6.2.3.1 Goal Definition and Scoping

The LCA process can be used to determine the potential environmental impacts from any product, process, or service. The goal definition and scoping phase will determine the time and resources needed. The defined goal and scope will guide the entire process to ensure that the most meaningful results are obtained. Every decision made during goal definition and scoping impacts either how the study will be conducted or the relevance of the final results. Goal definition and scoping will result in the determination of the following:

1. The goal(s) of the project
2. The type of information is needed to inform the decision‐makers
3. How the data should be organized and the results displayed
4. What will or will not be included in the LCA
5. The accuracy required of the data
6. Ground rules for performing the work

Each decision made in the fleshing out of these six areas has an impact on the LCA process, as explained in Sections 6.2.3.1 through 6.2.3.6.

6.2.3.2 Define the Goal(s) of the Project

The primary goal of the LCA is to choose the best product, process, or service with the least effect on human health and the environment. There may also be secondary goals for performing an LCA, which would vary depending on the type of project. Some typical secondary goals are as follows:

• To prove one product is environmentally superior to a competitive product
• To identify stages within the life cycle of a product or process where a reduction in resource use and emissions might be achieved
• To determine the impacts to particular stakeholders or affected parties
• To establish a baseline of information on a system's overall resource use, energy consumption, and environmental loadings
• To help guide the development of new products, process, or activities toward a net reduction of resource requirements and emissions

6.2.3.3 Determine the Type of Information Needed to Inform the Decision‐Makers

LCA can help answer a number of important questions. Identifying the questions that the decision‐makers care about will help define the study parameters. Some examples include the following:

• What is the impact to particular interested parties and stakeholders?
• Which product or process causes the least environmental impact (quantifiably) overall or in each stage of its life cycle?
• How will changes to the current product/process affect the environmental impacts across all life cycle stages?
• Which technology or process causes the least amount of acid rain, smog formation, or damage to local trees (or any other impact category of concern)?
• How can the process be changed to reduce a specific environmental impact of concern (e.g. global warming)?

Once the appropriate questions have been identified, the types of information needed to answer them will be apparent.

6.2.3.4 Determine How the Data Should Be Organized and the Results Displayed

LCA practitioners like to organize data in terms of a functional unit that appropriately describes the function of the product/process being studied. Comparisons between products and processes must be made on the basis of the same function, quantified by the same functional unit. This ensures that the activities being compared are true substitutes for each other. Careful selection of the functional unit to measure and display the LCA results will improve the accuracy of the study and the usefulness of the results.

An LCA study comparing two types of wall insulation to determine environmental preferability must be evaluated on the same function, the ability to decrease heat flow. Six square feet of 4‐in. thick insulation Type A is not necessarily the same as 6 ft² of 4‐in. thick insulation Type B. Insulation type A may have an R factor equal to 10, whereas insulation type B may have an R factor equal to 20. Therefore, types A and B do not provide the same amount of insulation and cannot be compared on an equal basis. If Type A decreases heat flow by 80%, you must determine how thick Type B must be to also decrease heat flow by 80%.

6.2.3.5 What Will and Will not Be Included

Ideally, an LCA includes all four stages of a product or process life cycle: raw material acquisition, manufacturing, use/reuse/maintenance, and recycle/waste management. These product stages are explained in more detail in the following. To determine whether one or all of the stages should be included in the scope of the LCA, the following must be assessed: the goal of the study, the required accuracy of the results, and the available time and resources. Figure 6.3 presents a set of life cycle stages that could be included in a manufacturing project related to treatment technologies. Note the “system boundary,” which encompasses all aspects of the LCA. Additional examples of the four life cycle stages are explained in more detail in the following.

Raw Materials Acquisition 

The life cycle of a product begins with the removal of raw materials and energy sources from the Earth. For instance, the harvesting of trees or the mining of nonrenewable materials would be considered raw materials acquisition. Transportation of these materials from the point of acquisition to the point of processing is also included in this stage (USEPA 1993).

6.2.3.6 Accuracy Required of the Data

The required level of data accuracy for the project depends on the use of the final results and the intended audience. (Will the results be used to support decision making in an internal process? In a public forum?) For example, if the intent is to use the results in a public forum to support product/process selection to a local community or regulator, then estimated data or best engineering judgment may not be accurate enough to justify basing policy decisions on them. In contrast, if the LCA is for internal decision‐making purposes only, then estimates and best engineering judgment may be applied more frequently. This may reduce the overall cost and time required to perform the LCA, as well as enable completion of the study in the absence of precise, first‐hand data. The criticality of the decision to be made and the amount of money at stake also come into play in determining the required level of data accuracy.

6.2.3.7 Ground Rules for Performing the Work

Prior to moving on to the inventory analysis phase it is important to define some of the logistical procedures for the project.

1. Documenting assumptions – All assumptions or decisions made throughout the entire project must be reported alongside the final results of the LCA project. If assumptions are omitted, the final results may be taken out of context or easily misinterpreted. As the LCA process advances from phase to phase, additional assumptions and limitations to the scope may be necessary to accomplish the project with the available resources.
2. Quality assurance procedures – Quality assurance procedures are important to ensure that the goal and purpose for performing the LCA will be met at the conclusion of the project. The level of quality assurance procedures employed for the project depends on the available time and resources and how the results will be used. If the results are to be used in a public forum, a formal review process is recommended. Evaluators might include internal and external LCA experts and interested parties whose support of the final results is sought. If the results are to be used for internal decision‐making purposes only, then an internal reviewer who is familiar with LCA practices and is not associated with the LCA study may effectively meet the quality assurance goals. A formal statement from each reviewer documenting his or her assessment of each phase of the LCA process should be included with the final project report.
3. Reporting requirements – To ensure that the LCA meets appropriate expectations, participants should know from the outset how the final results are to be documented and exactly what is to be included in the final report. When reporting the final results, or results of a particular LCA phase, it is important to thoroughly describe the methodology used in the analysis. The report should explicitly define the systems analyzed and the boundaries that were set. The basis for comparison among systems and all assumptions made in performing the work should be clearly explained. The presentation of results should be consistent with the purpose of the study. The results should not be oversimplified solely for the purposes of presentation.

6.2.4.1 Step 1: Develop a Flow Diagram

A flow diagram is a tool to map the inputs and outputs to a process or system. The system boundary varies for the LCA, as established in the goal definition and scoping phase and expanded to include process inputs and outputs serves as the system boundary for the flow diagram. Unit processes inside the system boundary link together to form a complete life cycle picture of the required inputs and outputs (material and energy) to the system. Figure 6.4 illustrates the components of a generic unit process within a flow diagram for a given system boundary. The more complex the flow diagram, the greater the accuracy and utility of the results. Unfortunately, increased complexity also means more time and resources must be devoted to this step, as well as the data collecting and analyzing steps.

Flow diagrams are used to model all alternatives under consideration (e.g. both a baseline system and alternative systems). For a comparative study, it is important that both the baseline and alternatives use the same system boundary and are modeled to the same level of detail. If not, the accuracy of the results may be skewed.

6.2.4.2 Step 2: Develop an LCI Data Collection Plan

An LCI data collection plan ensures that the quality and accuracy of data, characterized as part of the goal definition and scoping phase, meet the expectations of the decision‐makers.

Key elements of a data collection plan include the following:

• Defining data quality goals
• Identifying data sources and types
• Identifying data quality indicators
• Developing a data collection worksheet and checklist

Define data quality goals – Data quality goals provide a framework for balancing available time and resources against the quality of the data required to make a decision regarding overall environmental or human health impact (USEPA 1989a). Data quality goals, which are closely linked to overall study goals, both aid LCA practitioners in structuring an appropriate approach to data collection and serve as data quality performance criteria.

Although the number and nature of data quality goals necessarily depend on the level of accuracy required for a given LCA, the following list of hypothetical data quality goals is typical. Site‐specific data are required for raw materials and energy inputs, water consumption, air emissions, water effluents, and solid waste generation. Approximate data values are adequate for the energy data category. Air emission data should be representative of similar sites in the United States. A minimum of 95% of the material and energy inputs should be accounted for in the LCI.

Identify data quality indicators – Data quality indicators are benchmarks against which the collected data can be measured to determine if data quality requirements have been met. Selection depends on which of the available indicators are most appropriate and applicable to the specific data sources being evaluated. Examples of indicators are precision, completeness, representativeness, consistency, and reproducibility.

Identify data sources and types – For each life cycle stage, unit process, or type of environmental release, the data source and/or type that will provide sufficient accuracy and quality to meet the study’s goal is specified. Doing this prior to data collection helps to reduce costs and the time required to collect the data. Data sources include

• meter readings from equipment operating logs or journals; industry data reports, databases, or consultant’s laboratory test results; government documents, reports, and databases; other publicly available databases or clearinghouses; journals, and papers, books; reference books; patents; and trade associations. Related LCI studies are also useful, as are equipment and process specifications best engineering judgment.

Examples of data types include full measured, modeled, and sampled data; non site‐specific (i.e. surrogate) data; non‐LCI data (i.e. not intended for use in an LCI); and vendor data.

The required level of aggregated data should also be specified. For example, the reader should be able to ascertain quickly whether data are representative of one process or several processes.

Develop a data collection worksheet and checklist – The LCI checklist should cover most of the decision areas in the performance of an inventory. This document can be prepared to guide data collection and validation and can enable construction of a database to store collected data electronically. The following general decision areas should be addressed on the inventory checklist:

• purpose of the inventory
• system boundaries
• geographic scope
• types of data used
• data collection procedures
• data quality measures
• computational model construction
• presentation of results

All inputs and outputs for each process modeled in the flow diagram should be recorded on an accompanying data worksheet.

