![[logo] US EPA](../gif/logo_epaseal.gif){kind=logo}
Intel
Henry S. Cole and Associates' Comments
I. Introduction, EPA, the state of Arizona, Maricopia County Intel Company are in the process of finalizing an agreement under EPA's XL program that will have a large impact on community members in Chandler, Arizona and that will set an important precedent for EPA's XL program. The XL program allows companies to have an unprecedented degree of flexibility with regard to permitting. However, the agreements are supposed to lead to public benefits -- including superior environmental performance, pollution prevention, public accountability and stake holder involvement.
The proposed Final Project Agreement (FPA), while giving the company a great deal of flexibility, falls short of the mark with regard to environmental improvement and public accountability. For example, the agreement, as it now stands:
allows significant increases in emissions of hazardous air pollutants and other toxic contaminants to the air,
may result in exceedances of Arizona's Ambient Air Quality Guidelines for phosphine,
encourages the company to use poorly investigated chemicals which are not currently regulated as Hazardous Air Pollutants (HAPs) in order to avoid limits which apply only to HAPs.
Moreover, the company's ability to change process and emissions without chemical specific approval under the FPA, will make it difficult for the public and government agencies to ensure that the company is actually complying with the terms of the agreement and meeting Arizona Air Quality Guidelines for hazardous air pollutants in the neighborhoods surrounding the facility.
Furthermore, Intel and EPA have developed the FPA without public stakeholder process needed to establish public accountability and to respond to serious public concerns. The flexibility that the FPA provides to Intel, introduces many serious questions and potential problems that must be resolved to build public confidence not only in the precedent-setting Intel Agreement, but in Project XL as a whole. This set of comments includes specific recommendations for a stakeholder process. For example, technical assistance funding should be provided directly in order to allow a coalition of environmental and community stakeholders to hire experts of their own choosing. It is clear that to implement a program of this type, EPA and Intel will need to modify the current timetable and recognize both the needs and benefits of an expanded stakeholder involvement process.
II. The Agreement allows a substantial increase in plant emissions
The FPA would allow large increases in emissions of toxic chemicals over currently permitted levels. These chemicals include hazardous air pollutants which can result in risk to community members and workers even at very low levels of exposure. The increases can be clearly seen in Figure 1 through 3.
A. Phosphine. Figure 1 compares the emissions allowed by the current permit and by the proposed FPA for phosphine, and extremely toxic hazardous air pollutant (HAP). Phosphine has the lowest (strictest) 1 and 24 hour AAAQG (Arizona Ambient Air Quality Guidelines) of routine emissions (See Table 1 from the FPA proposal). The FPA would allow annual emissions to increase by a factor of 400, from 25 lbs to 10,000 lbs. The proposed increase in emissions is highly significant from the standpoint of public health. Table 1 compares the results of air quality modeling conducted by the Maricopia County Environmental Services Division (MCESD) with the AAAQGs for routine emissions from the plant. Notice that the estimate for the maximum 24 hour concentration is 3 ug/m3 which represent 99.4 percent of the AAAQG for 24 hours. Moreover, although, the model used, EPA's SCREEN 3 is designed to be conservative (tends to over estimate concentrations in order to be protective), the assumption of uniform emissions used may actually result in an underestimation of maximum short-term concentrations. (See section on modeling below.) In short, the 99.4 percent estimate is to close for comfort, especially when one considers that residents will be exposed to a number of toxic chemicals from the plant.
B. Total Volatile Organics. Figure 2 compares total organic compounds (VOCs) emissions allowed under the existing permit and the proposed FPA. The FPA would allow an additional 15 tons of VOCs to be emitted per year.
C. Hazardous Air Pollutants. It is not possible to directly compare emissions for specific hazardous air pollutants (HAPs) allowed under existing permit and proposed FPA. The FPA specifies a combined limit of 10 tons per year for all organic HAPs and 10 tons per year for all inorganic HAPs. There is no way to predict in advance the quantities of each chemical that the company will emit during a year.
The following analysis demonstrates the increase in the inorganic HAP emissions that could be allowable by the proposed FPA. Under this agreement, the facility could in theory consume the entire limit for inorganic HAPs by emitting 10 tons of a single chemical, e.g. hydrogen fluoride. This would allow HF emissions to increase from the 500 lbs under the current permit to 20,000 pounds under the FPA (a 40 fold increase).
