Grantee Research Project Results
Final Report: Life Cycle Analysis of Biolubricants for Aluminum Rolling
EPA Grant Number: R831521Title: Life Cycle Analysis of Biolubricants for Aluminum Rolling
Investigators: Theis, Thomas L.
Institution: University of Illinois at Chicago
EPA Project Officer: Aja, Hayley
Project Period: November 1, 2003 through October 31, 2006
Project Amount: $249,944
RFA: Technology for a Sustainable Environment (2003) RFA Text | Recipients Lists
Research Category: Sustainable and Healthy Communities , Pollution Prevention/Sustainable Development
Objective:
This project examined issues associated with the substitution of pertroleum-based lubricants with bio-based (i.e., plant-derived) lubricants for industrial applications. Such an approach appeared to be an attractive material substitution: biolubricants are renewable, relatively nontoxic, biodegradable, and more easily extracted and processed than petrolubricants. There are, however, several aspects of the use of biolubricants that have to be addressed further in order to estimate the societal, environmental, and technological benefits and impacts of the widespread use of these substitutes. These aspects fall into three areas: performance, regulatory, and life cycle.
Performance. Knowledge of the lubrication properties of biolubricants formulated for various applications must be measured and compared with their petro-based counterparts in order to determine such factors as material wear, lubricant stability, and quantitative needs.
Regulatory. There is reason to suggest that volatile organic compound (VOC) emissions and solid waste issues for biolubricants will be less severe than for petrolubricants. In addition to lower toxicity and higher biodegradability, the composition of biolubricants tends to consist of higher molecular weight/lower vapor pressure components. However, there is a need to measure emission rates and compositions, again in comparison with petro-based lubricants intended for similar uses.
Life Cycle. The extent to which they can be reused (either directly or for secondary uses), and the degree to which their production is itself dependent on petroleum products are important matters that must be addressed before it becomes clear that biolubricants are the preferable choice. Given the heavy dependence of U.S. agriculture on petroleum products, a significant quantity of petrochemicals will be used to produce plant feed stocks. In addition, widespread expansion of agricultural production of plants dedicated as feedstocks for biolubricants may have implications for elemental cycling (particularly nitrogen), carbon sequestration, and soil erosion.
Summary/Accomplishments (Outputs/Outcomes):
A comparative life cycle assessment (LCA) examining soybean and petroleum-based lubricants has been compiled using Monte Carlo Analysis (MCA) to determine system variability. Comparative inventory emissions between soybean and mineral oils show that if they have similar use rates during performance, soybean oils have greater life cycle emissions of VOC, NOx, SOx, N2O, NO3-, and total P, and lower emissions of CO2, CH4, and PM10 as well as decreased fossil fuel consumption; however, experimental data obtained from an aluminum manufacturing facility indicate that significantly less soybean lubricant is required to achieve similar or superior performance. With improved performance and a lower use rate, a transition to soybean oil results in lower aggregate impacts of acidification, smog formation, and human health from criteria pollutants. Regardless of the quantity consumed, soybean-based lubricants exhibit significant climate change and fossil fuel use benefits; however, eutrophication impacts are much greater due to nonpoint nutrient emissions. Fundamental tradeoffs in the carbon and nitrogen cycles are addressed in the analysis, demonstrating that a transition to soybean oil may result in climate change benefits at the expense of regional water quality.
Introduction
Commodities derived entirely or partially from biomass, known as bio-based products or biocommodities, have been proposed as substitutes for petroleum in applications ranging from transportation fuels to plastics. While many studies have focused on transportation fuels, the potential exists for successful implementation of biocommodities in niche markets, such as lubricants, plastics, and specialty chemicals (Lynd, et al., 1999; Mohanty, et al., 2002; Wilke, 1995; Dale, 1999). Bio-based lubricants are particularly feasible for adoption within industry; currently they are used as hydraulic fluids and once-through operations such as cutting and stamping. Biolubricants have yet to appear in continuous-loop operations, such as aluminum rolling; however, experimental evidence from this project suggests that the potential exists for implementation.
Over 2 billion gallons (~7.5 billion liters) of lubricants are produced annually in the United States (Energy Information Administration, 2004). Of these, over 900 million gallons (~3 billion liters) are used for industrial purposes, with approximately 100 million gallons (~380 million liters) dedicated to metalworking operations including aluminum rolling (Honary, 1996). The U.S. agricultural sector produces approximately two and a half billion gallons of vegetable oils annually, with 2% of these stocks currently used in nonfood applications (Ash and Dohlman, 2006; Kinney, 1998). Biolubricants are increasing in popularity due to superior technical properties and environmental concerns associated with petroleum lubricants (Pearson and Spagnoli, 2000; Honary, 2001). Recent chemical modifications improve the oxidative stability of vegetable oils, demonstrating their potential to compete with mineral oils in longer-term applications (Honary, 1996; McManus, et al., 2003; Kassfeldt and Goran, 1997; Pal and Singhal, 2000).