The checklist and worksheet are valuable tools for ensuring completeness, accuracy, and consistency. They are especially important for large projects when several people collect data from multiple sources. The checklist and worksheet should be tailored to meet the needs of a specific LCI.

6.2.4.3 Step 3: Collect Data

The flow diagram(s) developed in step 1 provides the road map for data to be collected. Step 2 specifies the required data sources, types, quality, accuracy, and collection methods. Step 3 consists of finding and filling in the flow diagram and worksheets with numerical data. This may not be a simple task. If some data are difficult or impossible to obtain, and available data are difficult to convert to the appropriate functional unit; therefore, the system boundaries or data quality goals of the study will have to be refined to describe the results that can reliably be obtained from the data available. This iterative process is common for most LCAs.

Data collection efforts involve a combination of research, site‐visits, and direct contact with experts which generate large quantities of data. An electronic database or spreadsheet can be useful to hold and manipulate the data. Alternatively, it may be more cost effective to buy a commercially available LCA software package (see Section 6.3). Prior to purchasing an LCA software package the decision‐makers or LCA practitioner should insure that it will provide the level of data analysis required.

A second method to reduce data collection time and resources is to obtain non‐site specific inventory data. Several organizations have developed databases specifically for LCA that contain some of the basic data commonly needed in constructing an LCI. Some of the databases are sold in conjunction with LCI data collection software; others are stand‐alone resources. Many companies with proprietary software also offer consulting services for LCA design.

6.2.4.4 Step 4: Evaluate and Document the LCI Results

When the data have been collected and organized, the accuracy of the results must be verified. In documenting the results of the LCI, it is important to thoroughly describe the methodology used in the analysis, define the systems analyzed and the boundaries that were set, and state all assumptions made in performing the inventory analysis. Use of the checklist and worksheet (see step 2) supports a clear process for documenting this information. The outcome of the inventory analysis is a list containing the quantities of pollutants released to the environment and the amount of energy and materials consumed. The information can be organized by life cycle stage, by media (air, water, land), by specific process, or any combination thereof that is consistent with the ground rules.

If the sensitivity of the LCI data collection efforts has not been properly determined before the next stage, life cycle impact assessment (LCIA), is begun, the LCA itself may have to be repeated because the data are found to be insufficient to permit the drawing of the desired conclusions.

6.2.5.1 Why Conduct an LCIA?

Although much can be learned about a process by considering LCI data, an LCIA provides a more precise basis to make comparisons. Thus, we know that large releases of both carbon dioxide and methane are harmful; an LCIA can determine whether 9000 T of CO₂ or 5000 T of methane would have the greater potential impact. Using science‐based characterization factors, an LCIA can calculate the impacts each environmental release has on problems such as smog or global warming. An impact assessment can also incorporate value judgments. In an air non‐attainment zone, for example, air emissions could be of relatively higher concern than the same emission level in a region with better air quality (ISO 2000).

6.2.5.2 Key Steps of a LCIA

The following steps comprise an LCIA.

1. Selection and definition of impact categories – identifying relevant environmental impact categories (e.g. global warming, acidification, terrestrial toxicity).
2. Classification – assigning LCI results to the impact categories (e.g. classifying CO₂ emissions to global warming).
3. Characterization – modeling LCI impacts within impact categories using science‐based conversion factors (e.g. modeling the potential impact of CO₂ and methane on global warming).
4. Normalization – expressing potential impacts in ways that can be compared (e.g. comparing the global warming impact of CO₂ and methane for the two options).
5. Grouping – sorting or ranking the indicators (e.g. sorting the indicators by location: local, regional, and global).
6. Weighting – emphasizing the most important potential impacts.
7. Evaluating and reporting LCIA results – gaining a better understanding of the reliability of the LCIA results.

The International Organization of Standardization standard for conducting an impact assessment states that impact category selection, classification, and characterization are mandatory steps for an LCIA and data evaluation (step 7) (ISO 1998a). Whether the other steps are used will depend on the goal and scope of the study.

Step 3: Characterization 

Impact characterization uses science‐based conversion factors, called characterization factors, to convert and combine the LCI results into representative indicators of impacts to human and ecological health. Characterization factors also are commonly referred to as equivalency factors. Characterization factors translate different LCI inputs – for example, the toxicity data for lead, chromium, and zinc – into directly comparable impact indicators. With such data in hand, estimates of the relative terrestrial toxicity of these metals could be made.

Impact category	Scale	Relevant LCI data (i.e. classification)	Common characterization factor	Description of characterization factor
Global warming	Global	Carbon dioxide (CO2) Nitrogen dioxide (NO2) Methane (CH4) Chlorofluorocarbons (CFCs) Hydrochlorofluorocarbons (HCFCs) Methyl bromide (CH3Br)	Global warming potential	Converts LCI data to carbon dioxide (CO2) equivalents Note: global warming potentials can be 50, 100, or 500 year potentials
Stratospheric ozone depletion	Global	Chlorofluorocarbons (CFCs) Hydrochlorofluorocarbons (HCFCs) Halons Methyl bromide (CH3Br)	Ozone depleting potential	Converts LCI data to trichlorofluoromethane (CFC‐11) equivalents
Acidification	Regional local	Sulfur oxides (SOx) Nitrogen oxides (NO) Hydrochloric acid (HCl) Hydrofluoric acid (HF) Ammonia (NH4)	Acidification potential	Converts LCI data to hydrogen (H+) ion equivalent
Eutrophication	Local	Phosphate (PO4) Nitrogen oxide (NO) Nitrogen dioxide (NO2) Nitrates Ammonia (NH4)	Eutrophication potential	Converts LCI data to phosphate (PO4) equivalents
Photochemical smog	Local	Non‐methane hydrocarbon (NMHC)	Photochemical oxidant creation potential	Converts LCI data to ethane (C2H6) equivalents
Terrestrial toxicity	Local	Toxic chemicals with a reported lethal concentration to rodents	LC50	Convert LC50 data to equivalents
Aquatic toxicity	Local	Toxic chemicals with a reported lethal concentration to fish	LC50	Convert LC50 data to equivalents
Human health	Global, regional, local	Total releases to air, water, and soil	LC50	Convert LC50 data to equivalents
Resource depletion	Global, regional, local	Quantity of minerals used Quantity of fossil fuels used	Resource depletion potential	Converts LCI data to a ratio of quantity resource used versus quantity of resource left in reserve
Land use	Global, regional, local	Quantity disposed of in a landfill	Solid waste	Converts mass of solid waste into volume using an estimated density

The impact categories listed in Table 6.1 have many possible endpoints, including the following:

Global impact
Global warming	Polar melt, soil moisture loss, longer seasons, forest loss/change, and change in wind and ocean patterns
Ozone depletion	Increased ultraviolet radiation
Resource depletion	Decreased resources for future generations
Regional impacts
Photochemical smog	“Smog,” decreased visibility, eye irritation, respiratory tract and lung irritation, and vegetation damage
Acidification	Building corrosion, water body acidification, vegetation effects, and soil effects
Local impacts
Human health	Increased morbidity and mortality
Terrestrial toxicity	Decreased production and biodiversity and decreased wildlife for hunting or viewing
Aquatic toxicity	Decreased aquatic plant and insect production and biodiversity and decreased commercial or recreational fishing
Land use	Loss of terrestrial habitat for wildlife and decreased landfill space

Impact indicators are typically characterized using the following equation:

For example, all GHGs can be expressed in terms of carbon dioxide equivalents by multiplying the relevant LCI results by a CO₂ characterization factor and then combining the resulting impact indicators to provide an overall indicator of global warming potential.

The Intergovernmental Panel on Climate Change provides conversion factors for a number of industrial pollutants. In the following example, the global warming impacts of different amounts of chloroform (CHCl₃) and methane (CH₄) are characterized. The value of the conversion factor, called GWP, for global warming potential, is 9 for CHCl₃ and 21 for CH₄.

Thus, we write:

Use of the conversion factors shows that 10 lb of methane has a larger impact on global warming than 20 lb of chloroform.

The key to impact characterization is using the appropriate characterization factor. For some impact categories, such as global warming and ozone depletion, there is a consensus on acceptable characterization factors. For other impact categories, such as resource depletion, a consensus is still being developed. Table 6.1 includes descriptions of possible characterization factors for some of the commonly used life cycle impact categories. A properly referenced LCIA will document the source of each characterization factor to ensure that they are relevant to the goal and scope of the study. For example, many characterization factors based on studies conducted in Europe cannot be applied to American data unless it can be verified that they are appropriate to conditions in the United States.

6.2.6.1 Key Steps to Interpreting the Results of the LCA

The guidance provided thus far summarizes the information on life cycle interpretation in the ISO draft standard Environmental Management – Life Cycle Assessment – Life Cycle Interpretation (ISO 1998b). The ISO draft standard covers the following steps in conducting a life cycle interpretation:

1. Identify significant issues
2. Evaluate the completeness, sensitivity, and consistency of the data
3. Draw conclusions and recommendations

Figure 6.5 illustrates the steps of the life cycle interpretation process in relation to the other phases of the LCA process.

Step 2: Evaluate the Completeness, Sensitivity, and Consistency of the Data 

The evaluation step of the interpretation phase establishes the confidence in and reliability of the results of the LCA. This is accomplished by performing completeness, sensitivity, and consistency checks to ensure that products/processes are fairly compared.

Completeness check – The completeness check ensures that all relevant information and data needed for the interpretation are available and complete. A checklist should be developed to indicate each significant area represented in the results. Using the established checklist, it is possible to verify that the data comprising each area of the results are consistent with the system boundaries (e.g. all life cycle stages are included) and that the data is representative of the specified area (e.g. accounting for 90% of all raw materials and environmental releases).

