Intel
EPA Response
Response to Comments on
Intel's Final Project Agreement
Table of Contents
Summary Table of Comments and Responses
Clarification of Underlying Facts/Assumptions/Questions
Key to Identification of Commentors
Summary of Issues Requiring Attention During FPA Implementation
Major Concerns on FPA/Process
Table 1: Comparison of current regulatory requirements to Intel XL Performance
Table 2: Emissions for selected HAP compounds for semiconductor facilities
In establishing Project XL, the United States Environmental Protection Agency offers regulated entities the opportunity to develop pilot projects that receive regulatory flexibility while generating superior environmental performance. An early participant in the XL process, Intel Corporation spent 9 months as part of a stakeholder team-composed of representatives from the company, the local community, Gila River Indian Community and state, local, and federal regulators-developing a pilot project at its Ocotillo site in Chandler, AZ. The details of this project are described in a Final Project Agreement (FPA) which was released for public comment on August 15, 1996, amended by an October 9, 1996 version and finalized on October 18, 1996.
EPA would like to thank those who submitted comments during the public comment period. As a result of some of the comments, EPA believes that the Intel XL project has become an even better pilot project. EPA offers these additional responses to complement Maricopa County's official response to comments on the air permit and the stakeholder's response to comments on the FPA.
During a review of all the public comments by the stakeholder team, the public comments were grouped into four categories (FPA, Intel Process, XL General, and Air Permit comments) to determine how best to respond to comments. Those that are numbered are considered to be major comments and are addressed in more detail after the summary table. These commentors are identified by their initials-a "key" appears at the end of the next section. The issues are presented in the following order, by general category.
Summary Table of Comments and Responses
(see attached key to identification of commentors)
CATEGORY
ISSUE: Raised by
RESOLUTION:
FPA COMMENTS:
2. The FPA does not represent superior environmental performance: ED, SN, SL
FPA does represent superior environmental performance.
5. The provisions for water conservation and re-use are inadequate: Roger Peterson, SN
Wastewater used for irrigation and ground water recharge rather than sent to POTW.
6. There is completely inadequate community and employee involvement: MG, RH, SN
Refer to Administrative Record on public participation process
7. There is inadequate pollution prevention: TS, SL
The FPA contains both pollution prevention as well as DFE principles.
12. There are problems with the method for generating the emission per production unit: TS
Intel generating company wide number. PSEL plus emission per production unit will prevent extra emissions if there is no plant expansion.
Termination procedures should include all stakeholders: NRDC
Stakeholder notice and participation for ADR are provided for in procedures for termination of the FPA.
The consolidated report should be expanded to include chemical use information: Cole, HR, TS
Consolidated report was constructed by stakeholder group including community members. Report is not intended as a chemical inventory, but a report on environmental goals in the FPA. CAP will review report can make recommendations to change format at any time. A copy of EPA's ANPR on expanding TRI was sent to stakeholders with a recommendation to wait for the outcome, or at least initial comments on ANPR, of the national debate before discussing.
Use the Silicon Valley Toxics Gas Ordinance for emergency reporting requirements
The Intel stakeholder team believed that the City of Chandler provided excellent resources to develop an electronic, state-of-the-art Hazardous Materials and Management Plan. In addition, the TGO is for controlling storage of chemical outside of a facility that does not apply to the FAB-12 plant.
INTEL PROCESS
Intel should reuse waste water: SN
Re-use of treated wastewater causes bacterial buildup in the system. Intel is funding SEMATEC to find out why this occurs.
Will there be an odor impact: RH
FAB-12 does not use any odiferous chemicals
XL GENERAL:
8. Independent technical/financial support is needed for communities: MG, SN, DWA, Cole, TS, RH, SL
EPA will examine different methods to provide resources to community groups in XL projects, and will fund a pilot for community groups to assist with implementation of the Intel FPA.
9. The transferability of this FPA should be done carefully: RH, DWA
XL process will support good actors.
10. Employee protection should be improved: MG, NRDC, SL
Info. to Admin. Rec. on safety teams plus OSHA VPP info. sent to Interested Parties/Stakeholders.
11. Intel should not claim CBI: TS, NM, SL, CEP
State laws give Intel the right to protect its "recipe". MCESD must verify all data, including emission factors, claimed as CBI.
13. The FPA should be enforceable: SL, NRDC, CRT
Air permit now includes enforceable provisions for public participation and reporting requirements.
Communities need assistance on understanding EPA models: SN
See Technical Assistance Issue
Synergistic effects are not evaluated for exposures: SN, NRDC
Current toxicological methods do not have an accepted method of evaluating for synergistic effects
AIR PERMIT
1. The Community and workers could be subjected to higher exposures of extremely toxic chemicals than the current permit allows. The standards (air emission) set in the FPA are too weak. The FPA allows for significant risk shifting to higher levels of exotic, untested chemicals. Includes desire for public participation on Strategic Chemical Committee (SCC): SN, SL, NRDC, NM, TS
Changed reporting limit for phosphine, and sulfuric acid to reflect actual amount release. New permit reflects current operating requirements and goes beyond requirements by using AAAQGs. Pollution per production unit will also keep emissions down in case there is no plant expansion.
Intel to keep SCC an internal-only process.
3. FPA or permit should include third party verification/On-line testing/stack testing for new chemical: HR, DWA, MG, SN, Cole,TS
Refer to CAP for discussion/evaluation. EPA will provide information from pilots on third-party verification.
4. The Arizona Ambient Air Quality Guidelines do not reflect chronic health information: Cole, CRT
Chronic information is included for carcinogens
10. Air permit levels do not protect workers: SN, SL , NRDC
Limits on routine emissions are protective of employees by using the 1-hour AAAQGs. Emergency upsets are covered under the HMMP.
BACT should be applied at the tool level rather than at the end-of-the-pipe: NRDC
As a minor source, Intel is only required to have BACT at the end-of-the-pipe. Through DFE, Intel is committed to reducing chemical use.Clarification of underlying facts/assumptions/questions
The following were questions sent in during the public comment period on the underlying assumptions and facts for the Intel XL project. The following responses are meant to clarify these facts and assumptions:
1. "Water recycling commitments in FPA are less stringent than in contract with City"
(Roger Peterson): Intel has a contract with the City of Chandler to accept a certain amount of treated water from the City for use in landscaping and cooling towers. The FPA is the only document that spells out increased recycling of fresh City water.
2. "The AAAQGs do not take into account off-gassing from soils" (RH): There are no sources of contaminants in the soils at the Chandler site.
3. "The FPA does not include product stewardship" (SL, HR): The FAB-12 plant makes wafers that are the building blocks for other components. Intel commits in the FPA to continue its Design for the Environment (DFE) and other similar activities at the FAB-12 site.
4. "The consolidated reporting is not multi-media oriented" (HR): The consolidated report does look at all media in an integrated report.
5. "There are inadequate standards for protecting the aquifers receiving Intel's wastewater" (SN): Water that leaves the City of Chandler reverse osmosis treatment plant exceeds drinking water standards before injection. This requirement is part of the City's Aquifer Protection permit issued by ADEQ.