It is often assumed that bio-based products are environmentally preferable to petroleum products due to their renewable nature. To determine the validity of this assumption, a comparative LCA allows a quantitative comparison of the energy and material flows throughout the stages of each product, from creation to disposal or reuse (Vignon, et al., 1992). A significant body of work is available on the life cycles of bio-based transportation fuels, such as ethanol and biodiesel, and various databases catalogue the environmental impacts of agriculture (MacLean, et al., 2000; Sheehan, et al., 1998; Fu, et al., 2003; Shapouri, et al., 1995; Lave, et al., 2000). One study focusing on rapeseed oil for use in hydraulic applications shows greenhouse gas emissions reductions for biolubricants but increased impacts of eutrophication, smog production, and energy use (McManus, et al., 2004).
Agricultural systems exhibit considerable variability and uncertainty in emission profiles because of differences in geography, climatology, and agricultural practices. The use of average data to characterize agricultural systems may not adequately generate an accurate system depiction of emissions occurring during “extreme” years (such as rainy or drought years), and the subsequent environmental impacts. This paper incorporates data variability to provide a more comprehensive system description. The use of variable data allows the LCA practitioner to compare alternative products using average values and also best and worst case scenarios given probability estimates.
The purpose of this project was to conduct a comparative LCA for biolubricants and mineral oils, using MCA to incorporate data variability into the assessment. Unlike most metalworking operations that currently use bio-based lubricants in once-through operations, rolling processes recycle lubricants in a continuous loop. Experimental data from aluminum rolling is used to determine lubricant behavior during the process, and convert to an appropriate functional unit. Use phase data are not often incorporated into LCAs due to lack of data, which results in a cradle-to-gate analysis that may overlook important inventory flows. By incorporating experimental data within the analysis, the results are more relevant to actual applications. Although this paper focuses on aluminum rolling as an example, the reported data are applicable to a range of soybean-based products.
Methods
Boundaries. The boundaries for the life cycles of the mineral oil and soybean lubricants modeled in this project are presented in Figure 1. All primary and secondary inputs related to upstream manufacturing are included; however, the contributions from the manufacture of capital equipment are assumed to be negligible. It is assumed that nonmodified soybean oil is used in the process due to its acceptable performance in this application. When nonmodified soybean oil can be used, it is preferable for economic considerations, since additional processing increases the cost. Applications using chemically modified or heat-treated soybean oil require incorporation of additional inventory data as appropriate, which will increase the energy consumed along the life cycle.
-
Figure 1. Flow Diagrams of Soybean and Mineral Oil Life Cycles, Depicting Relevant Material and Energy Flows Documented in the Inventory. Transportation between processes is also included.
The amount of lubricant used in aluminum rolling is determined by experimental results, which are described below. It is assumed that all of the emissions associated with metal rolling will be the same for both lubricants, except for VOC emissions during the process. The lubricant system consists of oil-in-water emulsions with additive packages that are assumed to be similar for both types of lubricants. Other emissions associated with the use phase, such as energy use by the mill, are the same regardless of lubricant and are excluded from this study.
Aluminum Rolling Trials. Soybean lubricant performance was tested in an aluminum rolling manufacturing facility. The experimental data indicate that soybean lubricants consistently outperform the mineral oil standards (Miller, et al., 2006). In addition to obtaining greater reductions, the soybean lubricants were able to perform at temperatures and pressures that induced failure in the mineral oils. Surface quality was also improved by the soybean material.
Inventory Analysis. Inventory data on agricultural operations are available from numerous sources; however, values of mass and energy flows for agricultural operations can differ substantially, depending on annual fluctuations in crop yield, weather patterns, and agricultural operations. Evaluation of three databases and their underlying assumptions was necessary in order to establish an inventory for this study, and is described in detail in a prior publication (Miller and Theis, 2005). From the findings of this study, the Greenhouse gases, Regulated Emissions, and Energy use in Transportation (GREET) model is the basic framework for the inventory data in this study, due to its transparency and adaptability to user assumptions (Wang, 1999). Data gaps and disparities identified in the earlier study are included in this study. These include incorporation of nonpoint nitrogen emissions and variability in VOC emissions from hexane extraction, and inclusion of upstream emissions from soybean-specific agricultural chemicals.
From the analysis of available models, it is evident that significant variability can exist. Variability data were obtained via MCA, which is a tool that repeatedly and randomly selects values from probability distributions assigned to system parameters. This study used Crystal Ball Version 7.0 to run the MCA simulations. In addition to providing transparency and adaptability to user assumptions, GREET Version 1.6 contains variability information in a Crystal Ball format for energy generation processes. Modifications to the GREET model as well as determination of variability distributions for individual parameters are described in Miller, et al. (2006).