It is more likely that the facility would emit a number of inorangic HAPs regulated under the existing permit (the existing permit lists 8 inorganic chemicals) and includes phosphine (which is separately limited at 5 tons per year). In order to estimate allowable emissions of inorganic HAPs under the FPA, I have assumed that the plant would emit 5 tons of phosphine and that the remaining 5 tons is divided equally among the following 7 inorganic chemicals for which specific limits exist under the existing permit (diborane, boron trifluoride, boron trichloride, hydrogen chloride, hydrogen fluoride, chlorine and silane).
The results of the analysis for this scenario are shown in Figure 3. For highly toxic hydrogen fluoride, the proposed FPA would allow nearly a three fold increase from 500 lbs per year to more than 1400 lbs per year. For boron trifluoride, the increase is 14-fold. Similarly, diborane emissions increase by a factor of 28.
In summary, the evidence presented in Figure 1 through 3 indicates that the FPA will encourage increased emissions rather than pollution prevention.
III. Protection of Public Health - Not assured by the Air Permit
A. Increased Risks. The proposed FPA uses a screening model (EPA's SCREEN 3) generally assumed to be conservative (tends to over predict). Table 1 presents the results of the SCREEN 3 modeling. The object of the analysis is to obtain 1 hour and 24 hour maximum concentration estimates that can be compared to the 1 hour and 24 hour AAAQGs. (There are no annual AAAQGs.) Since the maximum allowable annual emissions of specific HAPs will not be known (with the exception of phosphine's 5 ton/year limit), the screening uses an annual emissions rate of 10 tons for each HAP other than phosphine. This approach is conservative on a chemical-by-chemical basis since no chemical is likely to reach the 10 ton per year limit (for both inorganic and organic HAPs). However, there are many different ways to meet the HAP limits, some for riskier than others. Moreover, there is no provision to limit total risk imposed across all chemicals, nor the potential synergistic effects associated with exposure to multiple chemicals. As discussed above, it is quite possible that the new system will allow significant increases in exposure to certain chemicals and to the total risks imposed on area residents.
B. Phosphine. In the special case of phosphine the modelling result is not reassuring. The maximum allowable rate for the highly toxic chemical is 5 tons per year. This limit was selected on the basis of SCREEN 3 modeling such that the maximum fencaline concentration estimate just meets the 24 hour AAAQG. Intel used this method rather than conduct refined modeling. Given the approach and the narrow margin between the AAAQG and maximum concentration estimate, is especially important to scrutinize the methods and assumptions. The following section discusses the impact of variable emissions on short term concentrations.
C. Variable Emissions and short term concentrations. An important assumption that is not explicitly discussed in FPA analysis, missions, is that emissions are emitted uniformly throughout the year. The screening analysis makes this assumption without providing documentation on its validity. The analysis fails to consider the case where emissions of one or more chemicals vary substantially. There is nothing in the FPA air permit that would prevent the plant from emitting its annual limit on a third, a half or three quarters of the hours in a year. For this reason, the model may fail to predict maximum 1 hour and 24 hour concentrations resulting from the combination of higher than average emission rates with atmospheric conditions that maximize concentrations.
The effect of variable emissions is shown in the following example. Assume the company emits the annual emission limit for phosphine (5 tons per year). If the emissions occur evenly over the entire year the uniform emission rate is:
(10,000 lbs/year) / 8760 = 1.142 lbs per year or 0.144 grams/scc.
The same rate assumed by Maricopia County Environmental Services (MCESD).
However, if the process emitting phosphine operates on only 3/4 of the hours (i.e. 6570 active hours), the emission rate for operating hours would be:
(10,000 lbs/year) / 6570 hours/year - 1.52207 lbs/hour or 0.182 grams/scc.
For non-operating hours the rate would be zero.
MCESD calculated a maximum fenceline 1 hour phosphine concentration of 7.45 ug/m3 (at a distance of 479 meters from the plant.). In order to estimate the impact of the 3/4 active time scenario, assume that: (1) the annual emission rate is the same, (2) the emission rate for active hours is 0.192 g/sec, (3) meteorological conditions associated with the highest concentration estimates occur at least once during the 3/4 of the year that the process is active. To calculate the maximum 1 hour concentration estimate multiply MCESD's estimate for uniform emissions by the proportional increase in emissions:
0.192/0.144 x (7.45) = 9.93 ug/m3
Using MCESD's method to adjust the 1 hour estimate for maximum concentration to a 24 hour estimate:
9.93 x 0.4 = 3.97 ug/m3
The 3.97 ug/m3 estimate exceeds the 24 hour AAAQG.