The result of this effort will be a checklist indicating that the results for each product/process are complete and reflective of the stated goals and scope of the LCA study. If deficiencies are noted, an attempt must be made to remedy them. If this is not possible because data are not available, areas inadequately characterized because of insufficient data must be highlighted in the final results and their impact on the comparison estimated either quantitatively (percent uncertainty) or qualitatively (alternative A's reported result may be higher because “X” is not included in its assessment).

Sensitivity check – The objective of the sensitivity check is to evaluate the reliability of the results by determining whether the uncertainty in the significant issues identified in step 1 affect the decision‐maker's ability to confidently draw comparative conclusions. Three common techniques for data quality analysis can be used in performing sensitivity checks.

1. Gravity analysis – Identifies the data that has the greatest contribution on the impact indicator results.
2. Uncertainty analysis – Describes the variability of the LCIA data to determine the significance of the impact indicator results.
3. Sensitivity analysis – Measures the extent that changes in the LCI results and characterization models affect the impact indicator results.

Additional guidance on how to conduct a gravity, uncertainty, or sensitivity analysis can be found in the EPA document entitled “Guidelines for Assessing the Quality of Life Cycle Inventory Analysis” (USEPA 1995). If one of these analyses has been conducted as part of the LCI and LCIA phases, these results can be used. Then the sensitivity check will serve to verify that the goals for data quality and accuracy defined early in have been met. If deficiencies exist, additional efforts are required to improve the accuracy of the LCI data collected and/or impact models used in the LCIA. If better data or impact models cannot be obtained, the deficiencies for each relevant significant issue must be reported and its impact on the comparison estimated either quantitatively or qualitatively, as with the completeness check.

Consistency check – The consistency check determines whether the assumptions, methods and data used throughout the LCA process are consistent with the goal and scope of the study, and for each product/process evaluated. Verifying and documenting that the study was completed as intended at the conclusion increases confidence in the final results. A formal checklist should be developed to communicate the results of the consistency check. Table 6.2 lists seven categories and provides examples of inconsistencies that can creep into the data. The goal and scope of the LCA determines which categories should be used.

If, after completion of steps 1 and 2, it is determined that the results of the impact assessment and the underlying inventory data are complete, comparable, and acceptable as bases for drawing conclusions and making recommendations then stop! If any inconsistency is detected, document the role it played in the overall consistency evaluation. Although some inconsistency may be acceptable, depending upon the goal and scope of the LCA, the presence of inconsistencies usually means that it is necessary to repeat steps 1 and 2 until the results are able to support the original goals for performing the LCA.

6.2.6.2 Reporting the Results

When the LCA has been completed, the materials must be assembled into a comprehensive report documenting the study in a clear and organized manner. This will help communicate the results of the assessment fairly, completely, and accurately to others interested in the results. The report presents the results, data, methods, assumptions and limitations in sufficient detail to allow the reader to comprehend the complexities and trade‐offs inherent in the LCA study.

If the results will be communicated to parties who were not involved in the LCA study (e.g. stakeholders), the report will serve as a reference document, and it can help prevent any misrepresentation of the results.

6.2.6.3 Conclusion

Adding LCA to the decision‐making process provides a level of understanding of human health and environmental impacts that traditionally has not been available to those responsible for selecting a product or process. This valuable information provides a way to account for the full impacts of decisions, especially those made off‐site, that are directly influenced by the selection of a product or process. As emphasized earlier, LCA is a tool to better inform decision‐makers and other decision criteria such as cost and performance must be weighed to reach a well‐balanced decision.

As we have seen, LCA and LCI can be valuable tools in environmental analysis. To make sense of the many large data sets collected in any such study is clearly a task too complex to be undertaken without the assistance of computers. Section 6.3 describes some of the software tools that have been developed for this purpose.

6.4.3.1 Mechanical Pulping

Mechanical pulping, as the name implies, relies on mechanical energy to convert wood to pulp. Current mechanical pulp manufacturing processes constitute several high‐energy grinding and refining systems, including refiner mechanical pulping (RMP) process, the thermomechanical pulping (TMP) process, the chemimechanical pulping process, and the chemithermomechanical pulping process (Someshwar and Pinkerton 1992; Smook 1992). Mechanical pulping accounts for about 25% of the wood pulp production in the world today. Mechanical pulping, with its high yield, is viewed as a way to extend these resources. However, mechanical pulping is electrical energy–intensive and yields paper with lower strength compared with that produced by chemical pulping which are used in the pulping process. Figure 6.6 shows the basic flow diagram for TMP (Smook 1992). TMP is a modification of RMP; it involves steaming the raw material for a short period of time before and during several refining stages. The steaming serves to soften the chips, with the result that the pulp produced has a greater percentage of long fibers and less shives than RPM. Most often, heating and refining are both done under pressure (TMP), but some systems refine under atmospheric pressure. Chemicals are sometime added at the various stages of refining. In the TMP process, the chips are usually steamed under a pressure of 20–40 psi for 2–4 minutes before refining.

Table 6.4 presents an average quantity (in %) of make‐up chemicals that are typically needed for mechanical/thermomechanical pulping (USDA, personal communication). Table 6.5 presents the energy consumption in various operations of TMP and papermaking process (US Pulp and Paper Industry 1988). Table 6.6 shows a range of effluent loads and other environmental impacts of the TMP process (Hynninen 1998; Sundholm 1999). The data from Tables 6.4–6.6 are used to quantify the environmental and human health impacts that will be discussed later in Section 6.4.3.7.

A study by the National Council of the Paper Industry for Air and Stream Improvement (NCASI 1983) on estimating volatile organic compounds (VOCs) are likely to be released along with the steam during the cooking and refining process. The range of VOC emissions can vary between 1.0 and 7.0 lb of carbon per ton of pulp, depending on the wood species. No emissions other than VOC are expected from the mechanical pulping process.

6.4.3.2 Chemical Pulping Processes

The following discussion centers on chemical pulping processes, which supply more than two‐thirds of world's wood pulp. The most widely used chemical pulping process in papermaking is the kraft, or sulfate, process. Other chemical‐pulping processes (mainly using acid sulfite and soda) are sometimes combined with various chemical‐recovery subprocesses. First used over a century ago, these processes now result in recovery of 90% of the inorganics, which are used in the pulping process. Nearly 100% of the dissolved organics are converted to energy. The typical kraft process involves turning logs into wood chips, which are then pulped (Figure 6.7) (Someshwar and Pinkerton 1992). The major steps in chemical pulp and papermaking processes include wood yard operations, pulping, deknotting, washing and screening, chemical recovery, pulp bleaching, pulp drying, and paper making (Smook 1992; Someshwar and Pinkerton 1992). Various gaseous, particulate, liquid, and solid waste streams are produced within a kraft mill (BOD, COD, TSS, NO_x, SO₂, CO, CO₂, PM, TRS, odor, and ash). A wealth of information in the existing literature addresses various treatment methods and pollution prevention and process optimization techniques for these waste streams. Some of the pollution‐prevention advancements and a summary of reduction in total environmental discharges (wastewater, air emissions, and solid and hazardous wastes) from kraft and sulfite mills across the United States over the last 20–25 years are discussed in the literature cited in the reference section (Das 2005; Das and Jain 2001).

Table 6.4 presents an average quantity (in percentage) of make‐up chemicals that are typically needed for chemical pulping. Chemical pulping uses 63% more chemical than mechanical pulping, and 53% more than biopulping. Table 6.5 presents the energy consumption in various operations of chemical pulping and papermaking processes. Table 6.6 shows a range of effluent loads and other environmental impacts of chemical pulping and papermaking processes. The data from Tables 6.4–6.6 are used to quantify the environmental and human health impacts that will be discussed later in Section 6.4.3.7.

6.4.3.3 Biopulping: A Review of a Pilot Project

Biopulping is the treatment of wood chips and other lignocellulosic materials with lignin‐degrading fungi before pulping. Ten years of industry‐sponsored research has demonstrated the technical feasibility of the technology for mechanical pulping at a laboratory scale. Two 50‐T outdoor chip pile trials conducted at the USDA Forest Service, Forest Products Laboratory (FPL) in Madison, WI, have established the engineering and economical feasibility of the technology. After refining the control and the fungus‐treated chips through a thermomechanical pulp (TMP) mill, the resulting pulps were made into papers on the pilot‐scale paper machine at FPL. In addition to the 30% savings in electrical energy consumption during refining, improvements in the strength of the resulting paper were seen as a result of fungal pretreatment. Because of the stronger paper, we were able to substitute at least 5% kraft pulp in a blend of mechanical and kraft pulps. This recent work has clearly demonstrated that economic benefits can be achieved with biopulping technology through both the energy savings and substitution of the stronger biopulped TMP for more expensive kraft, while maintaining the paper quality.

6.4.3.4 Introduction to Biopulping

Biopulping, which uses natural wood decay organisms, has the potential to overcome these problems. Fungi alter the lignin in the wood cell walls, which has the effect of “softening” the chips. This substantially reduces the electrical energy needed for mechanical pulping and leads to improvements in the paper strength properties. The fungal pretreatment is a natural process; therefore, no adverse environmental consequences are foreseen (Kirk et al. 1993, 1994; Sykes 1994). Based on the results of previous work and discussions with mill personnel, we envision a fungal treatment system that fits into existing mill operations with minimal disturbance. Figure 6.8 is a conceptual overview of the biotreatment process in relation to existing wood yard operations. Wood is harvested and transported to the mill site for debarking, chipping, and screening. Chips are decontaminated by steaming, maintaining a high temperature for a sufficient time to decontaminate the wood chip surfaces, and then cooled so that the fungus can be applied. The chips are then placed in piles that can be ventilated to maintain the proper temperature, humidity, and moisture content for fungal growth and subsequent biopulping. The retention time in the pile is 1–4 weeks. Recent efforts have focused on bringing the successful laboratory‐scale procedures up to the industrial level. Our laboratory process treats approximately 1.5 kg of chips (dry weight basis) at one time. In our scale‐up experiments that took biopulping from this lab scale to larger scales, certain implementations of each step were chosen. The chosen implementation allowed us to reach our two goals for this phase of the project: (i) to demonstrate that chips can be decontaminated and inoculated in a continuous process rather than as a batch process and (ii) to demonstrate that the process scaled as expected from an engineering standpoint. The following discusses the results obtained during the scale‐up trials in relation to industrial processes.