GREET 1.6 catalogues energy use and emissions for each stage of the agricultural process, including variability estimates. Detailed descriptions of the inventory calculations can be found in the GREET manual (Wang, 1999). While GREET provides valuable information pertaining to the agricultural sector, it is created for analysis of transportation fuels and needs to be modified for this analysis. Modifications to the GREET inventory include a separate nitrogen characterization model, which supplies aqueous nitrate emissions as well as N2O, NOx and NH3 emissions. A complete description of the nitrogen distribution model is detailed in an earlier paper (Miller, et al., 2006). Agricultural chemical manufacturing data for fertilizer production and application, lime, and crop-specific pesticides are also added (Landis, et al., 2006). VOC emissions resulting from the hexane extraction of soybean oil are modified from the original GREET values, using a distribution resulting from a best-fit regression of ten collected industry sources ( EPA, 1995). While other methods of oil extraction exist, hexane extraction is currently the most prominent and economically viable, so it is the extraction method used in this analysis. Carbon sequestration and end-of-life releases are also incorporated into the analysis.
A credit of 2.44 kg CO2/kg soybean oil is included in the farming inventory in this analysis, which is based on the assumption that 66% by mass of processed soybean oil derives from atmospheric carbon (Sheehan, et al., 1998). Both the Energy Information Administration (EIA) and Intergovernmental Panel on Climate Change (IPCC) assume that 50% of the carbon in lubricants is sequestered at the end-of-life in a solid state and 50% is released as emissions (EIA, 2004). This assumption seems valid since filter material containing waste lubricants is generally sent to sealed landfills. This study assumes that all end-of-life carbon emissions are CO2, although a study is currently underway to measure the respective VOC emissions of soybean and mineral oils during aluminum rolling. A sensitivity analysis on the final impact data shows this assumption does not affect the overall conclusions of the analysis. The end-of-life emissions for mineral oils are based on a lubricant carbon content equivalent to 1.95 kg CO2/kg mineral oil, with uncertainty bounds of -1%/+6%, as defined by the EIA (EIA, 2004).
Allocation. Many industrial processes generate more than one product; each product is responsible for a portion of the emissions generated during the process, although it is often unclear exactly how the inventory should be divided among the components in LCA. The choice of allocation scheme can be an important factor in LCA and can significantly impact the outcome (Vignon, et al., 1992). Allocation is usually conducted on a mass, energy, or market basis. In this study, all allocation is conducted on a mass basis at the process level and is described in detail below. Although market-based allocation is also a reasonable alternative, the interdependence of corn-soybean agriculture complicates the allocation scheme. Allocation of emissions on an energy basis is rejected for this study since the ultimate function of the product is not related to energy purposes.
Agricultural cycles provide some interesting dilemmas in the choice of allocation. Allocating on a mass basis, soybean oil is responsible for 18% of the inventory flows associated with soybean production (Sheehan, et al., 1998). Since soybeans are generally rotated with corn, allocating overall farm-scale emissions to soybeans is more complex. Corn and soybeans are grown in rotation to minimize fertilizer requirements for corn crops and maintain soil health. Even though fertilizers are generally applied only to corn, soybeans are an integral part of the nitrogen cycle in modern agriculture, due to their ability to fix nonreactive nitrogen, and reduce fertilizer requirements for the next corn crop. Inefficient fertilizer uptake during a corn rotation allows soybeans to utilize residual fertilizer and fix less nitrogen (Gentry, et al., 2001). To determine environmental impacts of corn and soybeans, a proportion of emissions must be allocated to each.
This project uses process-level allocation between corn and soybeans, which dictates that inventory flows particular to a certain product should be allocated only to that product (Wang, et al., 2003). Two methods for allocating nutrients between corn and soybean are presented:
- Treat the rotation as a system, and allocate total inventory flows for the rotation to corn and soybean according to an appropriate percentage.
- Treat corn and soybean growing seasons separately, allocating only inventory flows which occur within the growing season, irrespective of the nutrient cycle overlap.
The first method is a more precise system depiction; however, deciding on the appropriate percentage can be complicated and subjective. Even though fertilizer applied during corn years has a role in soybean nutrition, its primary purpose is for corn crops, and it is difficult to determine the impact fertilizer applied during corn years has on soybean growth. Ideally, this allocation would account for the nutrient flows based on soil nitrogen budgets, which would relate each crop’s contribution to total nitrogen fluxes. This type of information is not well understood, and difficult to convert to allocation schemes. In addition, allocating upstream emissions of applied fertilizer to soybeans does not seem appropriate, since fertilizers are generally applied only to corn crops.
The second scheme is easier to perform, although it may not reflect the mechanistic nature of the system. According to this allocation, the inventory would include only fertilizer and chemicals applied to each crop, and nutrient runoff and air emissions during their respective growing years. In this manner, all upstream emissions generated from fertilizer manufacturing and the subsequent nitrogen loads to the environment are allocated to corn. Soybeans are accountable for fertilizer applied during soybean years only, as well as nutrient leaching and air emissions during their growing season. While the synergistic nature of the corn-soybean rotation is not captured by this method, it is a reasonable allocation, since biological nitrogen fixation from soybeans is responsible for disruptions in the nitrogen cycle even without direct fertilizer application. Emissions associated with phosphorus fertilizer are allocated in a similar manner.
