site Cleanup Computational molecular science is a key technology for addressing the com- plex environmental cleanup problems facing the Department of Energy' s nuclear production sites as well as other polluted sites in the nation. Production of nuclear weapons at U.S. DOE facilities across the nation over four decades has resulted in
108 APPENDIX D the interim storage of millions of gallons of highly radioactive mixed wastes in hundreds of underground tanks, extensive contamination of the soil and ground- water at thousands of sites, and hundreds of buildings that must be decontami- nated and decommissioned. The single most challenging environmental issue confronting the DOE, and perhaps the nation, is the safe and cost-effective man- agement and remediation of these wastes. Environmental impacts at these sites range from minimal (e.g., near-surface contamination with uncontaminated groundwater aquifers) to extensive (e.g., surface, vadose zone, and groundwater contamination that extends off-site). The DOE invests approximately $6 billion per year in environmental cleanup activities at its former and present production sites. Even with this large expenditure expected over a 50-year time frame, it is difficult to see how remediation of the sites will be accomplished by using cur- rently available technology. The sole use of conventional approaches to remediation and control (e.g., excavation, treatment, recovery, and disposal of residual waste for contaminated soils; pump and treat for contaminated aquifers) is cost prohibitive. Remediation strategies for DOE sites will require the combi- nation of conventional remediation approaches with the increasing effectiveness and decreased cost of emerging technologies (e.g., in situ techniques such as natural bioremediation), to meet future remediation goals within budget con- straints. Incorporation of in situ remediation technologies into overall remediation strategies has the potential for significant cost savings by leveraging physical, chemical, and biological subsurface processes to enhance the natural recovery of vadose zone and groundwater systems. However, a key ingredient in the success as well as the cost-effectiveness of remediation efforts is fundamental knowledge about the chemical properties and interactions of the wastes with their environ- ment. This information is needed in order to develop and use in situ processes as unlike ex situ processes, there is usually no easy way to effectively halt an in situ process once it has begun. Computational molecular science in combination with experimental investigations provides this fundamental understanding of the com- plex interactions of (man-made) materials with the environment. To support the development of innovative technologies for remediating various DOE sites, we need to develop reliable models to investigate the impact of the technology and the appropriate level of risk of using the technology or of doing nothing. These models of contaminant fate and transport in the subsurface have to be built on a detailed understanding of the binding and reaction of con- taminants on soil particles as well as transport and reaction in groundwater. In addition, reliable models of the direct impact of mobile contaminants on humans and the risk of proposed remediation technologies will be critical for developing the safest and most cost-effective approaches to site cleanup and for public ac- ceptance of the cleanup process and results. It is clear that such research spans the interracial regime from bare solid surfaces to complex, solution-phase surface chemistry, and covers a range of time and length scales. Thus, one important focus of such a computational science effort is an understanding of the linkages
DAVID A. DIXON 109 between different temporal and spatial scales. No matter how many details of the physics are included in the transport models, if the critical underlying physical, chemical, and biological data describing the various reactions are missing or un- reliable, the accurate predictive capability of such models will be significantly lessened. Computational molecular science can provide the thermodynamic, ki- netic, and structural properties data needed for the models and can provide data that is difficult, or at times even impossible, to obtain in the laboratory or in the field due to the cost or danger of the experiment. The calculation of interactions of chemicals, including those containing radioactive elements, with environmen- tal matrices such as soils is incredibly complex and will require tens to hundreds of sustained teraflops to begin to reliably predict the molecular interactions of chemicals with environmental systems and to provide the underlying data needed for reactive transport models. High-quality data are needed, and great care must be taken to minimize the errors in the calculated underlying data used in a sophis- ticated environmental or chemical process model so that errors in the data do not accumulate, propagate, and ultimately invalidate the macroscopic-scale model. There is a time-criticality in the need for the data that this computational molecu- lar science effort will provide. Contaminants have been released into the environ- ment over the past 50 years, and their mobility has led, and is continuing to lead, to broad-scale soil and groundwater contamination of the DOE sites. Some of the most hazardous materials in the underground tanks, in the soils and groundwater, and in the buildings contain radioactive isotopes of the ac- tinides and lanthanides. These materials pose significant radiation and other health hazards. A basic understanding of actinide and lanthanide chemistry is required to develop the new technologies needed for remediating the sites. Because of the difficulty and expense involved in conducting experiments with radioactive ma- terials, it is important to employ computational methodologies that include the effects of Einstein's theory of relativity to support accurate calculations on mo- lecular systems containing the actinides and lanthanides in order to guide the choice of experiments and to reliably extend the available experimental data into all of the regimes of interest. Such a capability would minimize the need for experimental work on radioactive materials. This capability would support many other research programs of importance to DOE, including the development of highly efficient and selective catalysts for new industrial processes. In addition, such a computational capability is needed to help offset the nation's loss of ex- pertise in the area of actinide chemistry that is so critical to the multiple missions of the DOE. Computational molecular science also can be used to aid in the design of new processes, for example, the design of new compounds for efficient separation of radionuclides, such as technetium, or the rational design of enzymes to enhance the biodegradation of organic wastes or to immobilize radionuclides. Computational molecular science not only provides needed data but also can be used to provide new insights and answer "what if?" questions raised by scientists and engineers. Improving the response time to answer such questions will dra-
110 APPENDIX D matically shorten the design and development time for new remediation strate- gies. In summary, theory can, at enormous but feasible computational expense, reliably and safely predict the chemical and physical properties of radioactive substances, often more cost-effectively than performing an experiment. The sci- ence goals in this area for next-generation computer architectures are: 50 TFlops: accurate calculations for realistic models of lanthanides and ac- tinides on complex mineral surfaces (<1.0 kcal/mol thermodynamics) to develop parameters for reactive transport models of the vadose zone 250 TFlops: accurate calculations for realistic models of lanthanides and actinides on complex mineral surfaces interacting with aqueous solutions (<1.0 kcal/mol thermodynamics) to develop parameters for reactive transport models of the vadose zone 1000 TFlops: accurate calculations for realistic models of lanthanides and actinides on complex mineral surfaces interacting with aqueous solutions (<50% error in reaction rates) to develop parameters for reactive transport models of the vadose zone Not surprisingly, the ability to predict the transport of wastes and associated contaminants within the Earth's shallow crust has become one of the most impor- tant challenges facing the DOE and is one of the most daunting scientific and computational challenges. Reliable prediction of the disposition of contaminants is critical for decisions on virtually every waste disposal option, from remediation technologies such as in situ bioremediation to evaluations of the safety of nuclear waste repositories. Nuclear waste repository designs, for both high-level and low- level wastes, require knowledge of transport of fluids in the vadose and saturated zones. Such predictive capabilities also will impact the responsible utilization of our nation's energy resources, particularly with respect to oil and gas. Any such predictions must be based on a comprehensive understanding of processes in the Earth's shallow crust. The accurate computation of the transport and disposition over long spatial and temporal scales of fluid-borne contaminants in complex natural systems is well beyond current capabilities. We currently lack the scien- tific understanding and the computational resources necessary to address these issues with the desired accuracy. However, with future generations of computa- tional resources, it will be possible to address scientific and technical issues that are fundamental for quantitative analysis of fluid transport in terrestrial systems and essential for improving predictive capabilities. Computation is essential for predicting the intricate web of species, reac- tions, and interconnections of transport in the natural system. The few natural examples of the integrated record of these processes constitute experiments that cannot be offered as an alternative to computational approaches. In addition, com- putational simulations are the only means that we have for predicting how con- taminants will move over a long time period to predict the future state and the
DAVID A. DIXON 111 potential long-term impact of proposed remediation strategies. Computational modeling is the essential tool for translating scientific breakthroughs into practi- cal applications in the area of subsurface transport. There are two key challenges for predictive modeling of subsurface transport. First, the present generation of hardware and software is insufficient to address the comprehensive treatment of flow and transport of multiple fluid and solid phases in heterogeneous, three- dimensional, nonequilibrium, interactive systems such as those found in the natu- ral state. Second, as discussed above, accurate theoretical descriptions of the mechanisms of interaction are not yet fully developed. Overarching issues con- cern the ability to scale processes over many orders of magnitude in length and time, the ability to handle coupled complex relationships of processes at multiple scales, and the ability to predict and evaluate the results of processes over long time scales using data available only for narrow windows of observation. Over the next 10 years, simulation will play a major role in the development of new theory for field-scale simulations. A hierarchy of models designed for different scales is necessary to identify key features and processes at fundamental scales that have to be propagated to the field scale. Key scales where integration is needed include i, the microbial membrane and surface where molecular-level structure and binding reactions are important to the biogeochemistry; 2. the mineral surface in solution to provide the setting for the interaction of minerals, aqueous components, surface-complexed metals, and attached micro- bial populations; 3. a network of pores to put multiple mineral surfaces in contact with trans- porting fluids that move nonuniformly, allowing for varying rates of exposure to nutrients and reacting components; 4. at the Darcy scale (macroscopic scale for fluid volumes on the order of milliliters), similar degrees of process interaction are necessary, but the process models are based on bulk parameterizations to account for behaviors that cannot be resolved by the continuum approach; and 5. at the field scale (megascopic scale) process integration is similar in ap- proach to the Darcy scale but generally considers more range in the parameter space (e.g., heterogeneity) but less detail in spatial and temporal resolution. Subsurface simulations have actually begun to impact geochemistry, geo- physics, and environmental chemistry by encouraging holistic and mechanisti- cally detailed investigations where multiple processes and multiple reactive com- ponents are considered and monitored in complex subsurface materials. Over the next 10 years, these simulations are expected to be indispensable to the advance- ment of subsurface science through the systematic upscaling of processes at fun- damental scales to representative parameterizations useful to field-scale models.