6. " There is no 'stakeholder' support for this project by community" (SN): There are four public members on the Intel XL stakeholder team. These individuals have put in tremendous effort to the FPA. They have made public appeals to have other community members become involved in the process. There has not been overwhelming response due to the steep learning curve, time commitment, and, as stated by some public members during public meetings to EPA, their belief that the process is open, accessible and that they are well represented.
7. "Intel could terminate the FPA and have a more flexible air permit without the commitments of the FPA" (NRDC): If Intel were to terminate the FPA, they would revert to the existing regulatory systems. In addition, the air permit includes the public participatory process, as well as the option for the County to reopen the permit, as set forth in County rules.
8. "What is the method to alert community to disaster?" ( DWA): Emergency plan developed in conjunction with City of Chandler Fire Department.
9. "Is Gila River's Indian community participating including both the Maricopa and Pima Tribes?" Yes. The Gila River Indian Community is composed of the Maricopa and Pima tribes.
Key to Identification of Commentors
MG: Michael Gregory
SN: Southwest Network for Environmental and Economic Justice
CRT: Campaign for Responsible Technology (CRT)
DWA: Don't Waste Arizona
Cole: Henry S. Cole & Associates
NM: New Mexico Environmental Law Institute
SL: Sanford Lewis (Good Neighbor Project)
TS: Ted Smith (Silicon Valley Toxics Coalition)
RH: Raena Honoan (Sierra Club)
HR: Hampshire Research Institute
NRDC:Natural Resources Defense Council
EDF: Environmental Defense Fund
CEP: Council on Economic Priorities
Summary of Issues Requiring Attention During FPA Implementation
During the development of the FPA, a number of issues were raised that did not get fully discussed and resolved. This XL project is based on a process that will continue past the signing of the FPA into the implementation phase over the next 5 years of the FPA's duration. There will be a number of public forums for these issues to be discussed as well as the opportunity for the stakeholder team to continue the valuable work it has done to date in resolving a large number of issues. Part of the benefit of this project is the commitment to ongoing relationship and the information sharing as opposed to a permit-by permit ad hoc approach. Besides continuing these discussions within the stakeholder team, outstanding issues will be brought to the public at either the upcoming quarterly or semi-annual meeting. The following is a list of some of the issues that were raised during the FPA negotiations or as a result of the public comment period and will continue to be discussed during implementation:
-Reclassify hazardous waste to chemical waste
-Create a one-stop shop for permit issuance and regulatory oversight
-Explore the feasibility of farming pickleweed with treated wastewater
-Determine if there are any more precise analytical methods available for measuring air emissions that have been developed and proven for use in the semiconductor industry
-Report from Intel on their participation in the OSHA Voluntary Protection Program (an OSHA excellence program)
-Discuss any process/chemical changes and the effect these changes will have on emissions,
-The stakeholder team still needs to come to consensus on what information should be reported related to employee health at FAB-12.
EPA will also be following-up on certain issues being discussed nationally that have a direct impact on the Intel XL project. EPA will continue to communicate with the Intel XL stakeholder team and the public by presentations at the quarterly CAP meetings or semi-annual public meetings. These issues include:
-providing technical assistance to community stakeholders
-third-party verification of environmental compliance
-expanding TRI data collection to include chemical use information
There were a number of issues raised during the public comment period for the FPA. This response to comments will be provided to interested parties and included as part of the FPA's Administrative Record that is at the Arizona Department of Environmental Quality. Below is a summary of the major issues raised during the comment period, the parties that stated the issue and EPA's response:
ISSUE #1: The Community and workers could be subjected to higher exposures of extremely toxic chemicals than the current permit allows (1994 permit as compared to the permit recently issued). The FPA would allow an increase in the phosphine emissions by a factor of 400.
Commentors: SN, SL, NRDC, CRT, Cole
Discussion: An overview will first be presented and then a more detailed discussion will follow.
A. Overview: There have been a number of comments that the original air permit for the facility was more protective than the current permit. The Intel air permit issued by the Maricopa County Environmental Services Department (MCESD) in 1994 to allow the construction of FAB-12 should not be compared to the recent air permit issued by MCESD. The 1994 construction permit was based on estimates made by Intel in 1993 predicting future operations after construction. This is important since the 1994 permit did include specific emission limits for individual Hazardous Air Pollutants (HAPs) and criteria pollutants, many of which are lower than the aggregate emission levels in the 1996 permit. The FPA development process that led to the 1996 permit created a more comprehensive permit than a permit Intel would have obtained without Project XL.
There are three reasons why the emission estimate data supplied to MCESD in 1993 were no longer accurate for the 1996 processes at FAB-12:
1. Intel wanted to develop, under Project XL, a comprehensive environmental master plan for the entire Ocotillo site that would include air emissions for the FAB-12 plant and a possible second semiconductor plant.
2. Intel will be using different chemicals and tool configurations at FAB-12 as there has been a process change since the 1994 FAB-12 permit application was submitted. This change in processes resulted in Intel needing to add new chemicals and discontinue use of others. For example, Intel now expects to use methanol and no longer plans to use diborane and silane in a few process steps;
3. Intel stated that they were informed earlier this year by the pollution control manufacturer that the air emission control equipment would not operate as originally expected at FAB-12. The pollution controls (wet scrubbers) for the inorganic HAP emissions was not as efficient as had originally been estimated because of the low inlet concentrations.
Therefore, the 1994 permit was obsolete and Intel required permit modifications to reflect actual operations whether or not it decided to participate in Project XL. MCESD issued a modified permit in November, 1996. The new permit is more comprehensive protective than the 1994 permit and those normally issued by Maricopa County because:
1. The combined emissions of FAB-12 and any other air emission source built at the Ocotillo site will be a minor stationary air source. All emission limits are well below the minor source threshold and the criteria pollutants are less than half.
2. As an additional safeguard, emission limits for all chemicals are established by the use of Arizona Ambient Air Quality Guidelines (AAAQGs). As a legally-enforceable component of its air permit, Intel must apply the screen modeling analysis to any compound it uses that is listed among the 189 HAPs currently listed under federal law. Moreover, Intel will perform the same analysis for any of the 273 other compounds listed in the AAAQGs. Before using any chemical that generates air emissions that might pose a risk to human health or the environment in its processes, whether or not regulated under section 112 of the Clean Air Act, Intel will perform calculations to demonstrate that AAAQGs would not be exceeded beyond the property line. If the model demonstrates otherwise, Intel would limit annual emissions of any such chemical to ensure consistency with the AAAQGs.
The idea of this analysis is to estimate a worst case for each chemical, determine whether that case would pose any risk, and if so, to put in place limits on that pollutant that prevent such risk. As part of the FPA, Intel will perform the same modeling within the site to provide additional protection for employees and site visitors. Again, this kind of protection goes well beyond what Intel was required to do by the 1994 permit or would be required to do without XL;
3. Another voluntary commitment is the FPA is that the Intel XL project will establish an emissions per production unit limit that is more protective of human health and the environment than they would be under required or expected future performance.