This study includes only nutrient runoff and chemicals applied during the soybean growing season; it does not include pesticides applied to corn. The only pesticide included in the inventory is glyphosate, since it is the predominant herbicide used on soybeans. Inventory data associated with lime, a soil amendment used to adjust soil pH, is allocated on an annualized basis depending on the percentage of soybean acreage in the study area (Landis, et al., accepted).
Calculating emissions associated with mineral oils was also conducted at the process level. Most lubricants are produced through a pathway similar to that of Residual Fuel Oil (RFO), which derives from the heavier fraction of crude separated via the initial distillation process (Wang, et al., 2003). Based on the product streams of a generic U.S. refinery, lubricant stocks are assumed to make up 7% by mass of the total product (Wang, et al., 2003), and refinery emissions are allocated accordingly.
Functional Unit and Use Phase Calculations. In LCA, products are compared on the basis of a functional unit that appropriately compares the performance of the products. The functional unit for the analysis is area of aluminum rolled. The relative performance of soybean and mineral oil must be determined in order to complete the analysis.
The described aluminum rolling performance experiments indicate superior performance by the soybean oil and improved surface quality of the metal. Soybean oil achieves greater reductions in metal thickness at lower temperatures than the traditional mineral oil, indicating that approximately 75% less lubricant can be used in an emulsion to achieve similar or improved results (Alcoa, Inc., 2004). A summary of the experimental data obtained at the aluminum rolling trial can be found in Miller, et al., 2006.
Impact Assessment. Several software tools are available to perform life cycle impact assessments (LCIA) (Goedkoop, 1995; Dreyer, et al., 2003; Jolliet, et al., 2003; Itsubo and Inaba, 2003; Bare, et al., 2003; Bare and Gloria, 2006). These vary by the inclusion of different impact categories, characterization factors, and the use of normative factors, such as weighting the importance of impacts to normalize an analysis (Bare and Gloria, 2006; Landis and Theis, submitted; Toffel and Marshall, 2004). LCIA uses characterization factors to determine the amount of impact a given amount of emission will have.
The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts (TRACI) is used in this analysis because of characterization factors specific to location-specific U.S. data as well as its midpoint approach to impact assessment. TRACI provides characterization factors for 12 impact categories: ozone depletion, global warming, acidification, eutrophication, photochemical smog, ecotoxicity, human health criteria air pollutants, human health cancer, human health noncancer, fossil fuel, land use, and water use (Bare, et al., 2003). The method for measuring each environmental impact is different and is discussed in detail in the literature (Bare, et al., 2003; Bare and Gloria, 2006; Landis and Theis, submitted; Toffel and Marshall, 2004; Norris, 2003).
Due to constraints of data availability and consistency, the only impacts that are calculated in this analysis are not compound-specific. Impacts pertaining to aquatic and human health depend greatly on emissions of individual chemicals and are not calculated to avoid biasing the results of the assessment due to lack of uniformity in the data. Reported human health impacts pertain only to those resulting from exposure to the criteria pollutants and do not indicate the acute or chronic health issues from toxic or carcinogenic substances. This may be a significant issue from a worker exposure issue and should be addressed in the future when compound-specific data are available. Ozone depletion potential is also not considered since neither inventory tracks specific chemical information on ozone-depleting substances.
Characterization factors for climate change are site-generic, meaning the location of the emission does not affect the magnitude of the impact value. Other environmental impacts vary depending on the location of relevant emissions. Using TRACI’s site-specific characterization factors, better impact assessments can be made than with generalized factors (Bare, et al., 2003). Information for determining location-specific impact factors can be found in Miller, et al., 2006.
Results and Discussion
Inventory Results. MCA allows calculation of probability distributions for each inventory flow. Figure 2 presents the inventory data for mineral oil and soybean lubricants on a mass of emission/mass of lubricant basis. The contributions of each stage of the life cycle to total process emissions are shown with variability bars indicating the values within a 10–90% probability range, with the median value of the total indicated above each emission.
a)
b)
-
Figure 2. Contributions of Each Life Cycle Stage to Total Emissions, With Variability Bars Showing the 10–90% Probability Range. Indicated values represent the median of the distribution. The life cycle stages of mineral oils are: crude oil recovery, refining combustion emissions, refining noncombustion emissions (Refining NC), and transportation to production facility. For soybean oil, the life cycle stages represented are farming equipment, fertilizer production and transportation, other (upstream production and transportation of lime and pesticides), on-field emissions, transportation to processing plant, soybean processing, and transportation to production facility.