In reactor scale‐up studies, Kirk and others (1993) investigated two types of reactor systems: tubular reactors and chip piles. The tubular reactors have an advantage in obtaining the necessary engineering and kinetic data for scaling‐up the process. The one‐dimensional nature of the system is easy to analyze and model. The reactor also allows for well‐controlled air flow in the system with air flow patterns that are well known. Heat loss from the system is easily controlled with exterior insulation, thus achieving conditions that would be experienced in the center of large chip piles. Details on the configurations of these reactors and the chip piles have been published.

6.4.3.5 Large‐Scale Implementation of Biopulping

On a large scale, decontamination and inoculation must be done on a continuous basis and not batchwise as in a laboratory trial (Figure 6.9). To achieve this, the FPL investigators built a treatment system based on two screw conveyers that transport the chips and act as treatment chambers. Figure 6.9 is an overview of the continuous process equipment used in 5‐ and 50‐T trials. Steam is injected into the first screw conveyer, which heats and decontaminates the chip surface. A surge bin, located between the two conveyers, acts as a buffer. From the bottom of the surge bin, a second screw conveyer removes the chips, which are subsequently cooled with filtered air. In the second half of the second screw conveyer, the inoculum suspension containing fungus, unsterilized corn steep liquor, and water is applied and mixed with the chips through the tumbling action. From the screw conveyer, the chips fall into the pile or reactor for a two‐week incubation. In the first scale‐up trial, 5 T of spruce wood chips were inoculated and incubated at a throughput of approximately 0.5 T/h.

In successful outdoor trials with the biopulping fungus Ceriporiopsis subvermispora, nearly 50 T of spruce was treated at a throughput of about 2 T/h (dry weight basis) continuously for over 20 hours. During the two weeks, the chip pile was maintained within the temperature growth range for the fungus, despite the outdoor exposure to ambient conditions.

6.4.3.6 Economics and Environmental Benefits of the Biopulping Process

The economic benefits of biopulping, evaluated based on the process studies and engineering data, result from several effects. Energy reduction at the refiner was used as the primary criterion for the effectiveness of biopulping. For a two‐week process, the savings should be a minimum of 25% under the worst‐case conditions of wood species and minimal process control, whereas up to nearly 40% can be achieved under some circumstances. Additionally, mills that are currently throughput‐limited as a result of refiner capacity may achieve total capacity increases as a result of biopulping. The improved strength of the biomechanical pulps would allow the required strength of the blend to be achieved with a lower percentage of the kraft pulp in those cases where the pulps are blended. Finally, only benign materials are used, and no additional waste streams are generated. An economic analysis of the process yielded very favorable results. The analysis of a 200 T/day TMP mill are summarized here. Under different scenarios and assumptions for utility costs, equipment needs, and operating costs, the net savings can range from $10 to more than $26/T of pulp produced, with an estimated capital investment of $2.5 million. Mills that are refiner‐limited can experience throughput increases of over 30% from the reduction in refining energy by running the refiners to a constant total power load. Even a modest throughput increase of 10%, coupled with the energy savings of 30%, results in a payback in less than 1 year. This is equivalent to a savings of $34/T at a 15% rate of return on capital. Furthermore, many mills blend mechanical pulps and kraft pulps to achieve the optical and strength properties desired. Additional benefits of over $10/T can be realized when the anticipated stronger biopulped TMP is partially substituted for kraft at a 5% rate. The results of a 600 T/day analysis have also been published (Scott et al. 1998a, b). This preliminary analysis is subject to appropriate qualifications. The capital costs are subject to some variability, in particular the costs associated with integrating the new facility into an existing site. The additional advantages of biopulping, including the environmental benefits and pitch reduction, have not been quantified in this report. Finally, much of this analysis is site‐specific, depending on the operating conditions at the particular mill considering incorporating biopulping into its operations.

Although the environmental benefits of biopulping for mechanical pulping have not been quantitatively determined in an economic sense, we can discuss the benefits qualitatively. Sykes (1994) determined that there were no environmental problems with the effluent from a biopulping process. In terms of GHGs, there is a net benefit to the biopulping process. The process itself is oxidative, producing carbon dioxide as the fungus metabolizes the various wood components. However, based on the typical generation efficiency of electrical energy, the reduced energy requirements (~30%) significantly lower the overall generation of carbon dioxide when the entire system is analyzed. Finally, in terms of biological issues with using fungi in the wood yard, there are benefits to using the biopulping fungus. The biopulping fungus outcompetes the normal cohort of fungi and other organisms that will grow in an uncontrolled wood chip pile. Some of these other organisms, such as Aspergillis, can be detrimental to human health. The biopulping fungi are naturally occurring in the forests of the world and have no known health effects on humans. More information on benefit–cost analysis and economic input–output model (EIO‐LCA) for chemical process and manufacturing industries are available in Appendix F.

6.4.3.7 Quantitative Analyses Using the USEPA Model: TRACI

The tool for the reduction and assessment of chemical and other environmental impacts (TRACI) is described along with its history, the research and methodologies it incorporates, and the insights it provides within individual impact categories.

TRACI, a stand‐alone computer program developed by the US Environmental Protection Agency (USEPA), facilitates the characterization of environmental stressors that have potential effects, including ozone depletion, global warming, acidification, eutrophication, tropospheric ozone (smog) formation, ecotoxicity, human health criteria–related effects, human health cancer effects, human health noncancer effects, fossil fuel depletion, and land‐use effects. TRACI was originally designed for use with LCA, but it is expected to find wider application in the future (Bare 2003; UNEP‐SETAC 2000).

The categories of odor, noise, radiation, waste heat, and accidents are outside of the USEPA's purview and are usually not included within case studies in the United States for various reasons, including, perhaps, because the perceived threat from these categories is often considered minimal, local, or difficult to predict. The resource depletion categories are recognized as being of significance in the United States, especially for fossil fuel, land, and water use.

For the development of TRACI, each of the above impact categories was considered and its current state of development and perceived societal value were assessed. The traditional pollution categories of ozone depletion, global warming, human toxicology, ecotoxicology, smog formation, acidification, and eutrophication were included within TRACI because various programs and regulations within the USEPA recognize the value of minimizing effects from these categories. The category of human health was further subdivided into cancer, noncancer, and criteria pollutants (with an initial focus on particulates) to better reflect the focus of renewable fuels, and purchased electricity used. Emissions from wastewater treatment plant and other biological processes are not included. In Washington State, 60% energy used in the pulp and paper mills are generated from renewable sources (almost all from hydro) and 40% from purchased electricity generated from natural gas, coal, and oil. The CO₂ emissions from renewable energy sources assumed zero. Typical values of 70% combustion efficiency and 60% electricity transmission efficiency were assumed. For an overall efficiency factor of 42% in calculating CO₂ emissions from nonrenewable sources, power‐generating plants to pulp mills were used.

USEPA regulations and to allow methodology development consistent with US regulations, handbooks, and guidelines (USEPA 1989b, c). Smog‐formation effects were kept independent and not further aggregated with other human health impacts because environmental effects related to smog formation would have become masked and/or lost in the process of aggregation. Criteria pollutants were maintained as a separate human health impact category, allowing a modeling approach that can take advantage of the extensive epidemiological data associated with these well‐studied impacts.

6.4.3.8 Model Input and Output

It is important to note that data quality is fundamental to LCA and its interpretation. Some of the data quality issues such as reliability and consistency can be overcome by using standardized database, which are starting to emerge after years of data compilation and their incorporation into publicly and commercially available databases. Table 6.7 presents some chemical release data that are used in TRACI. Table 6.8 presents the overall environmental and human health impacts, including global warming, smog formation, human health criteria pollutants, acidification, eutrophication, and ecotoxicity. The results indicate that biopulping has a lower environmental impact than that of the TMP process because of its reduced energy consumption. Biopulping also has a lower impact than that of chemical pulping in all the measures except global warming. One should note that the total CO₂ production per ton was the highest for chemical pulping, but because kraft recovery furnaces are fired with lignin derived from the wood, a renewable resource, this contribution to GHGs was assumed to be net zero. Figure 6.10 shows global warming characterization results predicted by TRACI.

These results likely indicate that biopulping is the preferred process when the electrical supply is dominated by renewable energy sources. In cases where the fuel mix for producing electricity is largely fossil fuels, chemical pulping provides a process powered by renewable energy but at a cost of higher hazardous waste emissions.

In conclusion, this streamlined LCA clearly indicates that biopulping in the papermaking process has several advantages over chemical and mechanical processes as follows:

• Reduced electrical energy consumption (at least 30%) over mechanical pulping
• Potential 30% increase in mill throughput for mechanical pulping
• Improve paper strength properties
• The TRACI model indicates a significant reduction in environmental and human health impacts
• The biopulping process proves to be more sustainable in terms of not only economic advantage but also environmental and human health benefits

6.5.2.1 Limitations

Although chlorine disinfection is a largely reliable and effective process, it has certain limitations. For examples, chlorine reacts with certain chemicals in the wastewater, leaving only the residual for disinfection. Wastewater components that readily combine with chlorine include reduced iron and sulfur compounds, ammoniated‐nitrogen, organic nitrogen, tannins, uric and humic acid, cyanides, phenols, and unsaturated organics. Cysts of Entamoeba histolytica, Giardia lamblia, and Mycobacterium tuberculosis, some viruses, and eggs of parasitic worms show resistance to chlorine. Consistent disinfection in effluents containing organic nitrogen may pose problems, even when a measured free chlorine residual is present.