B. Detailed Discussion:
Maricopa County Environmental Services Department (MCESD) permits generally include numerous operating conditions that the permittee must follow including emission limits for criteria air pollutants and, if necessary, hazardous air pollutants (HAPs). These emission limits, expressed as lb./day or ton/year, are established by MCESD after careful review of the information the applicant submits in a permit application. Both the 1994 permit and the proposed permit under project XL limit Intel to emissions below the major source levels. EPA believes that this is a significant accomplishment for a semiconductor manufacturing facility of this size.
As noted above, an air permit was issued by MCESD in April 1994 for the FAB-12 plant and includes specific emission limits for HAPs and criteria pollutants. To develop a best estimate of the future emissions at the unbuilt FAB-12 plant, Intel's application for the 1994 permit used control technology information and chemical use data from its existing plants in New Mexico and Santa Clara, California.
It recently became apparent to Intel that there were several reasons why the emission estimate data supplied to MCESD in 1993 was no longer accurate for the 1996 processes at FAB-12. First, Intel no longer uses the same processes anticipated when the 1993 FAB-12 permit application was submitted. This change in processes includes adding new chemicals and discontinue use of others. For example, Intel now expects to use methanol (an organic HAP) and no longer plans to use diborane and silane in a few process steps (non-HAPs with specific limits in the 1993 permit).
Second, Intel discovered that the control technology (wet scrubbers) for the inorganic HAP emissions was not as efficient as originally estimated because of the low inlet concentrations. Based on tests of similar wet scrubbers and operations at the FAB-11 plant in New Mexico, Intel discovered that the removal efficiencies of the wet scrubbers at FAB-12 ranged from 20% to 50% - down from their original estimate of 90%, which is information provided by the scrubber manufacturer based on the assumption that the inlet concentration will be much higher. This lower removal efficiency is due to the very low concentration of chemicals in the process air (60,000 cubic feet each minute per scrubber) that passes through the FAB-12 plant. This low removal efficiency only occurs when contaminant concentration in the air stream is low; as contaminant concentration increases, the scrubber efficiency increases which can also dampen peaks or periods of higher emissions within the plant.
Finally, Intel wanted to develop, under Project XL, a comprehensive environmental master plan for the entire Ocotillo site that would include air emissions for the FAB-12 plant and a possible second semiconductor plant.
MCESD issued a modified permit to Intel in November, 1996 based on recommendations from the XL stakeholder team. This air permit has an additional safeguard by including the use of Arizona's Ambient Air Quality Guidelines (AAAQGs) . Before using any chemical that provides air emissions that might pose a risk to human health or the environment in its processes, whether or not regulated under the Clean Air Act, Intel would perform screen modeling at 10-ton/year levels to demonstrate that AAAQGs would not be exceeded at the property line. If the model demonstrates otherwise, Intel would limit annual emissions of any such chemical to ensure consistency with the AAAQGs. The idea of this analysis is to estimate a worst case for each chemical, determine whether that case would pose any risk, and if so, to put in place limits on that pollutant that prevent such risk.
Finally, Intel has committed to establishing emission limits for all chemicals of concern, whether regulated or unregulated. As a legally-enforceable component of its air permit, Intel must apply the screen modeling analysis to any compound it uses that is listed among the 189 HAPs currently listed under federal law. Moreover, Intel will perform the same analysis for any of the 273 other compounds listed in the AAAQGs. As an additional voluntary commitment in the FPA, Intel commits to working with state and local regulators to determine whether any chemical it uses present any concern to human health and the environment, and if so, to establish emission limits for such chemicals.
Based on public comments, the Intel XL stakeholder team agreed to change the reporting requirements to report the actual amount of phosphine and sulfuric acid emitted on a quarterly and annual basis. This amount will most probably be below the 1000 pound cut-off currently stated in the FPA. The amount of phosphine reported as released will be the amount brought onto the site less any not consumed during the reporting period.
Based on the 1-hr AAAQG for protection of employees, the allowable level for phosphine emissions in the permit will still be set at 4.0 tons. In addition, the sulfuric acid value will be set at 9.0 tons per year. This is done to be consistent with how emission levels are set for all of the other chemicals emitted from the plant.
For HAPs other than phosphine, comparing the various pound per year limits in the 1994 permit to the 9.9 tons/year value is inappropriate for three reasons. First, the 10 ton/year limit in the Project XL permit is the limit for the sum total of all inorganic HAPs. Not only is it unrealistic to assume that the entire 10 tons will be consumed by only one HAP, but the commentor also assumes that all compounds can be emitted at 9.9 tons/year. This is not accurate. The Stakeholder team agreed to use the 10 tons/year value as the most conservative emission rate over the entire year.
Second, the separate aggregate 10 tons/year limits for aggregate inorganic HAPs and organic HAPs represents expected emissions for the entire Ocotillo site including a planned second manufacturing plant.
Resolution: The FPA will be modified to reflect the new reporting levels and process for phosphine and sulfuric acid.
ISSUE #2: The FPA does not represent superior environmental performance.
Commentors: ED, SN, SL, NRDC
Discussion: The standard for superior environmental performance for Intel's XL project was articulated in the original solicitation for Project XL, published in the May 23, 1995 issue of the Federal Register. Intel's project must "achieve environmental performance that is superior to what would be achieved through compliance with current and reasonably anticipated future regulation." When measured against this standard, Intel's XL project will achieve superior environmental performance over its 5-year life lifetime.
Table 1 presents the various elements of the Intel FPA and Maricopa County air permit, the environmental performance that Intel is committing under Project XL. For purposes of comparison, the environmental performance required of Intel under current requirements is also included. As outlined in Issue #1, the original 1994 permit is no longer applicable. Therefore, the only way to compare Intel's XL project to current requirements is to look at the requirements for a minor stationary source in the Phoenix area since Intel has committed the facility, at full capacity with additional manufacturing plants, to stay as a minor source of air pollution.
As Table 1 indicates, Intel's XL project achieves a level of environmental performance that surpasses what Intel would have otherwise been required to achieve outside of XL and under existing regulatory requirements. In this project, Intel has committed-both in the FPA and the Maricopa County air permit-to complying with all applicable legal requirements. Within those requirements, the permitting authorities are providing flexibility to Intel in the form of pre-approval of operating changes, including the construction of a new FAB at the Ocotillo site, and aggregated limits for air pollutants.
Comparison of current regulatory requirements to Intel XL Performance for elements
Elements: Current Requirements
Intel Project XL PerformanceAir: VOCs: <100 tons/year
NOX: <100 tons/year (moderate)
CO: <100 tons/year
Particulate: <70 tons/year
Sulfur dioxide: <250 tons/year
Hazardous Air Pollutants: <25 tons/year aggregate; 10 tons for a single HAP
VOCs: 49 tons/yearNOX: 49 tons/year
CO: 49 tons/year
Particulate: 5 tons/year
Sulfur dioxide: 5 tons/year
Hazardous Air Pollutants: 20 tons/year aggregate; 10 tons total for both organic and organic HAPS and each HAP not to exceed the Arizona Air Quality Guidelines. Phosphine at 4 tons/year and sulfuric acid at 9 tons/year.