Median fossil energy consumption for soybean oil is 5.27 MJ/kg oil. Oil processing is responsible for 55% of energy consumption, followed by farming operations (28%), and transportation (11%). Upstream processes make up the remainder. Oil processing is the most energy intensive stage because of the natural gas consumed drying the beans and generating steam for the extraction process. Internal energy for soybean oil is not included since it derives from solar energy. In contrast, the life cycle of mineral oil is responsible for the consumption of 44.78 MJ/kg oil of fossil energy, almost ten times that of soybean oil. The majority (91%) is from fossil energy embodied within the product. The internal energy of the mineral oil is included because it derives from a source of fossil energy; soybean oil does not. Of the remaining 4.24 MJ/kg oil consumed throughout the manufacture of the lubricant, 51% occurs during the refinery stage, 42% during crude oil extraction, and 7% during transportation.
As seen in Figure 2a, crude oil extraction is responsible for the majority of air emissions in mineral oil production, especially for methane, where fugitive emissions from oil extraction and transportation to the refinery are abundant. Life cycle emissions for soybean oil are distributed among various processes. Use of farming equipment and transportation are major contributors to NOx, SOx, and CO2 emissions. Soybean oil processing is the dominant contributor to VOC emissions, due to process emissions during hexane extraction, which separates oil from meal. On-field processes make up the majority of N2O emissions, and contribute substantially to NOx. These emissions are the byproducts of denitrification and nitrification reactions of fertilizer and mineralized soil nitrogen. In addition, on-field processes include CO2 sequestration, which results in net negative CO2 emissions due to the assumption that 50% of the carbon is re-released to the atmosphere, while 50% stays in a stable solid state. Even if it is assumed that the carbon is re-released in its entirety at the end of life, the carbon emissions are offset by the initial carbon sequestration, resulting in only the CO2 emitted by fossil fuel combustion throughout the life cycle.
System Variability and Error. The variability bars depicted in Figure 2 demonstrate the 10–90% range of possible values for these data. In the mineral oil inventory, aggregate PM10 and SOx emissions vary by over 100% between the 10th and 90th percentiles. Inventory components from particular life cycle stages also demonstrate significant variability, but these do not greatly affect the cumulative variability of the life cycle. For instance, refinery emissions from noncombustion processes (which include fugitive emissions, blowdown systems, thermal and catalytic cracking, and catalyst regeneration) possess variability ranges over several orders of magnitude for VOC, CO, PM10, NOx, and SOx emissions; however, these do not greatly impact the aggregate inventory for these compounds since they have relatively small values when compared to combustion processes. Certain soybean oil components demonstrate significantly greater variability. For soybean oil, aggregate VOC and N2O emissions vary by over 300% over the 10–90% range due to process variability in soybean oil extraction, and variability in N2O emissions from on-field processes.
Three sources of variability are contained within inventory data: operational variability (e.g., differences in operational practices such as operating machinery more efficiently); natural variability (e.g., nutrient cycles in agricultural systems, composition of crude oil, and methane released from oil fields), and systematic variability within the model (e.g., different researchers compiling data and assigning probability distributions to values and the use of different models to compile data). The large variability in nitrogen compounds in the soybean inventory is an artifact of the natural variability in agricultural systems, whereas the variability associated with SOx emissions in the mineral oil inventory is largely due to operational variability. Systematic variability differences appear in this inventory from the combination of GREET and nutrient models, as well as other supplemental data that were added. Systematic variability also encompasses uncertainty in inventory ranges and inventory flows that are not assigned variability distributions in this model due to limitations in data availability. These include CH4 emissions from crude oil recovery, CO2 sequestration during on-field processes, and end-of-life CO2 emissions. While the authors have tried to limit the systematic variability wherever possible, it is an inevitable result of combining multiple data sources. Operational and natural variability cannot be reduced in this analysis, since they are a result of actual data occurring within the system.
Inventory Comparisons. Figure 3 presents life cycle inventory results on a mass of emission per area of metal basis. If performance data are neglected and comparison is conducted using similar use rates, soybean oil is responsible for greater life cycle emissions for VOC, NOx, N2O, NO3-, and total P, and for significantly fewer emissions of CO2 and CH4, as well as significantly less fossil energy consumption. Due to the intensive participation of agriculture in the nitrogen cycle, soybean oil has notably higher NOx, N2O and aqueous nutrient emissions than mineral oils on a mass basis. Hexane extraction, which dominates the VOC emissions for soybean oil, is responsible for significantly greater VOC emissions than the mineral oil. The lower CO2 and CH4 emissions for soybean oil are due to the sequestration of carbon during soybean farming, and the methane releases during crude oil extraction. The comparative emissions for CO, PM10, and SOx have similar values for soybean and mineral oils, and the variability ranges associated with these emissions overlap.