2 APPENDIX D Atmospheric Chemistry Problems in atmospheric chemistry are extremely challenging compu- tationally. They are highly nonlinear, require the ability to simulate processes occurring on a vast range of spatial and temporal scales, and require coupled simulation of complex systems (the ocean and atmosphere) involving dynamics, thermodynamics (including condensable gases), chemistry, and thermal radia- tion. They are compute-intensive, data-intensive, memory-intensive, and analy- sis-intensive. One area in which computational chemistry has played an impor- tant role in atmospheric chemistry is in the development of alternatives to the chlorofluorocarbons involved in stratospheric ozone depletion. Computational chemistry was essential for calculating missing thermodynamic data as well as correcting older experimental data on the CFCs, hydrochlorofluorocarbon (HCFCs), and hydrofluorocarbons (HFCs). Indeed, there are now far more calcu- lated values being used than experimental ones. Calculations were critical to un- derstanding the stability of an alternative in the environment and to the design of production processes and the actual refrigeration system. The calculated thermo- dynamic and kinetic properties have been used extensively in the design of chemi- cal plants for producing CFC alternatives as well as by researchers who are devel- oping new catalysts and alternatives. From calculations, it is now possible to reliably predict the thermodynamics and kinetics of the atmospheric degradation processes for the CFC replacements that are so critical to the design of environ- mentally safe alternatives. The ability to reliably predict rate constants for the initial reaction of an HFC or HCFC with a hydroxyl radical and the ability to then develop reliable degradation mechanism parameters was important in understand- ing and predicting the behavior of the alternatives in the troposphere and strato- sphere in terms of ozone depletion potential, and together with calculated infra- red intensities, a compound's global warming potential could be calculated. The computational chemistry results could then be used in much larger-scale atmo- spheric models involving fluid flow. DuPont used the results from computational chemistry to improve its process design for the manufacture of new fluorochemi- cals. By using thermodynamic and kinetic parameters calculated from first prin- ciples, the behavior of a pilot plant for the production of a CFC alternative was modeled correctly before actual operation began. Computational work was criti- cal to getting the CFCs replacements to market as soon as possible, helping to save the stratospheric ozone layer. Computational chemistry is continuing to be used in many areas of atmospheric chemistry as we improve our knowledge of how the atmosphere behaves. Modeling aerosol formation, attachment of chemi- cals to aerosols, and reactions on aerosols are very difficult computational prob- lems that are just beginning to be addressed. Atmospheric models are extremely complex especially when the correct physics (e.g., radiative transport and im- proved cloud parameterizations) is included, and they represent another computa- tional grand challenge crossing many scales in space and time.
WILLIAM H. FARLAND 113 Summary As discussed above, computational chemistry can play a key role in advanc- ing the scientific enterprise. It can provide the data input for many larger, more complex models and provide us with unique insights into molecular behavior so that we can design and construct new molecules for specific tasks. Computational chemistry has become an established tool in the chemist's toolbox and is being used in broad areas of chemistry to replace experimental measurements and to provide us with improved understanding of molecular behavior. Computation will be the major tool that enables us to cross the many temporal and spatial scales that characterize environmental science. CHEMICALLY RELATED R&D AT THE EPA'S OFFICE OF RESEARCH AND DEVELOPMENT William H. Farland U.S. Environmental Protection Agency This presentation will provide background information on programs at the Office of Research and Development (ORD) at the U.S. Environmental Protec- tion Agency. EPA's mission is broadly designed around protecting human health and safeguarding the natural environment aims that require the application of chemistry, biology, epidemiology, physical sciences, and engineering in an inte- grated fashion. The role of ORD is to provide scientific foundations to assist the EPA's work by: · conducting research and development to identify, understand, and solve current and future environmental problems; · providing responsive technical support to EPA's programs and regions; · collaborating with our scientific partners in academia and other agencies, state and tribal governments, private sector organizations, and nations; and · exercising leadership in addressing emerging environmental issues and advancing the science and technology of risk assessment and risk management. ORD is a relatively large organization, with three national laboratories and three national centers. In early November 2002, we announced a new center for homeland security research in Cincinnati, and EPA has the lead on drinking water infrastructure issues for the security of the nation's drinking water system. We also played a significant role in the decontamination of congressional buildings that suffered anthrax exposure in 2001. Overall, ORD has about 1900 employees in R&D, about 1200 of whom have graduate degrees in science. In addition to its role in both health and ecological research, EPA also has responsibility for research in pollution prevention and new technology activities.
4 APPENDIX D Although there are other national institutes and centers that work in this area, only EPA has a fully integrated, multidisciplinary, problem-directed research pro- gram. We rely on the peer review process and the scientific community to focus our program as we push the envelope for being state of the art. At the same time we have a responsibility to evaluate the broader scientific literature and collect additional information to ensure that science is credibly used in EPA' s decisions. This is accomplished through a combination of our in-house research program and an external grants program. Sound science is essential for EPA decisions: . Science is a critical component of credible decisions and actions that pro- tect human health and the environment. . Making EPA decisions with sound science requires relevant, high-qual- ity, cutting-edge research in human health, ecology, pollution control and preven- tion, and socioeconomics; proper characterization of scientific findings; and ap- propriate use of science in the decision process. . ORD is a leader in environmental research, focusing its efforts and re- sources on those areas in which EPA can add the most value to reducing uncer- tainty in risk assessments and enhancing environmental risk management. ~ appreciated the discussion during the workshop session on the MTBE is- sue. A decade ago at EPA, we were laying out research needs for oxygenates, and we were concerned . We strongly suspected that MTBE would be a problem in groundwater. The discussion at this workshop shows that the concern is still with us. This clearly illustrates the importance of life-cycle analysis when we evaluate substances that are introduced widely into the environment. Many of the decisions made by EPA are risk based and have significant uncertainties associated with them. The only way we can make better decisions is to fill the knowledge gaps that lead to the uncertainties. For the scientific basis that we provide for regulatory decision making, the research falls into two gen- eral categories, problem-driven research and core research, where the former in- cludes topics such as . fine particulates in air, · drinking water contaminants, · diesel engine emissions, and · mercury in air and water. The core research includes topics such as · ecological monitoring and assessment of ecological resources, · health risks to sensitive populations, · pollution prevention and green chemistry, and · environmental economics.
WILLIAM H. FARLAND 115 These topics illustrate the need to bring a multidisciplinary approach to prob- lem-driven research. The core research program is closer to what might be called basic research, and it provides a forward-looking approach to dealing with eco- logical issues, health risks, pollution prevention, and support for environmental economics that eventually will allow us to anticipate new problems at an early stage. The interaction between problem-driven and core research is reciprocal and iterative. Chemistry will play a large role in the research areas of our own strategic plan. For example, work on particulate matter will allow us to understand the nature of the particles and their behavior in the atmosphere, develop the modeling that will predict their fate and transport along with the resulting human exposure, and understand the transition from exposure to dose that will enable health as- sessment work. Drinking water provides another example in which the by-products of disin- fection illustrate ways that environmental chemistry is part of our everyday ac- tivities. We have also been looking at the nature of the toxin associated with Pfiesteria in order to understand the chemistry by which Pfiesteria produces le- sions in fish and potentially has the ability to poison humans. This is a very difficult chemistry problem. Similarly, endocrine disrupters have been discussed in this workshop, and we need to understand their mechanism of action and the structure-activity relationships that would allow us to make a judgment about interaction with specific types of cellular receptors. Clearly, research in pollution prevention and new technologies are heavily based in the chemical sciences. A more focused list of ORD research interests includes a large number of topics that came up in today's workshop discussions: Particulate matter Drinking water Global change Endocrine disrupters Ecological risk Human health Pollution prevention and new technologies As one example of endocrine disrupter work, EPA has been involved in di- oxin reassessment for the last 11 years. This is an incredibly complex issue, with a huge amount of available information from the chemistry of toxicity equiva- lents and understanding how dioxin-like congeners can play into the total load in the environment to an understanding of how dioxin works to affect cells essen- tially as an environmental hormone. - Persistent biocumulative toxic chemicals were discussed in the workshop as one of the top opportunities for advances in chemistry understanding that these hydrophobic chemicals have the ability to persist and move around in the envi-
116 APPENDIX D ronment and are particularly difficult for us to control and remediate. This is another of the areas in which we work on a day-to-day basis, and again chemistry plays a large role. In atmospheric chemistry, a particular focus of our R&D activities is the development of air quality models and scaling to models that can predict local, regional, and global impacts on air quality from sources that we have the ability to control. A new generation of air-quality models looks at community modeling of air quality (CMAQj, and this has provided a real advance for some of the regulatory activities that go on in EPA's program. Mercury has recently been added to the model, and certainly the biggest improvement will be addition of the NOx module, which is currently under development. This will be particularly important if we move forward with a multipollutant approach to air pollution using the Clear Skies initiative.) Understanding the interactions among mercury, NOx, SOx, and other chemicals will be particularly important for us. ORD has supported work in Riverside, California to build chambers in which chemical reactivity studies will provide input into these models and advance our under- standing of this issue. Our programs include several areas that have a focus on analytical chemis- try. For example, the Clean Air Act requires us to publish Federal Reference Methods for how one measures particulate matter. We have the responsibility for looking at methods that are functionally equivalent to the Federal Reference Meth- ods so we have to be in a position to look at new analytical approaches. As an- other example, genomics has become an important way to look at molds as in- door air pollutants or microbial contamination on beaches. We need to develop markers that can be used for rapid detection and characterization of beach con- tamination rather than growing cultures that might tell you tomorrow where you shouldn't have gone swimming yesterday. These applications combine genomics and chemistry, and they have important implications for children's health, and for determining the ability to inhabit residences after flooding or wa- ter damage from fires. This approach will allow us to identify the organisms, understand the remediation techniques that must be used, and evaluate the extent of cleanup. It is an important contribution from our laboratories. Another analytical topic of concern to us is continuous monitoring. We're interested in finding ways of continuous monitoring for dioxins, mercury, and combustion by-products. Some of this work is also taking place in our laboratory in Las Vegas using pattern recognition from a continuous gas chromatography- mass spectrometry (GC-MS) analysis of water sources to discover changes, char- acterize the materials responsible for the change, and ultimately, respond quickly by diverting them from drinking water. We have several hundred chemists on our research and development staff in- http://www. epa. gov/cl ea. rskies/