VOC/HAP emissions per production unit to be set by January, 1997.
Emergency Response Planning: A number of separate emergency planning reports are required by various statutes.
With the assistance of the City of Chandler Fire Department, Intel will be preparing a single Emergency Response Plan for better use and accessibility to the public and emergency responders.
Environmental Reporting: Reporting of air and water emissions, pollution prevention efforts, and hazardous waste storage and movement is sent to one or more regulatory agency and time-consuming to access by the public. There are no public meetings to discuss the data.
Intel has combined routine environmental reporting requirements and accountability measures for voluntary goals set forth in the FPA into a single, integrated report that will be publicly available and distributed to interested stakeholders on a quarterly and annual basis, both in hard copy at ADEQ and on the Internet. The public participation process and reporting requirements are enforceable in the air permit.
Pollution Prevention: ADEQ has a requirement for pollution prevention plan to be submitted by company. This plan is submitted annually and shows how the facility will implement pollution prevention to minimize waste. There are no enforceable aspects to ADEQ's statute.
Intel's DFE program aggressively supports pollution prevention requirements by redesigning each new process to use fewer and less toxic chemicals. Intel reduced VOC emissions in the new process at FAB-12 by 40% compared to the previous process generation. The goals for increased recycling of chemical and solid waste and water are described below.
Public Participation: Community involvement in pre-permit negotiations is typically limited to a 30-day comment period that may end in a public hearing. No community input for environmental operations that are not subject to permitting or statutory requirements.
Community members were intimately involved during the nine months of negotiating the FPA and associated documents. There were seven public meetings and a public hearing to discuss the work in progress. Meeting notes and internally reviewed and approved documents were placed on the Internet. Community members also had input into such issues as solid waste recycling goals and water recycling goals that are normally not open for their input. A significant effort was made to address concerns raised during the process. When the stakeholder team concluded their work in August, there was consensus on the FPA.
Employee Protection: Manufacturing facilities must meet OSHA PELs for employee exposure to chemicals. This is an 8-hour exposure scenario
Intel uses the more stringent ACIGH TLVs for setting employee exposure limits to chemicals. In addition, Intel will voluntarily limit stack emissions so that the 1-hour AAAQG is not exceeded anywhere within the site as additional employee protection.
Any new chemicals that are brought into the process are reviewed by the internal Strategic Chemical Council. Changes to the process at FAB-12, including the introduction of new chemicals, will be discussed at the quarterly CAP meetings or semi-annual general public meetings.
There are also existing employee safety teams that identify hazardous working conditions and recommend improvements. In addition, Intel is considering joining the OSHA VPP.
Water Use: No requirements for either recycling water or providing a beneficial use of wastewater.
Intel spent $28 million dollars for a treatment plant to treat wastewater to meet drinking water standards for reinjecting into the aquifer and sent to canals for irrigation. FPA calls for improved water recycling with a final goal of 65% recycling with interim milestones and reporting.
Intel's ultimate goal is to reuse as much fresh water as possible. At the moment, full recycling in this type of manufacturing facility is not possible. Therefore, Intel is also funding research with SEMATECH on overcoming these obstacles..
Stormwater: Stormwater must be managed on site, and it is usually sent to dry wells requiring an Aquifer Protection Permit.
Intel pumps stormwater to a retention basin and does not use dry wells. The retention basin decreases the impact on ground water (greater evaporative losses) and no need to send stormwater to Chandler POTW.
Solid Waste: No requirements for recycling solid waste.
FPA calls for increasing recycling over time with a final goal of 60% with interim milestones and reporting.
Chemical /Hazardous Waste: No requirements for recycling chemical (non-hazardous) waste.
The FPA calls for increasing chemical waste recycling over time with a final goal of 70% by reclassifying the hazardous waste as non-hazardous chemical waste. By seeking regulatory flexibility, this reclassification will make recycling the waste much easier and can provide a higher quality recycled product. The chemical waste will still be handled in a manner that is protective of human health and the environment.
Ultimately, Intel will significantly increase the total percentage of hazardous and chemical waste it recycles. Because the smaller amount of hazardous waste that is left over is more difficult to recycle, the percentage of hazardous waste to recycle will drop from 60% in 1997 to 40%.
"Outside the fenceline": The only current requirement is for trip reduction as implemented by the County.
Intel has committed to a 1000 ft. property setback between FAB-12 and residential areas, equipment donation, environmental education and mentoring. Intel has committed to investigate additional trip reduction requirements.
All environmental aspects of this project must be considered in determining the superior nature of this project's environmental performance rather than focusing on any single medium. Table 1 indicates the multi-media nature of Intel's XL project that establishes voluntary commitments to reduce waste generation and water use, along with legally-binding commitments to comply with risk-based air standards.
Most commentors who questioned the superior benefits provided by Intel's XL project focused on the air-related component of the project, particularly the set of commitments addressing hazardous air pollutants (HAPs). With regard to the air emissions, the Intel XL project goes beyond current requirements in the following manner:
1. The Intel XL project establishes HAP limits that are more protective of human health than they would be under required performance.
First, to merit minor source classification, the Clean Air Act establishes a 25-tons/year limit for emissions from all HAPs combined and a 10-tons/year limit for emissions from any one HAP. Under XL, Intel will establish a 20-tons/year limit for all HAPs and a 10-tons/year limit each for emissions from all organic and inorganic HAPs. The 10-tons/year limit for inorganic HAPs includes a 5-tons/year limit for emissions from phosphine, a chemical of particular concern.
Second, Intel would supplement these over-all HAP limitations by demonstrating that emissions from their site will not exceed Arizona Ambient Air Quality Guidelines (AAAQGs). The AAAQGs are risk-based standards that make a number of conservative assumptions: exposure at a continuous concentration for 24 hours a day/7 days a week (including chronic exposure for carcinogens); a safety factor of 30 for non-carcinogens; and a one-in-a-million excess lifetime cancer risk for carcinogens. AAAQGs have been developed for 1-hour, 24-hour, and some annual emission concentrations.
2. Use of AAAQGs minimizes use of more toxic HAPs.
Commentors have questioned whether aggregated limits for organic and inorganic HAPs are more protective of human health than individual limits for specific types of HAPs. They have questioned whether Intel might use more toxic HAPs in lieu of less toxic HAPs as long as Intel complied with the 10-ton limit. The stakeholder team considered this issue during FPA development and Intel commits to demonstrating that emissions from any single chemical of concern does not exceed the AAAQGs.
As a legally-binding component of its air permit, before using any chemical that might pose a risk to human health or the environment in its processes, Intel must perform screen modeling at 10-tons/year levels to demonstrate that AAAQGs are not exceeded. If the model demonstrates otherwise, Intel will limit emissions of any such chemical to ensure the AAAQGs will not be exceeded (as it has already committed to do for phosphine and phosphoric acid). This analysis would estimate a worst case emissions of 10-tons/year for each chemical, would determine whether that case would pose any risk, and if so, would establish limits on that chemical to reduce such risk.