-
Figure 3. Comparative Life Cycle Inventory Results for Soybean and Mineral Oils, Showing the 10–90% Probability Range Per 100 m2 of Aluminum Production. Two soybean cases are shown: no change in performance, as well as the expected 75% reduction in the amount of oil used to produce of a similar quantity of aluminum. All values are reported as kg/100 m2 rolled aluminum, except for CO2 and NO3- emissions, which are in Mg/100 m2.
Experimental data suggest that approximately 75% less soybean oil will be required in the rolling process, resulting in lower life cycle emissions proportional to the reduction in the total amount of oil consumed. Interestingly, the total net carbon sequestration is lower for the improved performance case, since there is less oil consumed throughout the process and therefore less potential for sequestration. LCA cautions against direct emissions comparisons, since the data can be misleading. The emissions must be first translated into appropriate impact metrics for proper comparison.
Impact Results. An LCIA is compiled using the inventory data for the life cycles of soybean and mineral oil and location-specific characterization factors from TRACI. Table 1 reports the impact results for producing 100,000 m2 of rolled aluminum, indicating the median and the 10–90% probability range.
Table 1. Impact Results for Mineral Oil and Four Use-Phase Soybean Oil Scenarios, for 100,000 m2 Aluminum. Median value is shown with values in parenthesis representing the 10-90% range.
|
Impact Units |
Mineral Oil |
Soybean Oil Mass Basis |
Soybean Oil Performance Basis – generated from experimental data |
Reduction in lubricant used during the use phase |
|
- |
No Change |
76% |
Acidification |
moles H+ equivalents/ 100,000m2 |
437 (374-525) |
753 (540-1026) |
166 |
Eutrophication |
kg N equivalents/ 100,000m2 |
0.26 (0.23-0.30) |
154 (52.7-291) |
27.9 |
Human Health |
DALYs/ 100,000m2 |
1.14 (0.92-1.43) |
1.38 (1.14-1.67) |
0.31 |
Smog Formation |
kg NOx equivalents/ 100,000m2 |
8.45 (7.59-9.59) |
15.5 (10.8-21.0) |
3.42 |
Global Warming |
kg CO2 equivalents/ 100,000m2 |
5731 (5577-5900) |
-2343 (-2924, -1583) |
-538 |
Fossil Fuel Use |
GJ/ 100,000m2 |
182 (179-184) |
20.7 (20.1-21.4) |
4.92 |
For easier comparison among the data, the results have been normalized to the impacts associated with mineral oil, as seen in Figure 4. In this figure, variability bars are included only for soybean oil data, since normalized results for mineral oil have a value of 1 by definition. The variability associated with normalized soybean oil scenarios depicts the variability of the impact categories relative to mineral oil. Soybean oil shows negative impact profiles in the Climate Change category. This is caused by sequestration of carbon dioxide from the atmosphere, and results in net improvement instead of an impact. As discussed earlier, if the assumption of net carbon sequestration is rejected, the climate change potential for soybean oil will still be significantly less than mineral oil due to soybeans’ participation in photosynthesis. In addition, fossil fuel consumption for soybean oil is less than 10% of mineral oil. Eutrophication is several orders of magnitude greater for soybean lubricants regardless of production improvements. This result is not surprising, given the participation of biomass in the nitrogen cycle, and direct emissions of NO3- and phosphorus into watersheds.
Figure 4. LCIA Results Normalized to Mineral Oil
Tradeoff Analysis. The results show that with improved performance, soybean oil lubricants result in significant climate change and fossil fuel use benefits, but have increased impacts in eutrophication potential.
Fossil fuel combustion is a controlling factor determining the relative environmental impacts of most products and processes due to the quantity of emissions. The inventory emissions for the criteria pollutants are directly related to fossil fuel combustion. In general, products that have greater fossil fuel combustion throughout their life cycle will generate more emissions and cause the most deleterious environmental impacts. When combustion of fossil fuels is greater, the resultant emissions and subsequent impacts are also greater.
Modern agriculture is one of the exceptions to this linkage between fossil fuels and the emissions inventory. Agricultural systems rely largely on natural fluxes of carbon and nitrogen to produce biomass. While combustion processes are an integral part of modern agricultural systems, natural cycles that extract carbon and nitrogen from the air can contribute greatly to the overall fluxes within the system. As this analysis shows, the more biomass used within a system that has net carbon sequestration, the greater the climate change benefit. The more soybean oil that is used, the more carbon sequestration and subsequent climate change benefit can be realized. Conversely, as consumption of biomass increases, the flux of nitrogen into the environment also intensifies. Nitrogen contributes to many impacts in its reactive forms, with emissions of NOx, N2O, and NO3- primary factors in eutrophication, acidification, and smog formation, and contributors to climate change and human health impacts (Galloway, et al., 2003).