WILLIAM H. FARLAND 117 house primarily in our National Exposure Research Laboratory and our National Risk Management Laboratory. These individuals are, for the most part, research chemists who work on the kinds of problems I have described. In 1995 we developed an approach to reaching out to the scientific commu- nity and investing funds in particularly important problems where we can engage some of the top scientists in the country. About a quarter of the academic faculty involved in this meeting are currently funded by our Science To Achieve Results (STAR) program, which provides about $100 million a year in support of basic research directed toward specific problems. STAR-funded research supports a broad range of issues that will have important payoffs for our future approaches to dealing with environmental problems. Some of the STAR solicitations that have a chemistry emphasis include: · Measurement, Modeling and Analysis Methods for Airborne Carbon- aceous Fine PM (2003) · Development of High-Throughput Screening Approaches for Prioritizing Chemical for the EDC Screening Program (2003) · Technology for a Sustainable Environment (National Science Founda- tion) (2001, 2003) · Assessing the Consequences of Global Change for Air Quality: Sensitiv- ity of U.S. Air Quality to Climate Change and Future Global Impacts (2002) Nutrient Science for Improved Watershed Management Program (2002) Environmental Futures Research in Nanoscale Science, Engineering and Technology (2001, 2002) . Mercury: Transport, Transformation, and Fate in the Atmosphere (2001) In a collaborative program called Technologies for a Sustainable Environ- ment, the National Science Foundation (NSF) and EPA have awarded approxi- mately $46 million for 164 projects addressing · environmentally benign solvents, · biotechnology for pollution prevention, · green chemistry-reaction modifications, · green engineering-process modifications, and · industrial ecology-environmentally benign manufacturing. Pollution-prevention research is an overarching activity for which we want a combined in-house and extramural program that is focused on finding ways to eliminate or minimize hazardous solvents, emissions, and waste. In other words, we want to clean up processes to reduce impacts on the environment; to examine renewable feedstocks as a process improvement; and to minimize water, energy, and materials use in industrial processes. The work that EPA and NSF have funded on pollution prevention has been dominated by the chemical engineering
8 Other Eng Mechanical/ Industrial/ Civil & Env Eng Chemistry Other Science Biological Eng Chemical Eng FIGURE 1 Chemical sciences dominate pollution-prevention grants portfolio. APPENDIX D and chemistry community (Figure 1~. We hope that there will be a continued commitment in the areas of pollution prevention, greener chemistry, and greener technologies. I think this is a success story that illustrates how problem-directed research in these disciplines will be important to us for the future. I agree with many of the workshop participants that we have an opportunity in terms of nanotechnology. Certainly there are indications that we can improve environmental sensing, treatment, and remediation if we can improve manufac- turing processing efficiency, reduce waste production and toxicity, and reduce materials consumption. What are the effects of nanomaterials on the environ- ment? · Currently, little is known about the potential effects of manufactured nanoparticles on human health and the environment. · Nanomaterials may enter the food chain and human body when released into the air or water or discarded on the land. · Nanomaterials could affect human health and the environment through Exposure to skin, Adsorption by the lungs, and B. lo -uptake and bio accumulation . The toxicity of nanomaterials is largely unknown. . All of this is chemistry and chemical engineering at its finest if we can find a way to make it work at a nanoscale level. This is the real challenge for us. We
JANET G. HERING 119 participated in an interagency National Nanotechnology Initiative that will invest about $710 million in this particular area this year. The vast majority of this work comes out of NSF, the Department of Defense (DOD), and DOE. But EPA in- vests about $5 million a year right now in this area. We are making an important contribution in understanding what needs to be done in integrating the issues with environmental problems. In summary, ORD has a dynamic program of research and development that really is based on integration of disciplines. We are dealing with problems for which solutions necessitate an integration of various science and engineering dis- ciplines. We have a balance between problem-driven and core or basic research that requires us to really build a strong relationship with our partners, both inside the agency and out in the general scientific community, to meet the agency's needs, and we hope to be able to continue to enhance that partnership. We are in a position to anticipate some of the emerging scientific issues and, hopefully, be poised to meet the challenges of the twenty-first century. BIOGEOCHEMICAL CONTROLS ON THE OCCURRENCE AND MOBILITY OF TRACE METALS IN GROUNDWATER Janet G. Hering California Institute of Technology Introduction Clean water has been called the "oil of twenty-first century," a phrase that reflects both the growing demand for water resources and the recognition that the quality of many water resources has been degraded by human activities. The importance of groundwater as a water supply in developed countries, such as the United States, is often overlooked, yet in 1995, 46% of the domestic water supply was provided by groundwater and 54% by surface water (Table 1~. Although groundwater is particularly important for self-supply in rural areas, it is also an important resource for public water systems. Demand for and dependence on groundwater supplies are expected to increase with increasing population in the United States, particularly in the West. There are some important differences between groundwater and surface wa- ter with respect to water quality. Surface water is considerably more vulnerable to pathogens; groundwater provides some level of "natural protection" against bac- teria and viruses that can cause outbreaks of human disease. The composition of groundwater, however, is more influenced than that of surface water by contact with soil and aquifer minerals, which generally leads to the accumulation in groundwater of constituents derived from geologic materials.
120 APPENDIX D TABLE 1 Use of Surface Water vs. Groundwater for U.S. Domestic Supply in 1995 Million Gallons per day (MOD) Million Cubic Meters per Day (m3/d) Surface water Self-supply 38 0.14 Public supply 14,100 53.4 Total 14,100 53.4 Groundwater Self-supply 3,350 12.7 Public supply 8,460 32.0 Total 11,800 44.7 SOURCE: Data from http://water.usgs.gov/watuse/pdfl 995/html. The natural occurrence of arsenic at elevated concentrations in groundwater in West Bengal, India, Bangladesh, and other parts of South Asia provides an unfortunate example of the potentially devastating effects of groundwater quality on human health.) In these areas, huge shifts in water resource utilization, from surface water to groundwater, have occurred over the past 30 years. Now tens of millions of tubewells providing drinking water for individual families and larger wells are used for groundwater-based irrigation. Arsenic occurs in the ground- water in these regions at concentrations up to the milligram-per-liter range, and concentrations between 200 and 800 ,uglL are common. The human health effects of consuming this arsenic-contaminated water range from skin lesions to fatal cancers. At the same time, it must be recognized that groundwater offers protec- tion from pathogens diarrhea! diseases are a major cause of death in infants in the developing world. Furthermore, groundwater-based irrigation has allowed substantial expansion of agricultural activities and has drastically improved the nutritional status of people in these areas. Nonetheless, the lack of attention to groundwater quality has had profound human health consequences, and these experiences illustrate the importance of evaluating the quality as well as the quan- tity of groundwater resources. Properties of Aquifers Influencing Groundwater Quality If we are to consider how the chemical composition of water withdrawn from a specific well has evolved, a key issue is the origin of the water and the contact iSmith, A. H.; Eingas, E. O.; Rahman. M. Contamination of drinking-water by arsenic in Bangladesh: A public health emergency. Bull. World Health Organ. 2000, 78(9):1093-1103.
JANET G. HERING 121 it has had with soil and aquifer minerals between the sites of recharge and with- drawal.2 A distinction should therefore be made between wells drilled into un- confined and confined aquifers. In unconfined aquifers, the water table is in con- tact with the unsaturated zone and is subject to local recharge; groundwater ages can be less than decades. In confined aquifers, however, a confining layer (usu- ally clay) prevents local recharge, which must therefore occur in upland areas that can be distant from the withdrawal site. Residence times can vary greatly and groundwaters can be tens of thousands of years old. Within an aquifer, water flows in response to the hydraulic gradient (i.e., from locations of higher to lower hydraulic head). However, flow velocities are also determined by properties of the aquifer materials, specifically hydraulic con- ductivity and porosity, as shown in the following equation: Average linear velocity _ K dh ~ dL dh where K= hydraulic conductivity, dL = hydraulic gradient, and ~ = porosity. For various aquifer minerals, porosity varies over a fairly narrow range (ca. 0.3 to 0.5) but hydraulic conductivity varies over many orders of magnitude.2 Even for a specific type of aquifer material, ranges of 1-4 orders of magnitude are common (e.g., 1O-85 to 10= m/s for fractured rock, 1O-5 to 1O-3 m/s for well- sorted sand). The lowest hydraulic conductivities are found for crystalline rock (10-~4 to 10-~° m/s) and the highest for well-sorted gravel (10-2 to 1 m/s) and clean sand or cavernous limestone (10-6 to 10-2 m/s). The prediction of groundwater flow is complicated by the heterogeneities of the subsurface environment, which occur on multiple scales. To understand the evolution of groundwater composition, it is also important to distinguish between flow through porous media and flow through channels (caused by dissolution of the rock matrix) or fractures (resulting from tectonic activity). Such differences in texture can result in differences in the contact between the fluid and mineral surfaces. Biogeochemical Processes Affecting Groundwater Composition In the physical context of water movement in subsurface and water-rock contact, we may then consider the biogeochemical processes that can affect the distribution of a constituent X between the immobile and mobile phases in the subsurface (Table 2~. Obviously, mobile constituents are of the most direct con- cern because of the potential for human exposure. Constituents in the mobile 2Langmuir, D. Aqueous Environmental Geochemistry; Prentice Hall: Upper Saddle River, NJ 1997.
22 APPENDIX D TABLE 2 Partitioning of Constituent X Between Immobile and Mobile Phases Immobile Phase Processa Mobile Phase X as component of Dissolution ~ aquifer mineral matrix ~ Precipitation X sorbed on surfaces of Desorption ~ ~ Adsorption aquifer materials X dissolved in immobile ~ Diffusion water trapped in micropores Dissolved X Dissolved X Dissolved X X sorbed-precipitated in Detachment-peptization ~ X sorbed-precipitated in mobile attached colloids ~ Attachment- filtration colloids aBoth precipitation-dissolution and sorption-desorption can be strongly influenced by microbial pro- cesses. Microbial redox transformations of constituent X may (especially for metals) significantly alter solubility, and microbial production of ligands can promote the release of X from immobile phases by dissolution and desorption and stabilize X in solution. phase (i.e., as dissolved or colloidal species) can migrate from their source region to a drinking water well. However, it must be remembered that immobile con- stituents may have the potential to be mobilized if conditions in the subsurface environment change. As indicated in Table 2, the type of process by which constituent X may be mobilized (or sequestered) depends on speciation of X in the solid phase. Thus, if X is an integral constituent of an aquifer mineral, dissolution of the mineral will be required to release X into solution, whereas mineral dissolution would not be required to mobilize X if it is sorbed onto a mineral surface. Environmental con- ditions (including both chemical composition of pore fluids and microbial activ- ity) will influence the extent of mobilization or sequestration of X. For metals, in particular, both the solubility and the affinity for surfaces can be strongly influ- enced by redox conditions and the presence of (biogenic) complexing agents. In order to gain insight into the evolution of groundwater composition, these biogeochemical processes must be examined, and subsurface materials character- ized, over multiple scales ranging from nanometers to kilometers (Figure 1~. Each level of investigation in Figure 1 corresponds to different types of process inves- tigation or material characterization. For example, at the nanoscale, x-ray absorp- tion spectroscopy (XAS) has been a powerful tool to characterize the local (i.e., coordination) environments of a range of elements in association with solid sur- faces.3 At the laboratory scale, macroscopic experiments have determined disso- 3 Brown, G. E.; Parks, G. A. Int. Geol. Rev. 1992, 43(11), 963-1073.