6.5.2.2 Human Health and Environmental Impact

Chlorine is toxic to aquatic, estuarine, and marine organisms. An additional hazard is the carcinogenic potential of chloro‐organic compounds. Chlorine gas is potentially toxic when inhaled, and chlorine transport poses a risk (see Section 5.10.5and Tables 5.8 and 5.9). Special handling is required and emergency response plans are required under right‐to‐know regulations for on‐site storage of gaseous chlorine. Chlorine gas and the hypochlorites are also highly corrosive. Chlorine gas concentrations of 15–20 ppm for 30–60 minutes are dangerous; higher concentrations for very brief periods can be fatal. Chlorination can result in the formation of carcinogenic chloro‐organics.

The EPA has established toxicity criteria for TRC in receiving waters. In freshwaters the acute level is 19 mg/l (1‐hour average) and the chronic level is 11 mg/l (4‐day average). The saltwater acute and chronic criteria are 13 mg/l (1‐hour average) and 7.5 mg/l (4‐day average), respectively. Due to the toxicity of chlorine residuals at such low concentrations and the high limit of analytical detection (50–100 mg/l), chlorine induced toxicity in the receiving stream is difficult to control.

6.5.3.1 Limitations

Chlorination/dechlorination is more complex to operate and maintain than chlorination alone. Major difficulties are the inability to measure residual SO₂ and problems in the continuous measurement of a zero or low chlorine residual. Many halogenated organics are also rapidly formed upon chlorine addition and are unaffected by application of SO₂. Heltz and Nweke who examined many plants, reported that the amount of residual chlorine in dechlorinated effluents considerably exceeded EPA criteria for receiving waters (Heltz and Nweke 1995).

6.5.3.2 Environmental Impact

Sulfuric acid and hydrochloric acid are products of SO₂ dechlorination in small amounts but are generally neutralized in the wastewater. Based on laboratory experiments, residuals of sulfite dechlorination are at least three orders of magnitude less toxic than chlorine or ozone.

No cases of sulfur compounds affecting dissolved oxygen consumption or pH change in receiving waters or in dechlorinated effluents have been reported. In pilot studies, no significant oxygen depletion occurred until sulfur dioxide overdoses exceeded 50 mg/l. It is not uncommon, however, to find plants with post aeration after dechlorination to assure that dissolved oxygen requirements are met. Dechlorination with SO₂ would significantly reduce toxicity due to chlorination disinfection process.

Sulfur dioxide, which is used for dechlorination purposes, is stored on site in small quantities. The maximum anticipated storage inventory is quite small, and there need be no concerns about threats to the health of workers or of residents of nearby populated areas.

6.5.3.3 Effects of Chlorine on Aquatic Life

The potential environmental and chemical effects of chlorine toxicity on 14 aquatic species are summarized in Table 6.9. Rainbow trout was the most sensitive of the species tested, followed by the golden shiner and three‐spined stickleback. A calculated chlorine residual of 0.03 mg/l, based on dilution of a measured concentration of 2.0 mg/l, reduced plankton photosynthesis by more than 20% of the value obtained with a dilution of effluent having no chlorine residual. Dechlorination with sodium bisulfite also eliminated chlorine‐related toxicity. Dechlorination with sulfur dioxide also greatly reduced the acute and chronic toxicity to fish and invertebrate species.

In considering these data, it should be borne in mind that the toxicity of chlorine wastes in rivers depends not on the amount of chlorine added but on the concentration of chlorine remaining in solution (Merkens 1958). The toxicity of this residual chlorine will depend on its composition (i.e. the relative proportions of free chlorine and chloramines). Arriving at this ratio, which in turn depends on at least five other variables, is a complex undertaking. Although free chlorine is more toxic than chloramines, research that has stood since the 1950s suggests that the difference between the toxicity to fish of free chlorine and chloramine is not very great (Dondoroff and Katz 1950; Merkens 1958; Hart 1973; Kozloff 1974).

6.5.4.1 UV Transmission or Absorbance

UV light's ability to penetrate wastewater is measured in a spectrophotometer at the same wavelength (254 nm) that is produced by germicidal lamps. This measurement is called the percent transmission or absorbance and it is a function of all the factors that absorb or reflect UV light. As the percent transmission gets lower (higher absorbance) the ability of the UV light to penetrate the wastewater and reach the target organisms decreases.

It cannot be estimated by simply looking at a sample of wastewater with the naked eye. The range of effective transmittances (T) will vary depending on the secondary treatment systems. In general, suspended growth‐treatment processes produce effluent with T varying from 60 to 65%. Fixed film processes range from 50 to 55% T and lagoons 35–40% T. Industries that influence UV transmittance include textile, printing, pulp and paper, food processing, meat and poultry processing, photo developing, and chemical manufacturing. A discussion on other major parameters affecting the UV disinfection efficiency of wastewater is given by Das (2004).

6.5.4.2 Disinfection Standards

The level of disinfection required to obtain an EPA National Pollutant Discharge Elimination System (NPDES) permit is commonly less than 200 fecal coliform unit per 100 ml as a 30‐day geometric mean. In general, a UV dose of 20–30 mW·s/cm² is required to achieve this level of disinfection in secondary‐treated wastewater with a 65% transmittance and total suspended solids (TSS) below 20 mg/l. The UV dose requirement to meet specific limits depends on the nature of the particle with respect to numbers, size, and composition. Therefore, UV dose requirements will vary (Table 6.11). A more stringent limit (<2.2 total coliform units per 100 ml) is required for water reuse in California and Hawaii. In such cases, filtered effluents with TSS of 2 mg/l or less and 65% transmittance may require UV dose as high as 120 mW·s/cm² to achieve this level of disinfection. The concentrations of solids, bacteria in the particles, and the particle size distribution are the main limiting factors in the design of systems that must meet stringent disinfection limits. It appears that the UV dose required to achieve the traditional coliform limits will achieve better virus inactivation than the comparable chlorine dose. Figure 6.11, which illustrates the relative doses of UV and chlorine required to inactivate selected organisms compared to fecal coliform indicator, also shows that the chlorine doses required to inactivate most organisms are much higher than comparable UV doses that would achieve the same level of disinfection (Das 2002, 2004; Trojan Technologies Inc. 2010).

6.5.4.3 Operation, Maintenance, and Worker Safety

The proper operation and maintenance (O&M) of a UV disinfection system ensures that sufficient UV radiation is transmitted to the organisms to render them sterile. All surfaces between the UV radiation and the target organisms must be cleaned, and the ballasts, lamps, and reactor must be functioning at peak efficiency. Inadequate cleaning is one of the most common causes of a UV system's ineffectiveness. In all cases, the quartz sleeves or Teflon tubes need to be cleaned regularly by mechanical wipers, ultrasound equipment, or chemicals. The cleaning frequency is very site‐specific: some systems needing to be cleaned more often than others.

Chemical cleaning is most commonly done with citric acid. Other cleaning agents include mild vinegar solutions and sodium hydrosulfite. A combination of cleaning agents should be tested to find the agent most suitable for the wastewater characteristics without producing harmful or toxic by‐products.

UV is generated on‐site and poses no significant safety concerns to surrounding communities. Worker safety requirements are directed to protection from exposure (primarily of the eyes to skin) from UV light, as well as to strict monitoring of electrical hazards and safe handling and disposal of expended lamps, quartz, and ballasts (US EPA 1996; US EPA 1998).

6.5.4.4 Costs

Specifically, consideration must be given to how environmental matters affect our economic thinking and, conversely, how economic decisions affect the environment. Cost considerations are an integral part of the decision‐making process at the stage of identifying potential improvements to a process, a product, or an activity.

Estimated capital costs for 1, 10, and 50 million gal per day (mgd) (1 mgd = 3785 m³/day) plants are $150 000, $700 000, and $3 000 000 respectively. O&M costs are estimated to be 2.6, 1.4, and 0.8 US cents/1000 gal (or 3.785 m³), respectively, at an average dose of 6 mg/l. These increase to 4.0, 3.0, and 1.5 US cents/1000 gal at a dose of 10 mg/l.

Chemical costs of chlorine gas and hypochlorite vary considerably depending upon the locality, demand, and availability. Current prices for Cl₂ are varying in the range: $0.0425–0.06/lb (or 454 g) for 90 T (or ~90 000 kg) tank cars; $0.10/lb for 55 T (or ~55 000 kg) rail cars; $0.25–0.275/lb for 1 T (or ~1 000 kg) cylinders; and $0.50–0.55/lb for 150 lb (or ~68 kg) cylinders. Liquid sodium hypochlorite (12.5%) prices quoted were $0.50–1.75/gal.

A brief cost analysis for both chlorination/dechlorination and UV disinfection processes treating a wastewater treatment facility is presented in Figure 6.12, capable of processing 18 mgd. These costs can vary considerably depending on the locality, demands, and supply. The annualized cost values for alternatives were based on the design average plant flow of 18 mgd over a 20‐year period and 5.5% interest rate (F. Soroushian, personal communication; Temmer et al. 2000).

Figure 6.12 indicates that, although the construction cost for chlorination system is considerably less than the UV, annualized costs are about the same for both systems. There is no more real economic benefit for pursuing chlorination disinfection system.

6.5.4.5 Environmental Impacts of Energy Sources and Implications of Renewables

Like chlorination, UV disinfection systems demand energy supplies. One of the most important factors that can contribute to achieving sustainable development is the requirement for a supply of energy resources that is itself fully sustainable. Effective and efficient utilization of energy resources calls for such resources to be readily available at reasonable cost utilized for all required tasks without causing negative societal impact. Clearly, there is an intimate connection between renewable energy sources and sustainable development (Das 2002).