Finally, Intel has committed to establishing emission limits for all chemicals of concern, whether regulated or unregulated. As a legally-enforceable component of its air permit, Intel must apply the screen modeling analysis to any compound it uses that is listed among the 189 HAPs currently listed under federal law. Moreover, as a voluntary commitment in the FPA, Intel will perform the same analysis for any of the 273 other compounds listed in the AAAQGs. As an additional voluntary commitment in the FPA, Intel commits to, in concert with the stakeholders, to determine whether any chemical it uses present any concern to human health and the environment, and if so, to establish emission limits for such chemicals.
3. Conservative factors minimize the risk that hourly and daily AAAQGs will be exceeded.
The screening model used to determine consistency with the AAAQGs assumes uniform, constant emissions of 0.288 grams of HAP/second (i.e. 10 tons averaged over a year) for most HAPs (for reasons that will be explained in the next section, the model assumes lower emission rates for phosphine and sulfuric acid). Commentors have questioned whether this assumption accounts for production spikes or peaks of HAP emissions that may exceed the 1-hour and 24-hour AAAQGs.
A number of other factors reflect the conservative nature of the screening model:
· the 0.288 grams/second emission rate for non-phosphine HAPs was ultimately derived from the 10 tons/year maximum emission rate established by the Clean Air Act for any single HAP; this rate is three times greater than the expected actual emissions based on engineering data from other Intel facilities;
· in Intel's proposed air permit, the 10 tons/year limit is an aggregate limit set for numerous HAPs; it is unlikely to assume that any single HAP will ever be emitted at this level-and preclude the emission of any other HAP-if Intel is to continue manufacturing semiconductors at the Ocotillo site;
· the screening model assumes that all emissions are from a single stack closest to the Ocotillo site's fence-line, resulting in over-estimations of fence-line concentrations; and
· based on the operating history of other Intel manufacturing facilities, the manufacturing process at the Ocotillo site is designed in a way that generates fairly constant emissions. It is intolerant of extreme variations between average and peak values.
4. The Intel XL project establishes HAP limits that are more protective of the health of employees than they would be under actual and required performance. The Intel XL project establishes HAP limits that are more protective of the health of employees than they would be under actual and required performance.
Commentors have questioned whether the HAP limits established in Intel's permits are sufficiently protective of the employee's health and safety. Employee are protected since Intel goes beyond all applicable OSHA requirements; legally enforceable commitments are included in the air permit not to exceed the 24-hour AAAQGs limits at the Ocotillo site fenceline; and Intel makes voluntary commitments in the FPA not to exceed the 1-hour AAAQGs limits within the site itself. If this occurs, Intel would either establish lower limits for such compounds or would develop a more sophisticated analysis in collaboration with state and local regulators to determine the risk posed by such compounds.
An example of this strategy is that the 1-hour AAAQGs for sulfuric acid and phosphine were exceeded when a theoretical maximum release of 10 tons per year was modeled. The result is that the permit level for phosphine will be 4 tons per year and 9 tons per year for sulfuric acid.
5. The HAP limits in the air permit and FPA compare favorably to the actual performance of similarly situated facilities.
One factor in assessing the superior environmental benefits of the Intel XL project's limits on HAPs is that the Ocotillo site is a new facility and no historical data exists against which proposed air limits can be easily compared. Moreover, the absence of benchmark data for the entire semiconductor industry makes assessments of actual superior environmental performance problematic. Commentors have cited 1993 Toxics Release Inventory (TRI) data to argue that the 20-tons/year HAP limits established under Intel's air permit exceed the majority of HAP emissions for facilities within the same SIC code. This comparison is imprecise because it compares emissions data for facilities that cover a wide range of size and production capacities.
Some HAP emissions that EPA believes somewhat reflect the industry norm are presented in Table 2 following. For six selected HAPs, EPA obtained TRI emissions data from the semiconductor industry for the years 1992-94. The average total emissions across all six HAPs is derived. Several caveats accompany this analysis: the data are self-reported with no prescribed methodology; the data pertain only to facilities with emissions equal to or greater than 10,000 lbs/year for each HAP and the sample from which average HAP emissions are being calculated may not necessarily be representative of the industry.
Emissions for selected HAP compounds for semiconductor facilities
Toxics Release Inventory Data 1992-94
Total Emissions
(tons/year)
Number of facilities
Emissions per facility (tons/year)
HAP Compounds
1992
93
94
1992
93
94
1992
93
94 Hydrochloric Acid 51 39 48 92 78 88 1 0 1
Hydrogen Fluoride
27
30
24
87
71
79
0
0
0
Glycol Ethers
114
127
102
28
27
22
4
5
5
Xylene(Mixed Isomers)
158
131
81
29
25
21
5
5
4
Ethylene Glycol
3
6
16
10
16
19
0
0
1
Methanol
124
83
97
21
16
14
6
5
7
TOTAL
16
15
18With these caveats in mind, it appears that the Ocotillo site's limit for all HAPs is very close to the industry's average 18 tons/year emissions for the selected six HAPs alone. Thus, Intel is well within, if not exceeding, the standard for the industry and the additional safeguards that Intel is providing through its HAPs modeling and comparison to the AAAQGs certainly provides a superior benefit.
6. The Intel XL project will establish a limit on emissions per production unit for VOCs and HAPs that is more protective of human health and the environment than would be under required or expected future performance. The Intel XL project will establish a limit on emissions per production unit for VOCs and HAPs that is more protective of human health and the environment than would be under required or expected future performance.
As a facility that has only recently started production, aggregate HAP emissions at the Ocotillo site may increase up to the PSEL limit as production increases. While the stakeholder team was sympathetic to Intel's need to expand production at its newly constructed Ocotillo site, the team wanted some assurance that increases in either VOC or HAP emissions would be solely due to increased production, rather than decreased environmental control.
As a voluntary commitment under the FPA, Intel committed to developing a production based performance standard by the end of this year. This standard would establish voluntary limits on the amount of VOCs and aggregate HAPs emitted per standardized unit of production. There are no federal, state, or local requirements for companies to commit to limit emissions per unit of production. Intel's XL commitments thus provide the public a measure of pollution prevention that goes beyond what Intel would have had to do outside of XL.
ISSUE #3: There is no independent method for measuring/verifying the environmental impacts.
Commentors: MG, SN, CRT, Cole, TS
Discussion: Intel controls its organic air emissions with an activated carbon unit and its inorganic emissions by wet scrubbers. Emissions of inorganic HAPs will be computed through the use of emission factors, developed through analytical testing of process tools. The methods for verifying the amount of air emissions pursuant to the air permit are consistent with the method adopted by Maricopa County.
Intel has asserted a confidential business information (CBI) claim for the tool specific emission factors with Maricopa County. It is the responsibility of Maricopa County to deem this information as CBI. However, even information deemed to be CBI has been reviewed and found to be accurate by the appropriate regulatory authorities.