JANET G. HERING Field Observations Field or"Mesocosm" Experiments Laboratory Experiments Nanosca~e Characterization ]23 FIGURE 1 Interrogation of multiscale subsurface biogeochemical processes. lution rates for a wide variety of minerals as well as their properties as sorbents for both inorganic and organic chemical species.4 The effects of factors such as solution composition, the type and surface area of the solid, and biological activ- ity on both dissolution and sorption processes have been studied at this scale. Field- or mesoscale experiments include investigations conducted at the U.S. Geological Survey (USGS) Cape Cod Toxic Substances Hydrology Research Site, which is intensively instrumented with multilevel samplers and where subsurface conditions can be manipulated by injection of hundreds of liters of water with modified composition.5 Similar manipulations have been performed as a "push- pull" experiment at an individual well.6 In field observations, evolution of 4Stumm, W. Chemistry of the Solid-Water Interface; Wiley-Interscience: New York, 1992. 5USGS. Cape Cod Toxic Substances Hydrology Research Site. U.S. Geological Survey, 2002; http://ma. water. usgs. gov/Cape Cod Toxics/. 6 Harvey, C. F.; Swartz, C. H.; Badruzzaman, A. B. M.; Keon-Blute, N.; Yu, W.; All, M. A.; Jay, J.; Beckie, R.; Niedan, V.; Brabander, D.; Gates, P. M.; Ashfaque, K. N.; Islam, S.; Hemond, H. F.; Ahmed. M. F. Science 2002, 298(5598), 1602-1606.
124 APPENDIX D groundwater quality can be related to aquifer characteristics and attributed to processes occurring within the aquifer.7 Linking these different scales presents a substantial challenge. A critical is- sue is to determine how information acquired at the nano- or molecular scale can be applied to elucidate macroscale processes in the environment. Modeling and simulation provide an avenue for the linkage of biogeochem~cal processes across multiple scales. Ab initio calculations and molecular dynamics simulations are powerful tools for linking nano- and molecular-scale observations to macroscopic behavior on the laboratory scale.8 A variety of transport codes have been devel- oped to model flow in the subsurface. Although the level of biogeochem~cal so- phistication incorporated in these codes vanes widely, a few do incorporate bio- geochemical kinetics as well as equilibrium reactions. The USGS code PHREEQC9 combines reaction kinetics with one-dimensional transport, and the code HydroBioGeoChem 123D includes both three-dimensional flow and trans- port in the vadose (i.e., unsaturated) zone.~° A key issue is to understand what information about the subsurface environment and the operative biogeochem~cal processes is needed to constrain calculations made using these codes so that they have real predictive value. Case Study: Arsenic Mobilization in Sediments As already mentioned, the mobilization of arsenic provides an example of the enormous impact that biogeochem~cal processes can have on groundwater quality and human health. In the Ganges-Brahmaputra delta (i.e., Bangladesh and West Bengal, India), arsenic mobilization has been attributed to the reductive dissolution of iron oxides and the release of arsenic associated with these carrier phases. Recent manipulation experiments have shown enhanced release of ar- senic with the injection of molasses (as an organic substrate for microbial iron reduction) and arsenic sequestration with the injection of nitrate (as an electron acceptor for microbial iron oxidation).6 Possible sources of organic matter whose oxidation supports iron reduction and the elevated arsenic concentrations ob- 2014. 7Plummer, L. N.; Busby, J. F.; Lee, R. W.; Hanshaw, B. B. Water Resour. Res. 1990, 26(9), 1981- ~Molecular Modeling Theory: Applications in the Geosciences; Cygan, R. T.; Kubicki, J. D., Eds.; Reviews in Mineralogy and Geochemistry, vol. 42; Mineralogical Society of America: Washington, DC, 2001. 9USGS. PHREEQC (version 2): a computer program for speciation, batch-reaction, one-dimen- sional transport, and inverse geochemical calculations. U.S. Geological Survey, 2002; http:// wwwbrr.cr. usgs.gov/projects/GWC_coupled/phreeqc/index.html. i°Gwo, J. P.; Frenzel, H.; D'Azevedo, E.; Hoffman, F. M. Hydrobiogeochem. 1999, 123D, v. 1.1, Oak Ridge National Laboratory. iiNickson, R.; McArthur, J.; Burgess, W.; Ahmed, K. M.; Ravenscroft, P.; Rahman, M. Nature 1998, 395, 338.
JANET G. HERING 125 served in these aquifers include buried peat depositsi2 i3 and newly infiltrated organic carbon derived from agricultural activities.6 The coupled biogeochem~cal cycling of iron and arsenic has also recently been studied in a system dominated by the addition of iron as an engineering practice.~4 Since 1996, ferric chloride has been added as a coagulant to the Los Angeles Aqueduct in order to control the levels of naturally occurring arsenic that reach the water distribution system for the City of Los Angeles. The iron- and arsenic-nch floe formed by this process is deposited in North Ha~wee Reservoir. Sediments and porewaters from this reservoir were examined to determine the extent to which arsenic and iron are remobilized in the sediments and to probe the speciation of arsenic in the solid phase and its possible effects on arsenic remobilization. Arsenic and iron concentrations in the sediment porewaters were found to be closely correlated. Although these concentrations are considerably elevated at depth in the sediment column reaching 17 ~ (1.3 mglL) for arsenic and 1.6 mM (90 mglL) for iron only a small fraction of the iron and arsenic deposited to the sediments needs to be remobilized to support these concentrations. XAS analy- sis of the sediments indicated that arsenic in the solid phase is reduced from As(V) to As(III) above the depth at which arsenic is released into the porewater. Iron in solid phase remains as Fe(III). XAS analysis showed no evidence of con- version to magnetite (though conversion of fernhydnte to goethite could not be excluded). Sequential extractions indicated that most of the arsenic can be re- leased from sediment by treatment with magnesium chloride or phosphate solu- tions; this treatment does not release iron, behavior that is consistent with sorp- tion as a mode of association for the majority of the arsenic with the sediment. The remainder of the arsenic is released, along with almost all the iron, by treat- ment with hydrochloric acid.~5 These observations indicate that reduction of As(V) to As(III) does not, in itself, result in the mobilization of arsenic. This conclusion is supported by labo- ratory adsorption studies showing similar affinities of As(III) and As(V) for hy- drous ferric oxide, goethite, and magnetite. However, outstanding questions remain regarding the factors that control the rate and extent of the reductive dis- solution of iron in these sediments and whether the arsenic (and iron) that is released into the porewater is (reabsorbed onto the residual iron oxyhydroxides in i2Nickson, R. T.; McArthur, J. M.; Ravenscroft, P.; Burgess, W. G.; Ahmed, K. M. Appl. Geochem. 2000, 15, 403-413. i3McArthur, J. M.; Ravenscroft, P.; Safiulla, S.; Thirlwall, M. F. Water Resour. Res. 2001, 37(1), 109-1 17. i4Kneebone, P. E.; O'Day, P. A.; Jones, N.; Hering J. G. Environ. Sci. Technol. 2002, 36, 381-386. i5Dempsey, D. Arsenic Cycling in Natural Sediments: Effects of Aging, SURF report: California Institute of Technology; Pasadena, CA, 2002. i6Dixit, S.; Hering, J. G. Environ. Sci. Technol., 2003, submitted.