Table 6.12 gives a brief summary of cost of generating powers, air quality, and environmental impacts of renewable and other forms of energy sources. It illustrates that renewable energy sources can make a significant contribution to reducing greenhouse and acid gas emissions. Renewable sources have their own environmental impacts but these are often small, site‐specific, and local in nature.

6.6.1.1 System Energy Balance

The total energy consumed by each system includes the fuel energy consumed plus the energy contained in raw and intermediate materials that are consumed by the systems. Examples of the first type of energy use are the fuel spent in transportation and fossil fuels consumed by the fossil‐based power plants. The second type of energy is the sum of the energy that would be released during combustion of the material (if it is a fuel) and the total energy that is consumed in delivering the material to its point of use. Examples of this type of energy consumption are the use of natural gas in the manufacture of fertilizers and the use of limestone in flue‐gas desulfurization (FGD). The combustion energy calculation is applied where nonrenewable fuels are used, reflecting the fact that the fuel has a potential energy that is being consumed by the system. The combustion energy of renewable resources, those replenished at a rate equal to or greater than the rate of consumption, is not subtracted from the net energy of the system. This is because, on a life cycle basis, the resource is not being consumed. To determine the net energy balance of each system, the energy used in each process block is subtracted from the energy produced by the power plant. The total system energy consumption by each system is shown in Table 6.13.

To examine the process operations that consume the largest quantities of energy within each system, two energy measurement parameters were defined. First, the energy delivered to the grid divided by the total fossil‐derived energy consumed by each system was calculated. This measure, known as the net energy ratio, is useful for assessing how much energy is generated for each unit of fossil fuel consumed. The other measure, the external energy ratio, is defined to be the energy delivered to the grid divided by the total non‐feedstock energy to the power plant. That is, the energy contained in the coal and natural gas used at the fossil‐based power plants is excluded. The external energy ratio assesses how much energy is generated for each unit of upstream energy consumed.

Because the energy in the biomass is considered to be both generated and consumed within the boundaries of the system, the net energy ratio and external energy ratio will be the same for the biomass‐only cases (biomass‐fired IGCC and direct‐fired biomass). In calculating the external energy ratio, we are essentially treating the coal and natural gas fed to the fossil power plants as renewable fuels, so that upstream energy consumption can be compared. The energy results for each case studied are shown in Figure 6.13.

As expected, the biomass‐only plants consume less energy overall, since the consumption of nonrenewable coal and natural gas at the fossil plants results in net energy balances of less than one. The direct‐fired biomass residue case delivers the most amount of electricity per unit of energy consumed. This is because the energy used to provide a usable residue biomass to the plant is fairly low. Despite its higher plant efficiency, the biomass IGCC plant has a lower net energy balance than the direct‐fired plant because of the energy required to grow the biomass as a dedicated crop. Residue resource limitations, however, may necessitate the use of energy crops in the future. Cofiring biomass with coal slightly increases the energy ratios over those for the coal‐only case, even though the plant efficiency was derated by 0.9 percentage points.

In calculating the external energy ratios, the feedstocks to the power plants were excluded, essentially treating all feedstocks as renewable. Because of the perception that biomass fuels are of lower quality than fossil fuels, the external energy ratios for the fossil‐based systems were expected to be substantially higher than those of the biomass‐based systems. The opposite is true, however, due to the large amount of energy that is consumed in upstream operations in the fossil‐based systems. The total non‐feedstock energy consumed by the systems is shown in Table 6.14. In the case of coal, 35% of this energy is consumed in operations relating to flue‐gas cleanup, including limestone procurement. Mining the coal consumes 25% of this energy, while transporting the coal is responsible for 32%. Greater than 97% of the upstream energy consumption related to the natural gas IGCC system is due to natural gas extraction and pipeline transport steps, including fugitive losses. Although upstream processes in the biomass systems also consume energy, shorter transportation distances and the fact that FGD is not required reduce the total energy burden.

6.6.1.2 Global Warming Potential

Figure 6.14 shows the net emissions of the three GHGs quantified for these studies: carbon dioxide (CO₂), methane (CH₄), and nitrous oxide (N₂O). The biomass IGCC system has a much lower GWP than the fossil systems because of the absorption of CO₂ during the biomass growth cycle. Sensitivity analyses demonstrated that even moderate amounts of soil carbon sequestration (1900 kg/ha/7‐year rotation) would result in the biomass IGCC system having a zero‐net GHG balance. Sequestration amounts greater than this would result in a negative release of GHGs, and a system that removes carbon from the atmosphere overall.

The base case presented here assumes no net change in soil carbon, since actual gains and losses will be very site specific. The direct‐fired biomass system in which the residue is used at the power plants has a highly negative rate of GHG emissions because the generation of methane associated with biomass decomposition is avoided. Based on current disposal practices, it was assumed that 46% of the residue biomass used in the direct‐fired and cofiring cases would have been sent to a landfill and that the remainder would end up as mulch and other low‐value products. Decomposition studies reported in the literature were used to determine that if the biomass residue had not been used at the power plant, approximately 9% of the carbon would have ended up as CH₄ and 61% as CO₂. The remaining carbon is resistant to decomposition in the landfill. Had all the residue biomass been decomposed aerobically, the CO₂ produced would have been 1.85 kg/kg biomass. If the biomass residue was not used at the power plant, the decomposition pathways described above would have resulted in total GHG emissions of 2.48 kg CO₂‐equiv/kg biomass (1.117 kg CO₂ + 0.065 kg CH₄). The net difference is the reason for the negative GHG emissions associated with the direct‐fired system.

The NGCC plant has the lowest GWP of all fossil systems because of its higher efficiency, despite natural gas losses that increase net CH₄ emissions. Natural gas losses during extraction and delivery were assumed to be 1.4% of the gross amount extracted. Because of the potency of methane as a GHG, nearly one‐quarter of the total GWP of this system is due to these losses. Cofiring biomass with coal at 15% by heat input reduces the GWP of the average coal‐fired power plant by 18%. The reduction in GHGs is greater than the rate at which biomass is cofired because of the avoidance of methane emissions associated with decomposition that would have occurred had the biomass not been used at the power plant. Biomass disposal and decomposition emissions for this scenario are the same as those used in the direct‐fired case.

6.6.1.3 Air Emissions

Figure 6.15 charts the following emissions: particulates, oxides of sulfur (SO_x), oxides of nitrogen (NO_x), CH₄, CO₂, and non‐methane hydrocarbons (NMHCs). Methane emissions are high for the natural gas case due to natural gas losses during extraction and delivery. The direct‐fired biomass and coal/biomass cofiring cases have negative methane emissions due to avoided decomposition processes (landfilling and mulching). CO and NMHCs are higher for the biomass case because of upstream diesel combustion during biomass growth and preparation. Cofiring reduces the coal system air emissions by approximately the rate of cofiring, with the exception of particulates, which are generated during biomass chipping and handling.

6.6.1.4 Resource Consumption

Figure 6.16 shows the total amount of nonrenewable resources consumed by the systems investigated. Limestone is used in significant quantities by the coal‐fired power plants for FGD. The natural gas IGCC plant consumes almost negligible quantities of resources, with the exception of the feedstock itself, including that lost during extraction and delivery.

6.7.1.1 Reasonably Available Control Technology

Reasonably available control technology (RACT) is a pollution control standard created by the EPA and is used to determine what air pollution control technology will be used to control a specific pollutant to a specified limit. RACT applies to existing sources in areas that are not meeting national ambient air quality standards on controlled air pollutants and is required on all sources that meet these criteria. The RACT standard is less stringent than the BACT.

6.7.1.2 Lowest Achievable Emissions Rate

The lowest achievable emissions rate (LAER) is used by the EPA to determine if emissions from a new or modified major stationary source are acceptable under SIP guidelines. LAER standards are required when a new stationary source is located in a non‐attainment air‐quality region (see Figure 4.14). It is the most stringent air pollution standard above the best available control technology and reasonably available control technology standards.

6.7.1.3 Life Cycle‐Based Environmental Law

Life cycle concepts are increasingly becoming a part of environmental regulation, especially those involving GHG emissions. Now LCA is often used in writing about carbon accounting. In these times of heightened concern over climate change, individuals, organizations, and governmental agencies alike are eager to measure the release and impact of GHGs. For example, a variety of governmental agencies are developing approaches for regulating the emissions of GHGs associated with the production and use of fuels. As an example, the Energy Independence and Security Act (HR 6) of 2007 (EISA 2007) aims to: move the United States toward greater energy independence and security; increase the production of clean renewable fuels; protect consumers; increase the efficiency of products, buildings, and vehicles. States are also implementing GHGs emission regulation (USEPA 2010). For example, California Global warming Solutions Act of 2006 resulted in regulations that establish a limit for life cycle GHG emissions of transportation fuels. Both the California low‐carbon fuel standard and Section 526 of EISA require a life cycle evaluation of the GHG emissions of transportation fuels and this is becoming a common approach to considering GHG emissions (Allen et al. 2009).

6.7.1.4 Life Cycle Best Control Technology

A number of quantitative tools are emerging that enable life cycle assessments and analyses, and subsequently provide guidance on selecting best available control technologies for PSD, NPDES, and other environmental permitting processes. Many of our environmental laws are based on risk frameworks. Now, the life cycle–based best control technology decisions in environmental permitting processes are made that incorporate not only health and environmental but also the economic and societal impacts. Next, we present a case study on BACT for a power‐generating facility.

6.8.2.1 Combustion Turbines and Duct Burners

NO_x is produced through two mechanisms: high temperature processes, which create thermal NO_x (products of the reaction of nitrogen and oxygen gases in the air) and combustion of nitrogen‐containing materials, which produces fuel NO_x. Table 6.16 lists the technologies that were identified for controlling NO_x emissions from gas turbines and their effective emission levels.