The carbon and nitrogen cycles in agriculture are inextricably linked. As atmospheric carbon uptake increases in soybeans, so does biological nitrogen fixation in order to reach the typical 30:1 C:N Redfield ratio for foliage plants (Megonigal, 2002). As shown in Figure 4, there is an inverse relationship between climate change benefits and eutrophication impacts. Essentially, the substitution of soybean oils for mineral oils results in a fundamental tradeoff between the impacts of carbon and those of nitrogen. Increased biomass substitution may assist in ameliorating the global issue of climate change, but may occur at the detriment of regional impacts, such as eutrophication and hypoxia. Consideration of increased nitrogen flux and the subsequent environmental impacts should be an important factor in decisions concerning widespread adoption of bio-based products.
A correlation between the aggregate amount of carbon sequestered and the nitrogen compounds emitted can be determined. Based on the assumptions of this analysis, the ratio of nitrogen compounds emitted in soybean fields during on-field processes to the rate of CO2 sequestration via photosynthesis is: 14 g NO3-/kg CO2 sequestered (10–90% range 4.4–27.0); 0.33 g N2O/kg CO2 sequestered (0.13–0.59); 0.51 g NOx/kg CO2 sequestered (0.13–0.59). For the complete inventory data specific to soybean lubricants and the relative amount used in rolling, including all process emissions and sequestered carbon released at the end of life, the ratio of nitrogen compounds emitted for carbon dioxide sequestered is: 36 g NO3-/kg CO2 sequestered (13.2–76.3); 0.61 g N2O/kg CO2 sequestered (0.37–1.69); 3.42 g NOx/kg CO2 sequestered (2.46–4.52).
Only durable goods that maintain carbon in a solid state at the end of life can claim credit for sequestration. If the carbon stored during photosynthesis is released into the atmosphere completely via combustion or biodegradation, there is no net sequestration. Instead, there is a net release of carbon from fossil sources used to produce the biomass. In these cases, there is no carbon sequestration, and the ratio of the amount of reactive nitrogen released to the amount of displaced fossil carbon becomes much greater. Instead of creating climate change benefits, combustion products derived from biomass merely slow the flux of fossil carbon into the atmosphere, while accelerating the flux of reactive nitrogen.
Nutrient fluxes are often neglected in agricultural LCAs because of their variability and difficulty integrating them into life cycle inventories (Sheehan, et al., 1998). However, as the results of this analysis indicate, inclusion of this information is essential to understand and evaluate the tradeoffs that arise for bio-based products. Eutrophication impacts are significantly greater and climate change is lower for soybean-based lubricants, but the extent of the remainder of impacts for soybean oil is largely dependent on the amount of oil used in the process.
References:
Lynd LR, Wyman CE, Gerngross TU. Biocommodity Engineering. Biotechnology Progress 1999;15(5):777-793.
Mohanty AK, Misra M, Drzal LT. Sustainable bio-composites from renewable resources: opportunities and challenges in the green materials world. Journal of Polymers and the Environment 2002;10(1-2):19-26.
Wilke D. What should and what can biotechnology contribute to chemical bulk production? FEMS Microbiology Reviews 1995;16(2-3):89-100.
Dale BE. Biobased industrial products: bioprocess engineering when cost really counts. Biotechnology Progress 1999;15(5):775-776.
Energy Information Administration. Annual Energy Review 2004. Department of Energy, 2004.
Honary LA. An investigation of the use of soybean oil in hydraulic systems. Bioresource Technology 1996;56:41-47.
Ash M, Dohlman E. Oil crops outlook. United States Department of Agriculture Economic Research Service. 2006, pp. 1-16.
Kinney AJ. Production of specialised oils for industry, in plant lipid biosynthesis: fundamental and agricultural applications. In: J.L. Harwood, Ed. Cambridge, United Kingdom: Cambridge University Press, 1998, pp. 273-285.
Pearson SL, Spagnoli JE. Environmental lubricants—an overview of onsite applications and experience. Lubrication Engineering 2000:40-45.
Honary LA. Biodegradable/biobased lubricants and greases. Machinery Lubrication, September 2001.
McManus MC, Hammond GP, Burrows CR. Life-cycle assessment of mineral and rapeseed oil in mobile hydraulic systems. Journal of Industrial Ecology 2003;7(3-4):163-177.
Kassfeldt E, Goran D. Environmentally adapted hydraulic oils. Wear 1997;207:41-45.
Pal M, Singhal S. Environmentally adapted lubricants, part II. Hydraulic fluids. Journal of Synthetic Lubrication 2000;17(3):219-224.
Vignon BW, et al. Life-cycle assessment: inventory guidelines and principles. U.S. Environmental Protection Agency, Battelle and Franklin Associates Ltd.: Cincinnati, 1992.
MacLean HL, et al. A life-cycle comparison of alternative automobile fuels. Journal of the Air & Waste Management Association 2000;50(10):1769-1779.
Sheehan J, et al. Life cycle inventory of biodiesel and petroleum diesel for use in an urban bus, final report. National Renewable Energy Laboratory: Golden, CO, 1998.