26 APPENDIX D the sediment. These are the types of questions that must be answered in order to predict the concentration of arsenic in the porewater and its response to possible changes in environmental conditions. Opportunities, Challenges, and Needs Groundwater is an important resource and one that will be critical in meeting future needs for water for human consumption, irrigation, and industrial uses. Groundwater quality is an important (though sometimes underappreciated) issue because naturally occurring groundwater constituents can adversely affect human health. Although arsenic has been highlighted here, other trace constituents of geologic materials, such as uranium and radon, also occur naturally in groundwa- ter and have known adverse health effects. Chromium(VI) has recently been iden- tified as a natural constituent of groundwater in California though its health ef- fects are less well understood. Aesthetic problems are posed by some common groundwater constituents such as iron, manganese, and sulfide. In addition, groundwater quality can be degraded by introduction of con- taminants from a wide range of human activities. These include increased salin- ity, nitrate, pesticides, and pathogens from agriculture; pathogens from septic systems; fuel hydrocarbons and oxygenates (e.g., MTBE) from underground stor- age tanks; chlorinated solvents, metals, and perchlorate from industrial manufac- turing; metals from mining, ore processing, and refining; and radionuclides from nuclear weapons and energy production. Because of the long residence time of groundwater and the inaccessibility and/or physical extent of sources of contami- nation in the subsurface, groundwater remediation is often a slow and difficult process. In order to use and manage groundwater resources productively and safely, it is necessary to understand the occurrence and mobility of naturally occurring groundwater constituents and the fate and transport of contaminants in the sub- surface environment. Insight into the biogeochemical processes by which chemi- cal species are mobilized from or sequestered into immobile phases in the subsur- face is crucial to this understanding. The chemical sciences offer insight into the molecular mechanisms of bio- geochemical processes, links between nano-, laboratory, and field scales through modeling and simulation, sensors for the detection and quantification of chemical constituents of groundwater, and chemical reagents and chemistry-based tech- nologies for in situ remediation of contaminated aquifers. Yet, various needs must be addressed if the challenge of providing safe and adequate water supplies is to be met. The interrogation of biogeochemical pro- cesses and subsurface materials at a fundamental level requires continued support that recognizes both the value of studying well-controlled model systems and the need to work directly in complex, environmental systems. Multidisciplinary col- laborations that examine complex systems must be supported and complemented
CHARLES E. KOLB 127 by focused, disciplinary research. Priority should be given to the development and application of sensors for use in the subsurface environment and of tools for the characterization of subsurface materials at the nano- and atomic scale. These latter efforts should support activities at shared or user facilities such as the Stanford Synchrotron Radiation Laboratory (http://www-ssrl.slac.stanford.edu~ and the Environmental Molecular Sciences Laboratory (http://www.emsl.pnl.gov:2080/~. Finally, a crucial resource is provided by facilities, such as the USGS Cape Cod Toxic Substances Hydrology Research Site,5 that allow field-scale experiments to be conducted. Support should be provided for activities at this and additional facilities, such as the Vadose Zone Research Park and the proposed Subsurface Geosciences Laboratory at the Idaho National Engineering and Environmental Laboratory (http://www.inel.gov/env-energyscience/geo/~. The occurrence and mobility of harmful chemical substances, whether of natural origin or anthropogenic contaminants, in the subsurface environment pose both an intellectual and fundamental scientific challenge and practical concerns for the use and management of groundwater resources. The chemical sciences offer powerful approaches toward understanding and mitigating the problems of groundwater contamination. Society has benefited and will continue to benefit from this important application of chemistry to environmental problems. Acknowledgments The collaboration of P. O'Day, S. Dixit, P. Kneebone, and D. Dempsey and the cooperation of the Los Angeles Department of Water and Power in projects described in this paper are gratefully acknowledged. MEASUREMENT CHALLENGES AND STRATEGIES IN ATMOSPHERIC AND ENVIRONMENTAL CHEMISTRY Charles E. Kolb Aerodyne Research, Inc. Atmospheric and environmental chemistry are rapidly evolving disciplines that play a critical role in a wide variety of current environmental issues spanning local, regional, continental, and global scales. Effective management of these issues generally requires knowledge of chemical and physical properties over a wide range of spatial and temporal scales, with processes involving fluid media (atmospheric, oceanic, surface, and groundwater) usually demanding frequent updating (seconds to months, depending on chemical lifetimes and/or transport rates). Chemical characterization methods that rely on traditional sample collec- tion and subsequent analysis are generally slow, labor intensive, time consuming, and costly. Alternatively, environmental systems may be monitored pseudo-
28 APPENDIX D continuously at fixed-site instrumented stations, reporting ambient chemical species concentrations with time resolutions ranging from minutes to months. However, cost restraints usually severely limit the number of fixed-site monitor- ing stations deployed so that impacted environments are sparsely sampled spa- tially. The constraints imposed by traditional environmental measurement methods generally lead to environmental systems being badly undersampled in the spatial and/or temporal domains. The undersampling of environmental systems often has several unfortunate consequences. First, it can lead to misdiagnosis of both the nature and extent of environmental problems, leading to either an over- or underestimation of the seri- ousness or extent of the problem. Second, environmental process models devel- oped to assess the effectiveness of regulatory or other environmental manage- ment challenges are often "validated" with insufficient experimental data to truly constrain the model, resulting in potentially unproductive or even counterproduc- tive management strategies. Third, misunderstanding the nature and/or extent of an environmental problem can obscure its interconnections with other environ- mental problems, creating the possibility that management strategies may help eliminate or mitigate the target issue but exacerbate connected problems. Environmental measurements are challenging for a number of reasons that are summarized in the following section. After presenting measurement chal- lenges, some promising strategies for addressing environmental management challenges are advanced. Atmospheric chemistry measurement challenges and strategies are emphasized, but many of the lessons learned there can be applied to environmental problems with water, soil, and ecological components. Environmental Measurement Challenges The world is a pretty large place. The Earth's surface area is ~500 million square kilometers, approximately two-thirds ocean and one-third land. For atmo- spheric issues we are typically concerned with the two lowest atmospheric re- gions, the troposphere and the stratosphere, which compose the first 50 km of the atmosphere, with a volume of ~25 billion cubic kilometers. So atmospheric chemical issues that are global in scope pose a real measurement challenge. Fur- thermore, as noted above, the atmosphere is dynamic, so chemical and physical characterization measurements must be repeated, sometime quite frequently, to well describe and predict the evolution of processes of interest. Also, because the atmosphere is highly dynamic, quantification of fluxes surface-atmosphere emissions and depositions, stratospheric-tropospheric exchange, boundary layer detrainment, transport across the Intertropical Convergence Zone (ITCZ) are often as or more important than ambient concentration measurements. Atmospheric measurements are also challenging because they must deal with low to extremely low concentrations of trace chemical species. The major com- ponents (>99.999%) of the lowest portions of the atmosphere (the troposphere up
CHARLES E. KOLB 129 to ~10 km in altitude and the stratosphere between ~10 and ~50 km) are molecu- lar nitrogen, molecular oxygen, argon, water vapor, and carbon dioxide. Chem- ists will recognize that all of these species are very stable, strongly bonded mol- ecules or atoms that are essentially inert gases at normal atmospheric temperatures (190-310 K). Indeed, without solar photons to break up selected molecules, atmo- spheric chemistry would be very dull indeed. Atmospheric chemistry is domi- nated by trace species, ranging in mixing ratios (mole fractions) from a few parts per million, for methane in the troposphere and ozone in the stratosphere, to hun- dredths of parts per trillion, or less, for highly reactive species such as the hy- droxyl radical. It is also surprising that atmospheric condensed-phase material plays very important roles in atmospheric chemistry, since there is relatively so little of it. Atmospheric condensed-phase volume to gas-phase volume ratios range from about 3 x 1O-7 for tropospheric clouds to ~3 x 1O-~4 for background stratospheric sulfate aerosol. Progress in understanding key atmospheric environmental issues well enough to identify and test effective management strategies is very dependent on our ability to measure the required range of chemical and physical parameters over adequate spatial scales and appropriate time scales. It has been recognized for more than a decade that the inability of available instrumentation to meet these needs is a serious issue in both atmospheric chemistry research) and assessments of the effectiveness of air quality regulations.2 Environmental Measurement Strategies It is important to recognize that environmental scientists make measurements for a variety of reasons, each of which imposes its own requirements and con- straints on the instrumentation and measurement systems to be used. General environmental measurement modes include exploratory mapping and surveying, process investigations, baseline establishment and trend monitoring, and emis- sion-deposition and other flux measurements. Table 1 lists some of the goals that drive each mode of measurement and some of the measurement system capabili- ties that they require. While the various environmental measurement modes have a range of re- quirements, the undersampling problem discussed above is endemic. We are fre- quently unable to make enough measurements under an adequate range of envi- ronmental conditions over a sufficient range of spatial dimensions and time scales to understand and describe the issue being investigated. Three general strategies have been identified that can address the environmental undersampling issue. ~ Albrittton, D. L.; Fehsenfeld, F. C. Tuck, A. F. Science 1990, 258, 75-81. Rethinking the Ozone Problem in Urban and Regional Air Pollution, National Research Council, National Academy Press, Washington, DC, 1991.
130 TABLE 1 Environmental Measurement Modes APPENDIX D Mode Goal(s) Typical Requirements Exploratory . . mappmg-surveymg Establish ranges of environmental variables Investigate spatial-temporal variability Check for surprises Easy sampling-analysis, mobility Process investigation Quantitative understanding Multispecies-multiparameter Predictive model measurements, high temporal and spatial resolution, flux measurements Baseline-trend monitoring Emission-deposition measurements Quantify current state and rate of change Determine if regulations- management strategies are working Establish pollutant sources and quantify fluxes Check estimation algorithms Improve environment models Determine if regulations- management strategies are working High precision, regular repetition, long-term stability, adequate spatial and temporal resolution and coverage Mobility, moderate to very high temporal and spatial resolution, ability to correlate with tracer species They are: (1) fast sensors on mobile platforms; (2) remote sensors; and (3) dis- tributed sensor networks. These strategies are listed in Box 1, which includes some descriptors of their current or potential implementations. Of course, these strategies are not mutually exclusive, remote sensing instruments can be mounted on mobile platforms, with airborne and space-based passive (radiometers, spec- trometers) and active (lidar, radar) remote sensing instruments playing a large role in our understanding of global-scale environmental issues such as strato- spheric ozone destruction and climate change. Further discussion of these three approaches, their application to a variety of environmental problems, and the instrument challenges they pose can be found in the report of a recent workshop on environmental instrumentation sponsored by the National Science Founda- tion.3 The environmental science community has made enormous advances in Instrumentation for Environmental Science 2000Report of a Workshop and Symposium, Na- tional Science Foundation, Arlington, VA, 2000.
CHARLES E. KOLB 131 the past two decades in designing, fielding, and utilizing rapid sensors on mobile platforms and remote sensing instruments. One point made in the NSF report is that anticipated advances in sensors based on micro- and nanotechnology- coupled with advanced information technology solutions to the problems of data collection, processing, dissemination, and display may make the distributed sen- sor network measurement strategy much more powerful and affordable. Over the past 15 years, the atmospheric science community has developed a series of mobile platforms with highly accurate and specific fast response instru- mentation that have revolutionized atmospheric chemistry field measurements. These include high-altitude aircraft, such as NASA's ER-2 and WB-57, and lower-altitude aircraft like the NASA DC-8, the National Oceanic and Atmo- spheric Administration (NOAA) and Center for Interdisciplinary Remotely-Pi- loted Aircraft Studies (CIRPAS) (Naval Postgraduate School) Twin Otters, the National Center for Atmospheric Research (NCAR) C-130, and the DOE G1. In addition, mobile surface laboratories are now being used for a wide variety of urban and regional air quality and emission source characterization studies.4 Typi- cal configurations for the ER-2 and the mobile laboratory are shown in Figures 1 and2. Interestingly, the mobile (comparatively) fast-sensor strategy has recently been implemented for subsurface studies in the form of a truck-mounted geotechnical probe equipped with miniaturized laser-induced fluorescence or la- ser-induced breakdown spectroscopy sensors to detect vertical and horizontal dis- 4See for example, Lamb, B., McManus, J. B.; Shorter, J. H.; Kolb, C. E.; Mosher, B.; Harriss, R. C.; Allwine, E.; Blaha, D.; Howard, T.; Guenther, A.; Lott, R. A.; Siverson, R.; Westberg, H.; Zimmerman, P. Environ. Sci. Technol. 1995, 29, 1468-1479; Jimenez, J. L.; McManus, J. B.; Shorter, J. H.; Nelson, D. D.; Zahniser, M. S.; Koplow, M.; McRae, G.J.; Kolb, C. E. Global Change Sci. 2000, 2, 397-412.