6.8.2.2 SCONO_X

SCONO_X is an emerging proprietary catalytic and absorption technology that has shown some promise for turbine applications. Unlike selective catalytic reductions (SCRs), which requires ammonia injection, this system does not require ammonia as a reagent; its parallel catalyst beds are alternately taken off line through means of mechanical dampers for regeneration.

Despite its advantages, however, the process SCONO_X catalyst is subject to the same fouling or masking degradation that is experienced by any catalyst operating in a turbine exhaust stream. There is also a small energy loss from the performance loss due to the pressure drop across the catalyst.

The vendor of SCONO_X guarantees performance to all owners and operators of natural gas‐fired combustion turbines, regardless of size or gas turbine supplier. The system is designed to reduce both CO and NO_x emissions from natural gas‐fired power plants to levels below ambient concentrations. Indeed, the EPA considers SCONO_X a technically feasible and commercially available air pollution control technology and expects its emission levels for criteria pollutants such as NO_x, CO, and VOC to be comparable or superior to previously applied technologies for large combined cycle turbine applications.

SCONO_X has been demonstrated successfully on smaller power plants, including a 32 MW combined‐cycle General Electric LM2500 gas turbine in Los Angeles (Das 2003). This facility uses water injection in conjunction with SCONO_X to achieve a NO_x emissions rate of 0.75 ppm on a 15‐minute rolling average. The SCONO_X technology has also been successfully demonstrated on a 5 MW Solar Turbine Model Taurus 50 at the Genetics Institute in Andover, Massachusetts. The system is reducing NO_x down to 0.5 ppm NO_x, on a one‐hour rolling average. The permit for the power plant was originally issued for 2.5 ppm NO_x.

The manufacturer guarantees CO emissions of 1 ppm and NO_x emissions of 2 ppm. According to one set of figures, when NO_x is reduced from 12.18 ppm (gas turbine with duct burner firing) to 2 ppm, the cost effectiveness is $13 627/T of NO_x removed (Catalytica Energy Systems Inc. 2004).

6.8.2.3 XONON

Several companies are reported to be working on a second technology for the control of NO_x. Introduced commercially by Catalytica Combustion Systems, it is being marketed under the name XONON. This technology replaces traditional flame combustion with flameless catalytic combustion. NO_x control is accomplished through the combustion process using a catalyst to limit the temperature in the combustor below the temperature where NO_x is formed. The XONON demonstrated to achieve near‐zero emissions. The XONON combustion system consists of four sections: (i) the preburner, for start‐up, acceleration of the turbine engine, and adjusting catalyst inlet temperature if needed; (ii) the fuel injection and fuel‐air mixing system, which achieves a uniform fuel‐air mixture to the catalyst; (iii) the flameless catalyst module, where a portion of the fuel is combusted flamelessly; and (iv) the burnout zone, where the remainder of the fuel is combusted.

The single field installation of the XONON technology at a municipal power company is being used to perform engineering studies of the technology at Silicon Valley Power, in Santa Clara, California. NO_x emissions are well below 2.5 ppm on the 1.5 MW Kawasaki M1A‐13A gas turbine. Catalytica has a collaborative commercialization agreement with General Electric Power Systems, committing to the development of XONON. In conjunction with General Electric Power systems, the XONON system was specified for use with the GE 7FA turbines at the proposed 750 MW natural gas‐fired Pastoria Energy Facility, near Bakersfield, California. The project entered commercial operations in 2003. Because the NO_x emissions limitations of 2.5 ppm have been demonstrated in practice by a commercial facility, this technology is considered commercially available at this time (Catalytica Energy Systems Inc. 2004).

6.8.2.4 Selective Catalytic Reduction

SCR systems selectively reduce NO_x by injecting ammonia (NH₃) into the exhaust gas stream upstream of a catalyst. NO_x, ammonia, and oxygen react on the surface to form molecular nitrogen (N₂) and water. The overall chemical reaction can be expressed as

Parallel plates or honeycomb structures, permeated with the catalyst, are installed in the form of rectangular modules, downstream of the gas turbine in simple‐cycle configurations, and into the heat recovery steam generator (HRSG) portion of the gas turbine downstream of the superheater in combined‐cycle and cogeneration configurations.

The turbine exhaust gas must contain a minimum amount of oxygen and be within a somewhat narrow temperature range in order for the SCR system to operate properly. The temperature range is dictated by the catalyst: if it is too low, the reaction efficiency drops and increased amounts of NO_x and ammonia are released from the stack; if it becomes too high, the catalyst may begin to decompose. Turbine exhaust gas is generally too hot to be passed through the catalyst, so it is cooled by the HRSG, which extracts energy from the hot turbine exhaust gases and creates steam for use in other industrial processes or to turn a steam turbine. In simple‐cycle power plants where no heat recovery is accomplished, high temperature catalysts (e.g. zeolite) are an option. SCR can typically achieve NO_x emission reductions in the range of about 80–95%.

SCR is the most widely applied post‐combustion control technology in turbine applications and is currently accepted as LAER for new facilities located in ozone non‐attainment regions. It can reduce NO_x emissions to as low as 4.5 ppmvd for standard combustion turbines without duct burner firing and as low as 2–2.5 ppmvd when combined with lean‐premix combustion (again without duct burner firing). SCR uses ammonia as a reducing agent in controlling NO_x emissions from gas turbines. The portion of the unreacted ammonia passing through the catalyst and emitted from the stack is called ammonia slip. The ammonia is injected into the exhaust gases prior to passage through the catalyst bed. There is also a potential for increased particulate emissions from formation of ammonia salts. SCR may also results in the generation of spent vanadium pentoxide catalyst, which is classified as a hazardous waste. In addition, there is an energy loss from the performance loss due to the pressure drop across the SCR catalyst.

Gas turbines using SCR typically have been limited to 10 ppmvd ammonia slip (emissions of ammonia that has not reacted with nitrogen) at 15% oxygen. However, levels as low as 2 ppmvd at 15% oxygen have been proposed and guaranteed by control equipment vendors. In addition, Massachusetts and Rhode Island have established ammonia slip LAER levels of 2 ppmvd. Massachusetts has permitted at least two large gas turbine power plants using SCR reduction with 2 ppmvd ammonia slip limits. California recommended that the establishment of ammonia slip levels below 5 ppmvd at 15% oxygen on the basis of guarantees from control equipment vendors of single‐digit levels for ammonia slip.

Data supplied by the vendor show that when NO_x is reduced from 15 ppm (gas turbine with duct burner firing) to 3.5 ppm, the cost effectiveness is $9473/T of NO_x removed.

6.8.2.5 Lean‐Premix Technology or Dry‐Low NO_x

Processes that use air as a diluent to reduce combustion flame temperatures achieved reduce NO_x by premixing the fuel and air before they enter the combustor. This type of process is called lean‐premix combustion and goes by a variety of names, including the dry‐low NO_x (DLN) process of General Electric, the dry‐low emissions process of Rolls‐Royce, and the SoLoNO_x process of Solar Turbines.

Lean premixed designs reduce combustion temperatures, thereby reducing thermal NO_x. In a conventional turbine combustor, the air and fuel are introduced at an approximately stoichiometric ratio and air/fuel mixing occurs at the flame front where diffusion of fuel and air reaches the combustible limit. A lean premixed combustor design premixes the fuel and air prior to combustion. Premixing results in a homogeneous air/fuel mixture, which minimizes localized fuel‐rich pockets that produce elevated combustion temperatures and higher NO_x emissions. A lean air‐to‐fuel ratio approaching the lean flammability limit is maintained, and the excess air serves as a heat sink to lower combustion temperatures, which lowers thermal NO_x formation. A pilot flame is used to maintain combustion stability in this fuel‐lean environment. Lean‐premix combustors can achieve emissions of about 9 ppmvd NO_x at 15% oxygen (~94% control).

To achieve low NO_x emission levels, the mixture of fuel and air introduced into the combustor must be maintained near the lean flammability limit of the mixture. Lean‐premix combustors are designed to maintain this air/fuel ratio at rated load. At reduced load conditions, the fuel input requirement decreases. To avoid combustion instability and excessive CO emissions that occur as the air/fuel ratio reaches the lean flammability limit, lean‐premix combustors switch to diffusion combustion mode at reduced load conditions. This switch to diffusion mode means that the NO_x emissions in this mode are essentially uncontrolled.

Lean‐premix technology is the most widely applied precombustion control technology in natural gas turbine applications. It has been demonstrated to achieve emissions of approximately 9 ppmvd NO_x at 15% oxygen (Catalytica Energy Systems Inc. 2004).

6.8.2.6 Steam/Water Injection

In steam/water injection, the technology commonly chosen to reduce the NO_x emissions in natural gas turbine, higher combustion temperatures are used to achieve greater thermodynamic efficiency. In turn, more work is generated by the gas turbine at a lower cost. However, the higher the gas turbine inlet temperature, the more NO_x that is produced. Diluent injection, or wet controls, can be used to reduce NO_x emissions from gas turbines. Diluent injection involves the injection of a small amount of water or steam via a nozzle into the immediate vicinity of the combustor burner flame. NO_x emissions are reduced by instantaneous cooling of combustion temperatures from the injection of water or steam into the combustion zone. The effect of the water or steam injection is to increase the thermal mass by mass dilution and thereby reduce the peak flame temperature in the NO_x forming regions of the combustor. Water injection typically results in a NO_x reduction efficiency of about 70%, with emissions below 42 ppmvd NO_x at 15% oxygen. Steam injection has generally been more successful in reducing NO_x emissions and can achieve emissions less than 25 ppmvd NO_x at 15% oxygen (~82% control).