For VOCs and organic HAPs, Intel will calculate air emissions by subtracting the amount used form the amount brought on site, otherwise known as mass balancing. This is a conservative method of calculating emissions
The independent verification of air emissions is still of concern to the stakeholder group and is slated for continued discussion at the semi-annual FPA public meetings that will also include the Community Advisory Panel. To the best of our knowledge, no technology currently exists that would provide such monitoring for all these specific concentrations of HAPs. If cost effective and reliable proven analytical methodology is presented to the CAP that has been demonstrated in the semiconductor industry, the FPA stakeholders will discuss its application to monitoring the FAB-12 site. Intel is also working with EPA to modify the analytical methods reliably measure air emissions from the FAB-12 scrubbers when the inlet concentrations are very low. Until this is done, we cannot independently monitor the actual emissions.
The need for independent, third party verification for the other environmental media emissions such as water will also be on the agenda for the CAP. At this point, the stakeholder team did not see the need for this level of independent verification for the voluntary commitments that are not subject to permit requirements in the air permit and the industrial pretreatment permit. However, Intel has committed to exploring this issue with the Community Advisory Panel. EPA will also be evaluating third-party review of environmental compliance as part of the Environmental Leadership Program.
As a final safeguard, each of the permitting agencies has retained its authority to conduct inspections of the facility to determine adherence with environmental statutes and above mentioned permits that do have enforceable emission levels for air and water pollutant levels.
Resolution: The CAP will be a forum to review any new information that comes in regarding this issue. In addition, EPA will provide any information that becomes available nationally for monitoring the semiconductor industry. EPA will also be evaluating and reporting its findings on third-party review of environmental compliance as part of the Environmental Leadership Program.
ISSUE #4: Claims that the Arizona Ambient Air Quality Guidelines do not reflect chronic health information.
Commentor: Cole, CRT
Discussion: As stated in CRT's September 11, 1996 comments (page 15): "Just based on this limited review of inorganic hazardous air pollutants, it appears that the Arizona guidelines for short term exposures are generally reasonable. The chronic exposure guidelines are more of a problem, however, and seem to be missing for many chemicals of interest".
In discussions with Arizona Department of Health Services (ADHS), EPA found that there are three sets of AAAQGs: 1-hour; 24-hour; and annual levels. The annual level does contain carcinogenic information by basing the level on a one-in-a-million risk of contracting cancer if exposed for 70 years. The 24-hour levels are the lower of either: 1) the OSHA exposure standard, or 2) 365 times the annual level. The second level is set in this manner to assume that a yearly exposure to a carcinogenic level is obtained in one day. The 1-hour level is set by the OSHA standard.
ADHS acknowledged that these levels should be updated to include EPA's risk standards know as the Preliminary Risk Guidelines (PRGs) that are based on the most recent acute and chronic information. Changing the AAAQGs would require state rulemaking and ADHS is not sure when that will take place.
Resolution: As more current health exposure information is available, it should be forwarded to ADEQ who will then ensure that the information is sent to ADHS toxicologists to include the information to reevaluate the AAAQGs.
ISSUE #5: The provisions for water conservation and re-use are inadequate.
Commentors: Roger Peterson, SN, CRT
Discussion: To our knowledge, the FAB-12 plant is the only semiconductor manufacturing plant in the world to have such an innovative system in place to produce wastewater that meets or does better than drinking water quality standards. Intel and the City of Chandler have won a national award for the design
Intel is committed to reducing water consumption at the Ocotillo site. All water needs for landscaping and cooling towers will be provided by treated water from the City of Chandler Public Owned Treatment Works. Water needs for other on-site processes will be provided by fresh water from Chandler at approximately 1,850 gallons per minute at full plant capacity.
Intel could have designed their water system to discharge all wastewater to the city for treatment at a cost of about $5 million. Instead, Intel has spent approximately $28 million to build a dedicated water treatment plant to be owned and operated by Chandler that will treat the company's effluent via reverse osmosis.
The reverse osmosis plant will allow the treated water to be either recharged into the aquifer to replenish groundwater use or for reuse by the Roosevelt Water Conservation District. Recharging the aquifer will decrease the per capita consumption figures for the City and allow for additional population growth. It should be noted that water recharged into the aquifer will be at least as clean as drinking water standards. This water reuse component is unprecedented in the semiconductor industry and has received national attention as a model for decreasing water use at an industrial facility.
Currently, the Ocotillo site's water recycle rate is determined as the volume of water transferred to Chandler's water treatment plant divided by the volume of water purchased from the City of Chandler. Intel has committed to increasing this recycle rate from 45 percent in 1997 to 65 percent in the year 2001. In addition, Intel is also funding research, possibly through SEMATECH, on why a bacterial growth occurs in the system when treated waste water is reused in the facility.
One issue associated with the wastewater generated by the Intel system and its operations is whether the Resource Conservation and Recovery Act (RCRA) requirements apply to the Intel waste management system as a whole (which includes the City of Chandler's treatment system). EPA has concluded, based on the facts described above, that for purposes of this XL project, the "elementary neutralization unit" exemption to RCRA Part 264 (relating to requirements for owners and operators of RCRA treatment, storage and disposal facilities(TSDFs)), Part 265 (relating to interim status requirements for owners and operators of TSDFs) and Part 270 (relating to RCRA permit requirements) applies to the entire Intel system. (See 40 CFR s264.1(g)(6), 265.1(c)(10), and 270.1(c)(2)(v).
The term "elementary neutralization unit" means a device which "(1) is used for neutralizing wastes that are hazardous only because they exhibit the corrosivity characteristic....; and (2) meets the definition of tank, tank system, container, transport vehicle, or vessel in 260.10 of this chapter. The term "tank system" means a hazardous waste storage or treatment tank and its associated ancillary equipment and containment system. The term "tank" means a stationary device, designed to contain an accumulation of hazardous waste which is constructed primarily of non-earthen materials (e.g., wood, concrete, steel, plastic) which provide structural support. These terms are described in 40 CFR s 260.10.
Intel's neutralization system meets the definition of "elementary neutralization unit" as used in EPA's regulations. As explained in greater detail above, the on-site treatment system is specifically designed to neutralize effluent generated from the process of making ultra pure water. In addition, this effluent is hazardous only because it exhibits the corrosivity characteristic. Further, the neutralization is undertaken in a unit which meets the definition of a hazardous waste treatment "tank and its associated piping, which transports the neutralized effluent to the adjacent City of Chandler treatment system. Finally, the City of Chandler's treatment system is solely dedicated to the Intel facility for Intel's particular waste stream and is itself a part of the tank system as ancillary equipment. Therefore, for purposes of this XL project, the elementary neutralization unit exemption will extend to the entire Intel system (including the City of Chandler's treatment system).