32 APPENDIX D FIGURE 1 View of NASA WER-2 High Altitude Aircraft With Stratospheric Chemistry Instrument Package. Aerosol Mass Spec . Size Resolvecl Composition 40 nm - 2 ,um nitrate, sulfate, ammonium Organic Carbon (OC) diesel exhaust signature . Proton Transfer Reaction Mass Spectrometer Rapid Measurement of selected VOC's ...; Condensation Particle Counter Number Density Tunable Diode Lasers Combinations of NO, NO2 CO, N2O, CH4, SO2, HCHO typical detection limit < 1 ppb ............ , ~ ~ ~ ~ ~ coo .< ~~ f ~ ~ . 2 ~ ~ ·~.'; it. a ; a., ,~ :.;. , ~ -. . ~ a ....... i. ~ . . ~ . an. Am,. ~ . ~ ~ ~~ .:: ~ ~ . ~ :.~....f .' | Isokinetic Inlet System | FIGURE 2 Schematic of aerodyne research mobile laboratory deployed for air quality studies in Mexico City.
CHARLES E. KOLB 133 tributions of aromatic organic and heavy metal contaminants, respectively.5 Box 2 lists the general goal and some of the desirable characteristics for exploratory and process study field measurement campaigns that are required to take full advantage of fast-sensor-mobile platform strategy. Box 3 illustrates some of the evolution in environmental analytical instrumentation (especially in atmospheric measurement technology) that has motivated the development of instrument suites such as those illustrated in Figures 1 and 2. In order to produce the instruments and platforms that meet the specifica- tions required by the measurement strategies listed in Box 1, advances in en- abling technologies must be exploited. Table 2 lists some of the enabling tech- nologies that are currently undergoing rapid development and presents examples of the improvements in environmental instrumentation and measurement plat- forms they have or may soon allow. 5See for example, Sinfeld, J. V.; Germaine, J. T.; Hemond, H.F. J. Geotech Geoenviron. 1999, 125, 1072-1077; Theriault, G. A.; Bodensteiner, S.; Liberman, H. Field Anal. Chem. Technol. 1998, 2, 117-125; Instrumentation for Environmental Science 2000Report of a Workshop and Symposium, National Science Foundation, Arlington, VA, 2000.
134 APPENDIX D TABLE 2 Impacts of Enabling Technologies Technology Examples of Evolving Impacts Structural materials Energy systems Electro-optics Semiconductor technology Ion optics Lighter, more robust sensors and sensor platforms Smaller, longer lasting off-grid power sources, enhanced platform propulsion systems Compact, efficient solid state lasers and detectors, compact long-path sampling cells, detector arrays Compact, robust electronics, throw-away sensors and data systems Ion traps, smaller TOF and quadrupole mass filter spectrometers Vacuum technology Fieldable sensors based on electron, ion, molecular, particle beam methods Information technology Real-time data processing and display, multisensor integration, data fusion and assimilation Control technology Autonomous instrument-platform operation, real-time Fluid dynamics Biotechnology Nanotechnology experimental design More efficient airborne platforms, better sampling systems Smaller, faster diagnostics for microbial and biologically active molecules Smaller, lighter, cheaper everything Figure 3 shows an airborne version of one recently developed fast response instrument made possible by recent advances in materials, vacuum technology, ion optics, fluid dynamics, information technology, and control technology.6 This aerosol mass spectrometer allows the real-time measurement and display of the nonrefractory, size-resolved (~30 nm to ~1500 nm) ambient aerosol particle mass loadings. Figure 4 shows the ambient fine aerosol nonrefractory composition, Wayne, J. T.; Leard, D. C.; Zhang, X.; Davidovits, P.; Smith, K. A.; Kolb, C.E.; Worsnop, D. R Aerosol Sci. Technol. 2000, 33, 49-70.
CHARLES E. KOLB 135 Physical Power: 600 watts Mass: 190 kg Size: ~36 " x 36 " x 24 " (plus half rack of electronics) Performance Measurement size range: 30 nm to 1 ,um Sensitivity: 0. 01 ~g/m3 (60s sec.) Sampling Rate: 100 cm3 mind Maximum Data Rate: 100 Hz G1 Package Integral size and composition information for non-refractory bulk constituents on seconds time scale FIGURE 3 An Aerosol mass spectrometer allowing real-time size-resolved chemical spe- ciation measurements of fine aerosols, free standing and packaged for airborne measure- ments. 70 60 50 40 30 20 10 Nitrate Sulphate Ammonium Organics Water _ ~ i']" ' | /4 , ,, ,,,,, ,.,,, , ,,,,,,.,, ,,,, ,A, u L _. 12:00 PM 12:30 PM 1:00 PM 1:30 PM 2:00 PM 7/22/2002 Date and Time 2:30 PM 3:00 PM 3:30 PM 8000 6000 D 4000 2000 o FIGURE 4 Aerosol mass spectrometer airborne measurements of fine particle composi- tional mass loading and size distributions showing two sulfate aerosol layers in air masses from the Ohio River valley.
136 APPENDIX D mass loading, and size distribution measured as a function of aircraft altitude of the coast of Massachusetts in July 2002. The two large sulfate peaks are sampling air with back trajectories that left the surface over the Ohio River valley two days earlier. Summary It is clear that innovative instruments and measurement strategies are re- quired to cure the undersampled environment problem. They can be developed and deployed if ongoing advances in a wide range of enabling technologies are adapted and exploited. For many environmental challenges, measurement require- ments favor real-time, mobile, autonomous instruments. The resulting quantum leaps in measurement capabilities will have the potential to revolutionize atmo- spheric and environmental chemistry. ENVIRONMENTAL BIOINORGANIC CHEMISTRY Francis M.M. Morel Princeton University The convolution of life and geochemistry on the Earth has resulted in an extraordinarily tight coupling between the cycles of bioactive elements at the surface of our planet and the growth of microorganisms. This is evidently true of major plant nutrients such carbon, nitrogen and phosphorus. It is also true of many essential trace metals, such iron, manganese, copper, cobalt, nickel, and zinc, that serve as active centers in enzymes that catalyze the transformations of carbon and nitrogen in terrestrial and aquatic systems. The biological availability and use of these metals are modulated by intracellular and extracellular binding compounds produced by microorganisms. The chemistry of bioactive metals in the environment, including their coordination to binding agents and to appropri- ate centers in proteins, thus controls the efficiency of critical environmental pro- cesses that govern the global cycles of carbon and nitrogen. A detailed under- standing of the bioinorganic chemistry of these metals in various environs and of their effects on geochemically important processes presents major challenges and opportunities to environmental chemists as illustrated in the three oceanographic examples below. The increase in O2 and concomitant decrease in CO2 caused by the evolution of oxygenic autotrophs on the Earth, has resulted in an undersaturation of the main (and ancient) carboxylating enzyme, ribulose-1,5-bisphosphate carboxylase, responsible for the first step in carbon fixation the dark reaction of photosyn- thesis. To palliate this difficulty, a number of species of marine phytoplankton have evolved carbon concentrating mechanisms that all involve some forms of
FRANCIS M.M. MOREL 137 the zinc enzyme carbonic anhydrase.i This enzyme catalyzes the hydration of CO2 and the dehydration of HCO3-. In the surface oceans, which are singularly depleted in zinc, a number of phytoplankton species have evolved the ability to replace zinc with cobalt and cadmium in carbonic anhydrase.2 3 In diatoms, which are arguably the most important primary producer in the modern oceans, it ap- pears that the activity of the external carbonic anhydrase is enabled by the forma- tion of a silica frustule that serves, in part, as a proton buffer at the surface of the cell.4 Unraveling the conditions that allow for an efficient Zn-Co-Cd replace- ment in carbonic anhydrase (and similar enzymes) in marine autotrophs and of the use of silica as a local proton buffer will provide a molecular understanding of the links between the global cycles of these trace metals and of CO2 and SiO2. The increase in O2 concentration in the oceans since the evolution of oxy- genic photosynthesis has also led to a massive decrease in the concentrations of metals such as iron and manganese, which form insoluble oxy-hydroxides. These metals, particularly iron, are important in the transfer of electrons in photosynthe- sis and respiration and as metals centers in a wide variety of redox enzymes. Not surpnsingly, there is mounting evidence that modern marine microorganisms wage a fierce "iron war" against each other and that many have acquired the ability to replace iron with other metals in key compounds. For example, the production of marine siderophores allows some marine bacteria to sequester iron and limit its availability to organisms that do not possess the appropriate transport proteins (Figure 1~.5 As a countermeasure, some eukaryotic phytoplankton ex- tract the iron from Fe(III)-siderophore complexes by reduction of Fe(III) to Fe(II) extracellularly.6 We need to elucidate the chemical forms of iron in seawater and the mechanisms of iron acquisition by venous m~croorgan~sns in order to assess the role of iron in limiting primary production in the oceans7 and in controlling the assemblage of planktonic species. Nitrogen is generally considered the most important limiting plant nutrient in the oceans. The processes that determine the total concentration of fixed nitrogen in surface seawater particularly N2 fixation and denitnfication thus exert a fundamental control on marine cnmary production and on the resulting seques- iBadger, M. R.; Hanson, D.; Price, G. D. Functional Plant Biology 2002, 29, 161-173. 2Morel, F. M. M.; Reinfelder, J. R.; Roberts, S. B.; Chamberlain, C. P.; Lee, J. G.; and Yee, D. Nature 1994, 369, 740-742. 3Lane, T. W.; Morel, F. M. M. Proc. Natl. Acad. Sci. USA 2000, 97, 4627-4631. 4Milligan, A. J.; Morel, F. M. M. Science 2002, 297, 1848-1850. 5Butler, A. Science 1998, 281, 207-210. 6Maldonado, M. T.; Price, N. M. Journal of Phycology 2001, 37, 298-309. 7Coale, K. H.; Fitzwater, S. E.; Gordon, R. M.; Johnson, K. S.; Barber, R. T. Nature 1996, 379, 621-624.