6.8.2.7 Summary of NO_x BACT for Turbines and Duct Burners

Table 6.17 provides information on the emissions, control effectiveness, economics, and environmental impacts measures for the control of NO_x discussed in Sections 6.8.2.2, 6.8.2.4, and 6.8.2.5_. The analysis was performed on a unit (turbine and duct burner) basis.

SCONO_X provides the highest level of NO_x reduction. However, this very new technology has yet to prove itself for long‐term commercial operation on large‐scale combined‐cycle plants. It is the relatively high cost per emission reduction of this control technology ($13 627/T of NO_x removed) that rules out SCONOx as a control option. The next most effective control technology for NO_x is a combination of DLN combustors and SCR. The adverse environmental impact of SCR is primarily from the emissions of ammonia which is on the EPA's list of extremely hazardous substances. Although these adverse impacts of the SCR process can be minimized with proper system design and operation, it is ruled out as a control operation because the cost is prohibitive ($10 191/T of NO_x removed).

The next most effective control technology is DLN combustors. At reduced loads, combustion instability requires that a switch to diffusion combustion mode, in which NO_x emissions are essentially uncontrolled. However, this can be minimized with proper system design and operation.

6.8.3.1 Catalytic Oxidation

The rate of formation of CO during natural gas combustion depends primarily on the efficiency of combustion. The formation of CO occurs in small, localized areas around the burner where oxygen levels cannot support the complete oxidation of carbon to CO₂. Efficient burners can minimize the formation of CO by providing excess oxygen or by mixing the fuel thoroughly with air. CO emissions resulting from natural gas combustion can be decreased via catalytic oxidation. The oxidation is carried out by the well‐known overall reaction:

Several noble metal‐enriched catalysts at high temperatures promote this reaction. Under ideal operating conditions, this technology can achieve an 80% reduction in CO emissions. Prior to entering the catalyst bed where the oxidation reaction occurs, the exhaust gas must be preheated to about 400–800 °F. Below this temperature range, the reaction rate drops sharply, and effective oxidation of CO is no longer feasible. Moreover, there is an energy loss because of the reduction in performance due to the pressure drop across the CO catalyst.

Sulfur and other compounds in the exhaust may foul the catalyst, leading to decreased activity. Catalyst fouling occurs slowly under normal operating conditions and may be accelerated by even moderate sulfur concentrations in the exhaust gas. The catalyst can be chemically washed to restore its effectiveness, but eventually, irreversible degradation occurs. Catalyst replacement is usually necessary every 5–10 years depending on the type and operating conditions.

An economic analysis for the catalytic oxidation of CO emissions based on vendor information estimates the cost at $5084/T of CO removed. This cost level is considered to be economically infeasible for BACT. In addition to cost, catalytic oxidation would lead to increased downtime for catalyst washing and would present hazardous waste concerns during catalyst disposal. Due to the high cost and concerns with downtime and hazardous material disposal, catalytic oxidation is not selected as BACT for control of CO emissions from the turbines and duct burners.

6.8.3.2 Good Combustion Practices

Clearinghouse (RBLC) data show that the majority of BACT determinations for CO relied on the use of good combustion practices. Since add‐on controls for CO were shown to be economically infeasible, the proposed BACT for CO emissions is the use of good combustion practices.

6.8.3.3 Combustion Control

Because CO results from the incomplete combustion of fuel, combustion control is an inherent design feature of combustion turbines and duct burner. Control is accomplished by providing adequate fuel residence time and high temperature in the combustion zone to ensure complete combustion. These control methods, however, also result in increased emissions of NO_x. Conversely, a low NO_x emission rate achieved through flame temperature can result in higher levels of CO emissions. Thus, a compromise is needed to set the flame temperature at the level that will achieve the lowest NO_x emission rate possible while keeping CO emissions rates at acceptable levels.

6.8.3.4 Summary of CO BACT for Turbines and Duct Burners

Table 6.18 provides information on the emissions, control effectiveness, economics, and environmental impacts associated with control of CO. The analysis was performed on a unit (turbine and duct burner) basis.

SCONO_X provides a higher level of CO reduction than a combination of DLN combustors and CO oxidation. The adverse impacts include an energy loss from the performance loss due to the pressure drop across the CO catalyst and emissions of sulfates condense as additional PM₁₀ or PM_2.5. However, at a cost of $5084/T of CO removed, and the second option is more than four times as cost‐effective as a control option.

Combustion control is selected as BACT for CO control with the limit of 9 ppm at 15% O₂ (annual average) without duct burners firing and 11.5 ppm at 15% O₂ with duct burner firing (annual average).

6.8.4.1 Step 1: Identify Potential Control Technologies

The first step in the BACT analysis is to identify all potential control technologies available for the control of PM/PM₁₀ emissions. The technologies identified for the control of PM/PM₁₀ emissions are as follows:

• Electrostatic precipitation (ESP)
• Fabric filter
• Good combustion practices
• Fuel specification: clean burning fuels

6.8.4.2 Step 2: Eliminate Technically Infeasible Options

The second step in the BACT analysis is to eliminate any technically infeasible control technologies. Each control technology for each pollutant is considered, and those that are clearly technically infeasible are eliminated. The following is a list of control options deemed technically infeasible for the control of PM/PM₁₀ emissions.

6.8.4.3 Electrostatic Precipitators

ESP technology removes particulate from an exhaust stream by electrically charging the particles and collecting the charged particles on plates. ESP performance is greatly affected by the particles' ability to accept and maintain an electrical charge. Because of the resistivity of gas turbine exhaust particles, ESP technology is ineffective for the control of PM/PM₁₀ emitted from the proposed PSE (Puget Sound Energy) turbine.

Electrostatic precipitators are also not considered technically feasible options for combustion turbines due to the high exhaust flow rate and the low concentration of particulate in the turbine exhaust.

6.8.4.4 Fabric Filters

Fabric filters remove PM/PM₁₀ from an exhaust stream by collecting the particulate on the filter as the air stream passes through the filter. Fabric filters typically cannot withstand high exhaust temperatures (>500 °F) and would be damaged by the high temperature of turbine exhaust. Thus, fabric filters are not technically feasible for the proposed PSE turbine.

6.8.4.5 Step 3: Ranking of Remaining Control Technologies by Control Effectiveness

The third step in the BACT analysis is to rank the remaining control technologies in order of control effectiveness. Table 6.19 presents the control technologies and their approximate control efficiencies.

6.8.4.6 Step 4: Evaluation of the Most Effective Emissions Controls

Because ESP technology and fabric filters have been deemed technically infeasible, the implementation of good combustion practices and the firing of clean burning fuels are the only remaining options for the control of PM/PM₁₀ emissions. The combination of implementing good combustion practices and the firing of clean burning fuels is the most effective emission control option.

6.8.4.7 Step 5: Select BACT for the Control of PM/PM₁₀ Emissions

Properly tuned burners firing natural gas and light oils inherently emit low levels of particulate matter. The RBLC database indicates that good combustion control is widely accepted as BACT for turbines firing natural gas. Thus, it is selected to use of clean burning fuels (natural gas and fuel oil with a sulfur content <0.05%) and good combustion practices as BACT for PM/PM₁₀ emissions.

6.8.5.1 Catalytic Oxidation

The formation of VOC in combustion units depends primarily on the efficiency of combustion. Inefficient combustion leads to the formation of aldehydes, aromatic carbon compounds, and various other organic compounds by several mechanisms. Catalytic oxidation decreases VOC emissions by facilitating the complete combustion of organic compounds to water and carbon dioxide. Prior to entering the catalyst bed where the oxidation reaction occurs, the exhaust gas must be preheated to about 400–800 °F.

The RBLC database shows few instances of catalytic oxidation being selected as BACT for VOC at any gas‐fired turbine power plant nationwide. An economic analysis for the catalytic oxidation of VOC emissions based on vendor information estimates the cost at $30 811/T of VOC removed. This cost level is economically infeasible for VOC removal. Thus, catalytic oxidation would not be selected as BACT for VOC emissions from turbines and duct burners proposed in the present time frame.

6.8.5.2 Good Combustion Practices

All the RBLC database BACT determinations for VOC outside California and New York show the use of combustion control or good combustion practices. Thus, in most areas good combustion practices are acceptable as BACT for VOC, with limits of 7.25 ppmvd at 15% O₂, as methane.

6.8.6.1 Step 1: Identify Potential Control Technologies

The first step in the BACT analysis is to identify all potential control technologies available for the control of SO₂ emissions. The technologies identified for the control of SO₂ emissions are as follows:

• FSD
• Spray dryers
• Fuel specification: low‐sulfur fuels

6.8.6.2 Step 2: Eliminate Technically Infeasible Options

Neither FGD systems nor spray dryers have been applied to the exhaust gases from turbines, and significant technological difficulties are envisioned to apply to either of these technologies with high exhaust temperatures. The high temperature of the turbine exhaust deems FGD units and spray dryers infeasible. Thus, both FGD systems and spray dryers are eliminated for the control of SO₂ emissions from the proposed PSE turbines.

6.8.6.3 Step 3: Ranking of Remaining Control Technologies by Control Effectiveness

All add‐on control options for the control of SO₂ emissions have been eliminated due to technical unfeasibility. For the control of SO₂ emissions, the use of low‐sulfur fuels is the only control method determined to be technically feasible.

6.8.6.4 Step 4: Evaluation of the Most Effective Emissions Controls

As stated above, all add‐on control options for the control of SO₂ emissions have been eliminated due to technical unfeasibility. Thus, the use of low‐sulfur fuels containing less than 0.05% by weight of sulfur is the most effective means for reducing SO₂ emissions.

6.8.6.5 Step 5: Select BACT for the Control of SO₂ Emissions

The low SO₂ emissions levels inherent with low‐sulfur fuel in a turbine constitutes BACT. Thus, the permittee proposes that BACT for the proposed turbines is the firing of low‐sulfur fuels.