Another issue is whether RCRA Land Disposal Restriction (LDR) regulations apply to the effluent from the Chandler treatment system. As explained above, most of the treated effluent, which meets safe drinking water standards, is used to recharge the local aquifer by reinjection through a Class V well regulated under the Safe Drinking Water Act. However, some of the treated effluent is "rejected" if the salt content is too high to be used for groundwater recharge. Instead, the rejected effluent is sent via enclosed piping to a salt evaporation pond. EPA has concluded that the LDR regulations do not apply to this rejected effluent because LDR regulations do not apply if the effluent is managed in a system that is a "zero discharger" that engages in "Clean Water Act-equivalent treatment before ultimate land disposal...." The term "CWA-equivalent treatment" includes a "treatment technology that can be demonstrated to perform equally or greater than [the technologies specifically listed in section 268.37(a)]." 40 CFR s 268.37. It is EPA's view that the Chandler treatment constitutes a CWA-equivalent treatment prior to the land disposal of the effluent in the salt evaporation pond, because it is a very effective form of wastewater treatment and is as effective as biological treatment, which is expressly listed in the regulations as an acceptable technology. Rejected effluent which is not sent to the salt evaporation pond is sent to the Lone Butte POTW, which is regulated under Section 402 of the CWA. After it is discharged from the Lone Butte POTW, the effluent is no longer regulated under RCRA. 40 CFR s 261.4(a)(2). Similarly, 40 CFR s 268.37(b) also exempts from LDR regulations effluent sent to a Class V injection well if it too is managed in a system that engages in "CWA-equivalent treatment before injection." Therefore, since the effluent is subject to reverse osmosis treatment at the Chandler system prior to its reinjection to groundwater via a Class V well, the LDR regulations do not apply to this effluent either. In sum, once the effluent exits the Chandler treatment system, LDR standards do not apply.
In addition, as noted above, the reinjection of the treated effluent into the aquifer is regulated by the Safe Drinking Water Act. The well through which the treated effluent is injected is regulated as a Class V well under the Act (see 40 CFR s 146.5), for which no permit is required under federal law. 40 CFR s 144.24 (activity authorized by rule). However, even though no permit is required under federal law, states may impose more stringent or protective requirements, and with respect to this facility, the reinjection of the treated effluent to the aquifer requires and has the benefit of an Aquifer Protection Permit under Arizona law.
Finally, as noted above, some of the treated effluent is diverted from reinjection for use for irrigation of local crops. Also as noted above, the RCRA LDR's no longer apply once the effluent exits the Chandler treatment system, and the treated effluent meets the drinking water standards under the Safe Drinking Water Act. Consequently, under all applicable regulations, the treated effluent is as appropriate for irrigation as for groundwater recharge.
Resolution: No changes required in the FPA.
ISSUE #6: There is completely inadequate community and employee participation.
Commentor: SN, CRT, MG, RH
Discussion: The Intel XL Stakeholder team has received a number of presentations on how Intel employees have had opportunities to provide input into the FPA development process through internal working groups, attending the public meetings, or providing comments directly to the stakeholder team. Intel employees briefed the stakeholder team on how they are informed on health and safety conditions at the plant and are active in plant decisions that affect exposure to potential pollutants by internal factsheets and electronic mail as well as safety teams reviewing plant operations.
On the issue of community involvement, four members of the local community participated in the meetings of the FPA working group. There has been a tremendous effort on the part of community members Barbara Knox (resident, local farmer), Jim Lemmon (resident, member of AZ Public Health Association), David Matusow (resident, software engineer), and Doug Ballard (CAP Member). Minutes from all the FPA working group meetings are posted on the Internet. The stakeholder team spent over 2,000 hours in 68 meetings over 11 months working on the FPA. Additionally, six public meetings and a workshop have been held that were announced in local papers, postings, and leaflets that were passed out from door to door. All public meetings were videotaped and are available to the public. Maricopa County held a public hearing on the air permit conditions.
There has been repeated requests asked for other residents to step forward to serve on the stakeholder team for this project. To date, no one else has volunteered. This lack of involvement is not because of a flawed public participation process, but the overwhelming effort required to learn about the issues and participate in the discussion.
Resolution: The CAP, in conjunction with the Intel XL stakeholder team, will have to grapple with this issue as the FPA calls for semi-annual meetings to review the progress of the elements within the FPA and will require continued involvement by the public.
There were also requests by a member of the stakeholder team for direct funding to obtain independent technical assistance. Although not provided during the development stage, EPA will provide direct funding for technical assistance to local citizens members as a pilot so that they can develop the knowledge they need to effectively evaluate the implementation of this project
s. EPA is currently determining the best mechanism to provide this funding. EPA is also examining approaches for use of public and private funding sources to provide resources to community groups involved with XL projects.
The Administrative Record will contain a more detailed description of the various opportunities described above that Intel uses to communicate with employees regarding implementation of the FPA.
ISSUE #7: There is inadequate pollution prevention built into the FPA.
Commentor: SL
Discussion: The FPA includes company commitments to reducing solid and hazardous waste, reducing the use of fresh-water, reducing employee commuting, and establishing property set-backs that surpass current requirements. The establishment of a VOC and HAP limits on a per unit of production basis, which Intel would commit to in the Final Project Agreement (FPA), is a key driver for toxics reduction and pollution prevention. Such a limit ensures that the pollution efficiency ratio does not increase as a result of operational modifications or expanded production.
As part of its corporate philosophy of Designing for the Environment, Intel has established a cross-corporation council to ensure that Intel's facilities, products and processes are designed with improvements in chemical, water and energy use for each generation of production technology. Since Intel is bringing new processes on line every 2 years, pollution prevention is built into the design of the process and not considered as an afterthought for an existing process. By using this approach, Intel has reduced the VOC emissions in the new process at the FAB-12 unit by 40% compared to the previous process generation. In addition, Intel has reduced hazardous waste generation by 50% since 1985.
In addition, Intel is working with ADEQ to determine that the FPA satisfy the State's pollution prevention planning and reporting requirements.
Resolution: No changes required in the FPA.
ISSUE #8: There is inadequate technical/financial assistance.
Commentors: CRT, SL, SN, Cole, MG, DWA
Discussion: A member of the stakeholder team made repeated requests for independent funding for technical assistance. The source of this funding should have been either EPA or from Intel. At the time, requests were forwarded to the national XL program office in EPA Headquarters to determine the if this type of funding was possible. During deliberations with the EPA Headquarters, regional EPA staff, EPA ORD staff provided technical assistance on air regulations and air modeling questions. In addition, Intel brought in John Gardner from the City of Chandler to assist on development of a single emergency response plan. Intel was willing to fund a technical assistant if either working group expressed interest.
Response: As mentioned in Issue # 6, although not provided during the development stage, EPA acknowledges that community members are at a disadvantage by not working directly in the environmental field and could benefit from resources to provide independent technical assistance on various issues. EPA will provide direct funding for technical assistance to local citizens members as a pilot so that they can develop the knowledge they need to effectively evaluate the implementation of this project
s. EPA is currently determining the best mechanism to provide this funding. EPA is also examining approaches for use of public and private funding sources to provide resources to community groups involved with XL projects.
ISSUE #9: The transferability of this FPA should be done carefully.
Commentor: DWA, RH
Discussion: The air permit conditions are experimental. The permit conditions have been crafted as part of the Environmental Protection Agency's Excellence in Leadership Project that stresses the experimental nature of such projects. These types of conditions, therefore, should not be considered immediately transferable to other permit applicants until all parties have had an adequate opportunity to evaluate if the goals of the Project are being achieved. At that time, which may be several years, you may be able to revise the County's rules and regulations to reflect the successful aspects of the revised Air Quality Permit in a way that is transferable to other facilities within the same industrial segment, or possibly even to other industries.