138 APPENDIX D FIGURE 1 Two examples of ampiphilic siderophores produced by marine heterotrophic bacteria. Redrawn from Martinez et al. (2000~. tration of CO2 in the deep oceans. All the steps in the nitrogen cycle are catalyzed by metalloenzymes (Figure 2~. For example, all forms of nitrogenase (the enzyme responsible for N2 fixation) contain a large number of iron atoms. Iron, molybde- num, and copper are also involved in the various enzymes that carry out denitrifi- cation. There is presently wide speculation that iron availability limits the overall
FRAN'(~OIS M.M. MOREL 139 FIGURE 2 A diagram of the nitrogen cycle with catalyzing enzymes and metal require- ments of each step. NOTE: AMO = ammonium mono-oxygenase; HAO = hydroxylamine oxidoreductase; NAR = membrane-bound respiratory nitrate reductase; NAP = periplasmic respiratory nitrate reductase; NR = assimilatory nitrate reductase; NIR = respiratory nitrite reductase; NiR = assimilatory nitrite reductase; NIT = nitrogenase; NOR = nitric oxide reductase; N2OR = nitrous oxide reductase. rate of nitrogen fixation in the oceans.8 9 There is also some evidence that, in some suboxic waters, the concentration of available copper may be too low for the activity of the copper enzyme nitrous oxide reductase and result in N2O accu- mulation and release to the atmospheres Understanding the nitrogen cycle of the oceans and the release of some important greenhouse gases such as N2O to the atmosphere thus requires that we elucidate the acquisition of metals such as iron and copper and their biochemical utilization by various types of marine microbes. As exemplified above, the major goals of environmental bioinorganic chem- istry are to elucidate the structures, mechanisms, and interactions of important "natural" metalloenzymes and metal-binding compounds in the environment and to assess their effects on major biogeochemical cycles such as those of carbon and nitrogen. By providing an understanding of key chemical processes in the biogeochemical cycles of elements, such a molecular approach to the study of global processes should help unravel the interdependence of life and geochemis- try on planet Earth and their convolution through geological times. ~Falkowski, P. G. Nature 1997, 387, 272-275. 9Berman-Frank, I.; Cullen, J. T.; Shaked, Y.; Sherrell, R. M.; Falkowski, P. G. Limnology and Oceanography 2001, 46, 1249-1260. i°Granger, J.; Ward, B. Limnology and Oceanography 2002, 48, 313-318.
140 APPENDIX D ENVIRONMENTALLY SOUND AGRICULTURAL CHEMISTRY: FROM PROCESS TECHNOLOGY TO BIOTECHNOLOGY Michael K. Stern Monsanto Company Agncultural practices are undergoing a transformation driven partially by advances at the interface of chemistry and biotechnology. This paper outlines some of the new technologies that are playing an integral role in catalyzing that change. Topics of discussion include new chemical process technology and chemical catalysis that allows for more efficient production of herbicides as well as transgenic crops and their benefits to agriculture and the environment. Glyphos ate is the active ingredient in Roundup herbicide. The process Monsanto uses to manufacture glyphosate relies heavily on chemical catalysis (Figure 1~. Improvements in our catalyst have driven profound environmental and economic benefits in the manufacturing of glyphosate. One of the main themes around environmental chemistry moving into the future will be a renewed focus on the development of novel homogeneous and heterogeneous catalysts. One of the beautiful things about catalysts is their ability to positively impact the economics of a process often without the requirement for major capital invest- ments. Steps DEA \ Catalyst or , HCN ~ CH2O ~ NH3 i'N into Na-O H O-Na DSIDA P4 + Chlorine I PCI 3 q~N - O DSIDA ~ Formaldehyde , HO ~ OH PO 3H 2 Gl Gl ~O Catalyst FIGURE 1 The current glypho sate process. H O PA -No Aft H OH Glyphosate
MICHAEL K STERN Copper Plating Technology FIGURE 2 Second-generation catalyst. 141 A key intermediate in the manufacturing of glyphosate is disodium iminodiacetic acid (DSIDA). Originally an HCN-based route was the only pro- cess used to manufacture this product. Some of the advantages of this technology are (1) it's proven, (2) it uses readily available raw materials, and (3) the process typically gives good yields. However there are a lot of challenges associated with this chemistry. For instance, a considerable amount of waste is produced and HCN is a difficult raw material to handle due to its toxicity. Accordingly, Monsanto wanted to explore other technologies when we were faced with the need to expand our DSIDA capacity in the early 1990s. We ultimately settled on a novel catalytic route for the production of DSIDA. The initial catalyst was a Raney copper composition that allowed us to convert diethanolamine, to DSIDA in a really interesting dehydrogenation reaction. This is an endothermic reaction that gives off hydrogen as a by-product. We were able to get the facility to work with this catalyst, but there were a lot of operational issues associated with this technology. Raney copper is a very malleable soft metal which resulted in catalyst stability issues. Ultimately we needed to go ahead and find something that was better. Copper is essentially the only metal that catalyzes this reaction. The technic cat challenge was to find a way to stabilize copper under the reaction conditions. We spent a lot of time looking at whether you could put copper directly on car- bon. It turns out that you really can't do that very well. Copper likes to move around, particularly under the reaction conditions. So a new catalyst technology was developed using platinum as an anchor that was then coated with copper. This resulted in a very stable catalyst (Figure 2~. The new catalyst technology had significant environmental benefits. These included the use of less toxic raw mate- rials and the elimination of nearly all of the waste produced by the older technol- ogy.
142 APPENDIX D Let me switch gears to another catalytic reaction. This is the reaction where we take glyphos ate intermediate and convert it to glyphos ate using a carbon cata- lyst. This appears to be a very simple reaction, but there are several technical issues related to the production of by-products. The problem is that the by-prod- ucts go ahead and react with your desired product to make other undesired by- products. What we needed was a catalyst that could do two reactions at the same time. The first reaction converts glyphos ate intermediate to glyphosate and the other is to react away the undesirable by-products. We were successful in devel- oping this type of catalyst which resulted in significant environmental and com- mercial benefits (Table 1~. The major benefits associated with this technology result from the more effi- cient use of water in the process. With the implementation of this new catalyst technology we have been able to reduce the amount of water flowing through our process by nearly 300 million gallons a year. This resulted in a concomitant re- duction in flows to our biotreatment systems that reduce the amount of biosolids we send to landfill. Overall this was a very successful project for Monsanto. I'd like to change focus from catalysis to another theme in green chemistry: that of atom efficiency. If you noticed in the glyphos ate process, we're not com- pletely atom efficient. This is due to the fact that putting on the phosphonomethyl group is challenging. The issue is that if you don't protect the primary amine, you end up doing two phosphonomethyl reactions that yield the undesirable product glyphosine. The current solution to this problem is to protect the primary amine with a carboxymethyl group, which ultimately gets removed in the last step of the process. The challenge was to find a more atom-efficient protecting group. A team investigated this and developed a whole new technology based on novel platinum catalysts. This technology allows for the selective demethylation of N-methylglyphosate to produce glyphosate acid directly. This would save one TABLE 1 Environmental Benefits of New Catalyst and Process Technology Annual Reductions Projected by 2002 Resources Steam (BTUs/yr) Demineralized water (gal/yr) Waste Flow to biosystems (gal/yr) SARA 313 deep well injection (lb/yr) Biosludge (lb/yr) Land-filled solid waste (lb/yr) SARA 313 air emissions (lb/yr) Carbon dioxide production (lb/yr) 880,000,000,000 380,000,000 800,000,000 52,000 8,000,000 1,380,000 17,600 100,000,000 NOTE: SARA = Superfund Amendments and Reauthorization Act.
MICHAEL K STERN H2 3 cat. A - ~N~CO2H A H2O3P\' N. ::CO2H 2 Moedritzer-lrani ~ H H2O3P\' N. ::CO2H + FIGURE 3 N-isopropyl glyphosate: an atom-efficient intermediate. 143 ~'N::CO2H N-isopropyl glycine By H2O3P\: N. ::CO2H N-isopropyl glyphosate carbon atom when compared to the traditional route. However it was also discov- ered that you could use other protecting groups besides methyl. In fact it was possible to use an isopropyl group, which under the reaction conditions could be removed to generate acetone and glyphosate. The acetone can be recycled back into the process, resulting in an extremely atom-efficient system (Figure 3~. As mentioned above, glyphosate is the active ingredient in Roundup herbi- cide. Roundup plays an important part in the new wave of agricultural products derived from biotechnology. This new technology has many economic and envi- ronmental benefits. Expansion of the global acreage planted with Roundup Ready crops has resulted in a reduction of the use of pesticides by nearly 50 million pounds per year. This can affect groundwater positively by reducing agricultural chemical contamination in watersheds where a large percentage of Roundup Ready crops are planted. When crops such as cotton and corn are protected against insect pest through biotechnology, we also see a benefit to nontarget organisms. So in summary, biotechnology has delivered significant environmental benefits. Many of these benefits are consistent with the EPA's guidelines and focus. In conclusion, the chemical industry is going to be even more dramatically transformed in the future, and key advances will be made at the interface of chem- istry and biology. Discovery and development of new environmentally beneficial catalyst and process technologies will be critical for the chemical industry to thrive in the United States. The new technologies will need to be relevant and have a positive impact on the earnings and competitiveness of the chemical in- dustry. Advances in catalysis will be a key driver in the development of cleaner and more efficient chemical processes. Finally, breakthrough discoveries are likely to be the products of interdisciplinary work teams.