Fu GZ, Chan AW, Minns DE. Life cycle assessment of bio-ethanol derived from cellulose. International Journal of Life Cycle Assessment 2003;8(3):137-141.
Shapouri H, Duffield JA, Graboski M. Estimating the net energy balance of cornethanol. U.S. Department of Agriculture, Economic Research Service: Washington, DC, 1995.
Lave L, et al., Life-cycle analysis of alternative automobile fuel/propulsiontechnologies. Environmental Science & Technology 2000;34(17):3598-3605.
Wang M. GREET—Greenhouse gases, regulated emissions, and energy use in transportation. Argonne National Laboratory, Argonne, IL, 1999.
Landis AE, Miller SA, Theis TL. A probability approach to incorporating natural variability and nutrient flows into the agricultural inventory for biobased product life cycle assessments. Environmental Science & Technology (accepted, 2006).
EPA. Emission factor documentation for AP-42 section 9.11.1 vegetable oil processing. EPA Office of Air Quality Planning and Standards, Research Triangle Park, NC, 1995, pp. 1-44.
Energy Information Administration. Documentation for emissions of greenhouse gases in the United States 2002. U.S. Department of Energy, Washington, DC, 2004, pp. 1-265.
Gentry LE, et al. Source of the soybean N credit in maize production. Plant and Soil 2001;236:175-184.
Wang M, Lee H, Molburg J. Allocation of energy use in petroleum refineries topetroleum products: implications for life-cycle energy use and emission inventory of petroleum transportation fuels.The International Journal of Life Cycle Assessment 2003;8(5):1-11.
Alcoa Inc. Pilot mill aluminum rolling trial. Alcoa Technical Center, Pittsburgh, PA, 2004.
Goedkoop M. The EcoIndicator 95. Pre consultants, Amersfoort, The Netherlands, 1995.
Dreyer LC, Niemann AL, Hauschild MZ. Comparison of three different LCIA methods: EDIP97, CML2001, and Eco-indicator 99. Does it matter which one you choose? International Journal of Life Cycle Assessment 2003;8(4):191-200.
Jolliet O, et al. IMPACT 2002+: A new life cycle impact assessment methodology. The International Journal of Life Cycle Assessment 2003;8(6):324-330.
Itsubo N, Inaba A. A new LCIA method: LIME has been completed. International Journal of Life Cycle Assessment 2003;8(5):305.
Bare J, et al. TRACI: the tool for the reduction and assessment of chemical and other environmental impacts. Journal of Industrial Ecology 2003;6(3-4):49-78.
Bare JC, Gloria TP. Critical analysis of the mathematical relationships and comprehensiveness of life cycle impact assessment approaches. Environmental Science & Technology 2006;40(4):1104-1113.
Landis AE, Theis TL. Comparison of impact assessment tools: TRACI, CML, and IMPACT 2002+.Journal of Industrial Ecology (submitted).
Toffel MW, Marshall JD. Improving environmental performance assessment: a comparative analysis of weighting methods used to evaluate chemical release inventories. Journal of Industrial Ecology 2004;8(1-2):143-172.
Norris GA. Impact characterization in the tool for the reduction and assessment of chemical and other environmental impacts. Journal of Industrial Ecology 2003;6(3-4):79-99.
Galloway JN, et al. The nitrogen cascade. BioScience 2003;53(4):341-356.
Megonigal JP. Global natural cycles. In: SER Center, ed. Encyclopedia of Life Support Systems, Edgewater, MD: EOLSS Publishers, 2002.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 2 publications | 2 publications in selected types | All 2 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Miller SA, Theis TL. Comparison of life-cycle inventory databases: a case study using soybean production. Journal of Industrial Ecology 2006;10(1-2):133-147. |
R831521 (2004) R831521 (Final) |
Exit |
|
Miller SA, Landis AE, Theis TL. Use of Monte Carlo analysis to characterize nitrogen fluxes in agroecosystems. Environmental Science & Technology 2006;40(7):2324-2332. |
R831521 (2004) R831521 (Final) |
Exit Exit |
Supplemental Keywords:
life cycle assessment, life cycle inventory, bio-based production, soybean oil, aluminum rolling, metal working, life cycle comparison, inventory analysis, soybean agriculture, product substitution, sustainable engineering,, RFA, Scientific Discipline, Air, POLLUTANTS/TOXICS, Sustainable Industry/Business, Chemical Engineering, Sustainable Environment, Environmental Chemistry, cleaner production/pollution prevention, Chemicals, climate change, Air Pollution Effects, Technology for Sustainable Environment, Ecological Risk Assessment, Atmosphere, life cycle analysis, environmental monitoring, environmentally conscious manufacturing, petrolubricant substitutes, air pollution control, VOC removal, aluminum rolling, metal casting industry, life cycle assessment, biolubricants, carbon emissions credit trading, Volatile Organic Compounds (VOCs)Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.