In addition, the air permit conditions were recommended to Maricopa County following a lengthy stakeholder process in which the interestsof Chandler residents and non-governmental organizations were represented. The flexibility in the Air Quality Permit conditions would not have been considered to be sufficient without the additional substantial efforts that Intel is undertaking to achieve environmental benefits. Intel has agreed to stringent permit conditions for modeling public health impacts at its property boundary and for reporting to the public on its air emissions. In addition to these permit requirements, Intel has set forth a number of extremely ambitious environmental goals that it will seek to achieve over the next five years.
While these additional goals are voluntary, Intel is accountable to the public for achieving or failing in these goals. These goals were also established through the public stakeholder process, and therefore, reflect the priorities of the residents who are directly affected by the FAB-12 site. For the Intel FAB-12 FPA, the public participation process was enforceable under the air permit.
Other permit applicants may request similar conditions allowing flexibility that are similar to those in the revised Air Quality Permit for Intel's FAB-12 site. EPA would strongly object to the inclusion of such conditions until two prerequisites are satisfied: 1) there has been an opportunity to evaluate the effectiveness of all of the air permit conditions to achieve superior environmental performance as anticipated; and 2) the conditions that an applicant is requesting have been subject to the extensive stakeholder participation that occurred with Intel's Project XL. Lastly, the applicant should have a clear understanding of what aspects of their project will be enforceable and which will be voluntary commitments.
Resolution: Given the screening process for Project XL proponents, only "good actors" will be granted approval for any project. For the Intel site, if the Ocotillo site is sold to another company, EPA and the state would add the new owner to the FPA if the agencies were satisfied that the new company met the criteria of being a good environmental actor. This determination would be based on past environmental performance and review of all enforcement dockets.
ISSUE #10: There are inadequate provisions to protect employee safety.
Commentors: SL, Cole, SN, CRT, NRDC
Discussion:
Intel currently goes beyond OSHA regulations for protecting workers from chemical exposure. OSHA standards for 8-hour exposure limits are set in regulations that call for Permissible Exposures Limits (PELs). Intel actually uses the more stringent limits set by the
American Conference of Governmental Industrial Hygienists. These limits are called the Threshold Limit Values (TLVs) and are also based on an 8-hour exposure that are more stringent than the PELs.
The XL process has provided additional employee protection by voluntarily committing in the FPA to set emissions so that the 1-hour AAAQGs limits within the site itself is not exceeded. If the screen modeling described in the previous section indicate that the 1-hour AAAQGs for any compound will be exceeded within the Ocotillo site, Intel would either establish lower limits for such compounds or would develop a more sophisticated analysis in collaboration with state and local regulators to determine the risk posed by such compounds. In its proposed permit, Intel is establishing limits for sulfuric acid, and phosphine to ensure that these emissions do not exceed the 1-hour AAAQGs.
The 1-hour AAAQG is protective based on an average annual air emission level. There may be emergency upsets that will result in air emission levels that exceed the 1-hour level. This emergency situation is planned for in the Hazardous Materials and Management Plan.
In addition, Intel employees have formed safety teams that are composed of employees within a specific task area. These teams evaluate all aspects of their production area to ensure that their safety is not compromised. The Intel XL stakeholder group was briefed by two such safety teams.
OSHA does have an excellence program called the STAR or Voluntary Protection Program (VPP). This program is very rigorous and involves the facility applying for inclusion in the program by justifying its excellent track record and systems that are in place to go beyond OSHA standards for employee health and safety concerns. Intel is evaluating this program to determine if their facilities will participate. Information on the OSHA VPP program was sent to the stakeholders and interested parties who commented on the FPA.
Resolution: A more complete description of the Intel employee safety teams will be added to the Administrative Record. Participation in the VPP program is being evaluated by Intel and will be reported to the CAP.
In addition, Intel's DFE program will evaluated any new process change to decrease the emissions from individual tools to protect employee health and safety. Employees do have input into this process through employee safety committees and the SCC.
ISSUE #11: Intel should not claim the emission factors as Confidential Business Information
Commentor: SL, CRT, NM
Discussion: A number of commentors felt that Intel should not be claiming CBI for certain aspects of the FPA, specifically the tool specific emission factors. They believed that the process should be as transparent as possible and not allow any information to be withheld from the public.
Resolution: Intel has stated that the information that is embodied in the tool specific emission factors is the "recipe" for their process and would provide an undue competitive advantage to other semiconductor manufacturers. Intel can only claim CBI on those items of information that satisfy Maricopa County's CBI criteria. Intel has asserted the CBI claim, Maricopa County will determine if this claim is valid and deem then information to be CBI. If the information is deemed to be CBI, it will not be available for public review.
ISSUE #12: The emission per production unit is not enforceable, verifiable and is of limited use
Commentor: TS
Discussion: In addition to plant site emission limits (PSELs), Intel has committed to determining an emissions per production unit. The air permit for the Ocotillo Site does not restrict Intel to its current level of production activity and provides pre-approval for changes and construction that may increase production, provided Intel's air emissions remain below the PSELs and all other permit conditions. Under the FPA, however, Intel commits that it will not increase the level of its emissions beyond the levels proportional to increases in production activities.
To demonstrate this commitment, Intel agrees to develop a production based performance standard by January 1, 1997. The production based performance standard will measure pounds of total VOCs and aggregate HAPs emitted per standardized unit of production. This measurement will allow the public to verify that any increase in VOC and HAP emissions at the Site is the result of increase production activities and not a consequence of decreased environmental control. Emissions will be reviewed with the CAP on a quarterly basis.
See Issue #13 for discussion on enforceability and Issue #3 discussion on verification.
Resolution: The commitment to include an emissions per production unit is in the FPA and will limit the emissions from FAB-12, especially given the situation where there is no increase in production.
ISSUE #13: All aspects of the FPA should be enforceable.
Commentors: SL, NRDC, CRT
Discussion: The aspects of the FPA that are enforceable are the attached air and water permits that include public participation and reporting. All other elements of the FPA, including the recycling commitments and the emissions per production unit, are voluntary. However, in the event that the FPA is not effective, Intel is subject to all current regulatory requirements.
The FPA, beyond those provisions contained in the attached air and water permits, cannot be made enforceable under EPA's authorities. However, EPA believes that there are strong incentives for Intel to meet the voluntary commitments contained in the FPA. First, there is the potential for the regulatory agencies to return Intel to otherwise applicable requirements if the FPA is not effective. Second, progress towards these voluntary commitments will be widely reported to the stakeholders. publicly in Chandler, and on the Internet.
Intel has provided a great deal of operating information to the stakeholder group, except for that which is protected under state confidential business information statutes. Representatives from federal, state, and local regulatory agencies can at any time inspect the facility and records, check calculations to determine whether the emission factors and reported emissions are correct.
Resolution: As a result of recent discussions with the stakeholder team, an agreement was reached that the public participation aspects and consolidated reporting requirements would be included in the enforceable air permit.
The FPA is a statement of all of the parties' commitment, seriously undertaken, to proceed with this project. Legal enforceability is not the sole measure of the seriousness of the commitment.