44 APPENDIX D STABLE ISOTOPES AND THE FUTURE OF ENVIRONMENTAL- CHEMICAL RESEARCH Mark H. Thiemens University of California, San Diego In 1947, Harold Urey and, simultaneously, Bigeleisen and Mayer developed the formalism for determination of the position of equilibria in isotope exchange reactions. These papers, for the first time, calculated at high precision the posi- tion of isotope exchange equilibria as a function of temperature. In the same year, Nier reported the development of the double collector isotope ratio mass spec- trometer, which allowed for measurements of isotope ratios at a precision suffi- cient to measure the modest isotope ratio changes associated with the temperature dependencies of exchange reactions. In this sense, 1947 represents the birth of stable isotope chemistry. Subsequently, an enormous range of applications has emerged utilizing isotope ratio measurements as a probe of natural processes that include studies of atmospheric chemical processes, paleoceanography and cli- mate, stable isotope geochemistry, and planetary sciences. Although the utilization of stable isotopes as a means to resolve environmen- tal processes has had many applications for more than a half-century, there have been some limitations on the extent to which specific processes, both chemical and physical, might be resolved. These limits arise because with only a single isotope ratio, there is a certain lack of specificity associated with the measure- ments. In 1983, Thiemens and Heisenreich reported a new isotope effect. This par- ticular effect was unique in that the isotope ratios (e.g., of oxygen) alter on a basis other than mass. For example, in the case of ozone formation, the isotopomers 160160170 160160180 form at essentially equally rates that exceed those associated with 16016ol6o. These reactions proceed in part on the basis of isotopic symmetry, with the asym- metric species forming at a rate greater than the purely symmetric species. Marcus and colleagues at Cal Tech have developed a chemical formalism that accounts for a large proportion of the laboratory experiments. There remain some funda- mental issues, however, with respect to a fully developed quantum-level theory. While additional theoretical formalisms for the isotopic fractionation event are still needed, the mass-independent isotopic fractionation process has provided a new and definitive mechanism by which an extraordinary range of environmental
MARK H. THIEMENS 145 processes may be resolved. The inclusion of the second isotope ratio measure- ment adds a sensitive probe of processes that may not be afforded by concentra- tion or single isotope ratio measurements. As such, the use of mass-independent isotope compositions of environmental molecular species has provided a new probe to understand natural processes and to characterize anthropogenic impacts. There will be a wide range of new and significant applications in the future that will provide a powerful complement to other measurement techniques and model- ing efforts. These include studies of the atmosphere, hydrosphere, and geosphere, as well as paleoenviroments and global environmental change. Such studies will be of critical importance in evaluating and predicting global environmental sustainability. Present and Future Applications There exist numerous recent review articles on the subject of mass-indepen- dent isotope effects and details are available in these articles.) 2 3 A key point of the culmination of observations is that the most important chemical issues in the environment today, and as expected in the future, may be studied utilizing mass- independent isotopic compositions. In the context of this report, this represents a future chemical frontier area of chem~cal-environmental research. There exist a number of gases that possess mass-independent isotopic com- positions, and with future measurements, the frontiers of environmental chem~s- try may be extended. The following are specific, though not inclusive, examples. Stratospheric and Mesospheric CO2 It was first observed by Thiemens et al.4 that stratospheric CO2 possesses a large and variable mass-independent isotopic composition. This composition was suggested as denying from isotopic exchange with O(iD), the product of ozone photolysis.5 6 As later confirmed by rocket-borne collection of stratospheric and mesosphenc air, this unique isotopic signature provides an ideal tracer of odd oxygen chemistry of the Earth's upper atmosphere, one of the most important upper atmospheric processes. There are, however, several features that require further measurement (laboratory and atmosphenc) and theoretical considerations. iThiemens, M. H. Science 1999, 283, 341. 2Weston, R. E. Chem. Rev. 1999, 99, 2115. 3Thiemens, M. H.; Savarino, J.; Farquhar, J.; Bao, H. Acct. Chem. Res. 2001, 34, 645. 4Thiemens, M. H.; Jackson T. L.; Mauersberger, K.; Schuler, B.; Morton, J. Geophys. Res. Lett. 1991, 18, 669. 5Yung, Y. L.; DeMore; W. B.; Pinto; J. P. Geophys. Res. Lett. 1991,18, 13. 6Yung, y. L.; Lee, A. Y. T.; Iriow, W. B.; Demaro, W. B.; Chen, J. J. Geophys. Res. 1997,102, 10857.
146 APPENDIX D This represents a future goal in need of pursuit because both climate and chemical budgeting considerations are impacted. As discussed in an earlier report,7 upper atmospheric CO2 possesses a mass-independent isotopic composition, while tro- posphenc CO2 is strictly mass dependent (as a result of equilibrium isotopic ex- change with water). This renders CO2 an ideal tracer of stratosphere and tropo- sphere mixing. Quantification of this process is also of significance for understanding global chemical budgets and lifetimes of several greenhouse spe- cies. Given that significant uncertainties remain, future measurements will be crucial in quantification of stratospheric and tropospheric mixing and in develop- ment of remediation policies associated with global climate change. Greenhouse Gas Characterization As discussed in the review article,8 several greenhouse gases have been ob- served to possess mass-independent isotopic compositions. These include O3, CO, and N2O. In each instance, this measurement has provided significant insight unattainable by other measurement techniques. The case of atmospheric N2O is particularly interesting. Nitrous oxide is a greenhouse gas with a warming capac- ity nearly 200 times that of CO2 on a per-molecule basis, and serves as a major sink for stratospheric ozone via photochem~cal destruction. In spite of decades of research, the N2O budget is still inadequately understood. Most recently, isotopic measurements of a new variety have proven to be particularly valuable. These observations utilize high-precision measurements of the isotopomenc fragments of NO in a mass spectrometer.9~0~2 From such measurements, the internal N2O isotopomenc distribution may be determined. It is now recognized that the isotopomenc distributions of 969. 15N14N160 15N14N180 14N15N160 14N15N180 7Thiemens, M. H.; Jackson, T.; Zipf, E.C.; Erdman, P.W.; Van Egmond, C. Science 1995, 270, ~Thiemens, M. H. Science 1999, 283, 341. 9Toyoda, S.; Yoshida, N. Anal. Chem., 1999, 71, 4711-4718. i°Brenninkmeijer, C. A. M.; Rockmann, T. Rapid Commun. Mass Spec. 1999,13, 2028. i~Rockmann, T.; Kaiser, J.; Crowley, J. W.; Brenninkmeijer, C. A. M.; Crutzen, P. Geophys. Res. Lett. 2001, 28, 503. i2Toyoda, S.; Yoshida, N. C.; Urabe, T.; Aoki, S.; Nakazawa, T.; Sugawara, S.; Honde, H. J. Geophys. Res. 2001,106, 7515.
MARK H. THIEMENS 147 are highly characteristic of specific processes, such as photolysis or defining indi- vidual point sources (biologic, abiologic). In the future, this particular variety of measurements will provide a new level of detail in understanding the global atmospheric cycles of N2O as well as the other gases. There is a clear need for expansion of such measurements in a variety of global environments and to obtain fundamental physical-chemical information simultaneously. With such concomitant developments, new details of N2O atmo- spheric processes may be obtained. Atmospheric Aerosol Sulfate and Nitrate Atmospheric sulfate aerosols are known to exert a significant influence on Earth's surficial processes. They mediate climate, both as cloud condensation nuclei and as light-scattering agents. Sulfate is also a pernicious respirable mol- ecule with well-known consequences for human health. It is estimated that there are more than 60,000 deaths a year from cardiovascular disease associated with aerosol inhalation. Recent studies have demonstrated a link to cellular damage. Additionally, following wet and dry deposition, sulfate destroys biota, alters biodiversity, and causes pervasive structural damage. The segregation of sulfur between gas- and aqueous-phase oxidative processes has implications for the de- gree of indirect effect of sulfate aerosols on climate due to its dependence on aerosol number densities. Sulfate derived from gas-phase oxidation results in new particle formation. However, for aqueous-phase oxidation, the sulfate is gener- ated on a previously existing particle and does not contribute to the total aerosol number. It has recently been shown that atmospheric sulfate possesses a significant and variable oxygen mass-independent isotopic composition. From a series of laboratory and atmospheric measurements it has been demonstrated that these measurements provide a highly sensitive mechanism by which the relative pro- portions of hetero- and homogeneous oxidative pathways may be quantified. This represents a significant observational advancement, and future measurements on a global scale will dramatically enhance understanding of this important atmo- spheric species. It has also been shown that sulfate oxygen isotopic measure- ments of aerosols collected during the Indian Ocean Experiment (INDOEX) re- vealed that the Intertropical Convergence Zone (ITCZ) is a source of new aerosol particles, which has significant consequences. First, this is a previously unrecog- nized process unaccounted for in any global climate model. Secondly, this is a source of large (micron-sized) particles and the mechanism associated with their formation is unknown. Gas-to-particle conversion processes produce submicron- rather than micron-sized particles. The likely mechanism for this process may be surface catalysis, possibly on carbonaceous particles. Should this be confirmed, there would be significant consequences. Climate models assume that sulfate par- ticles are white and reflective of visible light. If sulfur is catalytically oxidized on
148 APPENDIX D carbon surfaces, the sign of sulfate radiative forcing may be the reverse of what has been assumed, in which case there could be significant error in climate mod- els. It is therefore of considerable importance that the nature and magnitude of this process be resolved. Fully understanding the reaction pathways can be ac- complished only by an intensive combination of isotopic measurements, climate models, and laboratory studies of the relevant surface catalytic reactions. Such a program typifies the future needs of environmental chemical research. As is the case of sulfate, nitrate aerosols also possess large, mass-independent isotopic compositions. Nitrate concentrations may double in the next half-century with severe environmental consequences, including alteration of biodiversity, enhanced algal blooms, and loss of agricultural productivity. As in the case of sulfate, the large mass-independent isotopic signature has afforded new insights into a major global cycle the nitrogen cycle. There remains a large range of studies to be pursued, which will develop new understanding of the chemistry of the Earth's environment. Once more, this may be accomplished by combined laboratory physical-chemical measurements, field observations, and modeling efforts. The ultimate consequence is a significant advancement in understanding global interactions of the chemical environment. Summary The utilization of mass-independent isotopic measurements of a wide variety of atmospheric, hydrospheric, and geologic species has advanced understanding of a wide range of environmental processes. The future development of the utili- zation and understanding of this new technique clearly will have numerous appli- cations that should, and will, be advanced. Issues in climate change, health, agri- culture, biodiversity, and water quality all may be addressed. Simultaneous with the acquisition of new environmental insight will be enhanced understanding of fundamental chemical physics.