Similarly, if the process operated only 1/2 of the time (with the same annual level), the operating emission rate would double, as would the maximum concentration. Thus for the half-time operation scenario, the maximum estimated phosphine concentration estimate is 14.9 ug/m3, which exceeds the 1 hour AAAQG for phosphine of 11.4 ug/m3.
In reality, the plant may not operate at all hours or may operate with variable emissions over the course of the year, the maximum 1 hour and 24 hour concentrations will reflect the combination of meteorological conditions actually occurring at any point in time. However, the analyses used to set the 5 ton per year limit for phosphine and to ensure that other contaminants will meet the AAAQGs doesn't include the effect of emissions variability. The flexibility which the FPA allows in both process and emissions suggest that emissions could be quite variable.
A related issue which has not been considered in the modeling analysis is a highly elevated short term emission rate resulting from an accidental release or process upset. It is quite possible that the maximum average hourly rate (annual limit divided by 8760 hours per year) could be exceeded by many times over under this type of circumstance.
In order to assure that short term AAAQGs can be met, the FPA be amended to include the following measures:
1. The company should be required to meet an hourly emission rate for each HAP sufficiently low to ensure that fenceline concentrations resulting from the least favorable dispersion events are below the 1 hour and 24 hour AAAQGs. The company can submit such analyses as a planning procedure in advance of actual use to allow for flexibility. Such submittals should be subject to agency and public review.
2. For any process or chemical input change the company should be required to submit (for agency and public review) documentation demonstrating that the process would not cause emissions whose average or maximum rate would cause the exceedance of a AAAQG. Moreover, for chemicals for which ambient guidelines or criteria are not available, the company should be required to base maximum average and short term emissions rates based on risk assessment. The company can submit such analyses as a planning procedure in advance of actual use to allow for flexibility.
3. The company should be required to collect and provide verifiable information tracking processes and chemical use on a daily basis.
4. The company should be required to provide evidence in the form of mass balance information and stack monitoring demonstrating that emission rates meet the limits. Mass balance analysis (for emissions which are ingredients) can utilize daily chemical input rates, scubber removal rates, etc. and chemical-specific stack testing.
IV. Stakeholder Process and Technical Assistance
While EPA solicited comments and concerns of public stakeholders, community involvement needs to be strengthened substantially. The opportunity for technical assistance was provided to the working groups, but not to any group representing environmental or community stakeholders. Public stakeholders have little access to the kind of expertise available to both industry and government participants (e.g. scientists, and engineers). Without having direct access to consultants of their own choosing, public stakeholders will have difficulty evaluating and trusting claims made by industry. This type of technical expertise is necessary not only in the development of the agreement but to lend public accountability and credibility to its implementation. The following measures are recommended for (1) revision and finalization of the FPA to incorporate remaining concerns, (2) further protocol development, (3) tracking of Implementation:
1. Draft proposals are provided to environmental/community stakeholders in a timely manner for review and comment. For each significant proposal a meeting or phone conference is held to question the draft proposal, to present comments and to offer counter proposals.
2. Stakeholders are given the opportunity to provide written comments following this meeting.
3. Parties to the agreement should respond in writing to all comments.
4. Funding for technical assistance should be provided to a group of environmental and community stakeholders constituted for the purpose of participating in environmental negotiations on the Ocotillo Plant. This model respects the need for community and environmental stakeholders to independent technical assistance. In order to even the playing field, sufficient funding should be provided to allow community-hired experts to play a significant role in the process.
This model is followed in EPA's successful Superfund Technical Assistance Grant Program. In other cases, PRPs have funded community organizations to hire technical expertise. Examples are Ciba-Geigy, Toms River New Jersey and the Hilton Davis Site in suburban Cincinnati. In the latter, Kodak, one of the major PRPs and the city of Cincinnati are providing funds that have allowed a community organization to hire technical expertise. In all of these cases, the community selected its technical expertise independently of PRPs government agencies and the funding source.
As a consultant who has been hired by many community groups and via TAG funding and PRP funding, it has been my experience that enabling community members to hire their own expertise helps to sharpen the technical discussion in ways that are clarifying to all parties. The process does not always lead to consensus, but is more likely to result in the more rigorous review of proposals, the presentation of viable alternatives, and greater public accountability (one of the goals of XL).
In summary, if EPA wishes to replace command and control with greater public accountability it needs to strengthen the ability of public stakeholders to participate directly in the negotiations that lead to agreements, both by opening the process and providing technical assistance funding.
Go to EPA Response