Volume 2, Number 9
A Gossman Consulting, Inc. Publication
September 1996

Should You Test for HAP Emissions?

David Constans

The proposed HWC Regulation is a major use of the Clean Air Act's "Maximum Achievable Control Technology" (MACT) concept of emission limitation. In its briefest form, the CAA requires facilities within a category to meet the emission rates of the best performing 12 percent of the existing sources within that category.

The EPA has broadly interpreted a number of the provisions of this act and has melded these interpretations together with existing RCRA regulations to result in the proposed HWC regulation. The heart of the proposed HWC regulation, the emission standards, are based on the data that produced the MACT standard, i.e. the "best performing 12 percent of existing sources (for which the Administrator has emissions information)" (section (d)(3)(A) of the CAA).

Note that the Administrator is not required to seek out or even request data from the sources. Consequently, emissions data that was generated for a number of other reasons and already possessed by the EPA has been utilized in determining the MACT standards in the proposed HWC. In fact, this emission data has been used to justify the specification of hazardous waste combustors as "major sources of HAP emissions."

So in effect, data generated to demonstrate compliance to State or Federal emissions regulations or to comply with negotiated permit provisions, or accumulated as part of an EPA research project, etc. has been used to set MACT standards. Some of this data is genuinely suitable for establishing these MACT standards, some is most definitely not suitable. In addition, the use of some of this data as a basis for establishing the specification of hazardous waste combustors as "major sources" of individual HAPs is inappropriate. In many cases, especially data generated during trial burn tests, BIF certification of compliance tests and similar "operating envelope" performance tests, these emission data are representative of extreme operating conditions frequently with air pollution control devices operating at minimum performance levels. Clearly, such emissions data is inappropriate for establishing MACT emissions limits. Much more importantly, such data should not be used to establish the category as a "major source" of HAPs emissions. As an example, it is likely that a realistic accounting of dioxin emissions based on "normal" operation would not support EPA's contention that cement kilns are major sources of dioxin emissions.

In Volume III of the Technical Support Document for HWC MACT Standards, the EPA presents the data used to generate the MACT floor levels. Even a brief perusal of this data would cause one to be concerned about the appropriateness of the information. In establishing the dioxin emission limits for incinerators, the EPA relied on data from 24 of the 190 incineration units currently operating. Some of these units combusted PCB laden oils, another had consumed waste spiked with "high levels" of chlorinated POHC's such as carbontetrachloride and hexachlorbenzene, and were admittedly "trial burn" tests. For cement kilns, data from 27 of approximately 40 kilns using waste fuels was used in EPA's MACT floor determinations. As stated by the EPA: "the vast majority of this database is comprised of compliance test emission data generated as a result of the BIF requirements" (FR17390). For lightweight

aggregate kilns (LWAKs), the MACT database contained dioxin emissions data for only one facility of the approximately 15 LWAKs combusting waste fuel. For the purpose of establishing a MACT emissions standard, the data from the one LWAK and the data from the cement kilns were pooled together. In addition, the dioxin emissions database contains data presented in a number of different conventions, e.g. some of the data contains detection limit and EMPC values in the reported TEQ values, making the data non-comparable.

Even the database for particulate emissions from hazardous waste combustors was less comprehensive than one would expect. For incinerators, the EPA MACT database for particulate emissions includes 73 incinerators. For cement kilns and LWAKs, 34 and 12 units respectively. The mercury, semi-volatile metals (SVM), the low volatile metals (LVM), and total chlorine emissions databases used by the EPA to establish the MACT standards are as follows:

Incinerator Cement Kiln Lightweight Aggregate Kiln

Number of Units in database Number of Units in database Number of Units in database
Hg 29 25 10
SVM 42 34 10
LVM 41 34 10
Total Cl 59 33 10

As noted previously, the database for the BIF units is based almost entirely on compliance tests required by the BIF regulations and, as such, represent the limits of the operating envelope rather than normal routine operations.

In some cases the emissions data does not conform to the proper QA/QC standards for the sampling and analysis of the targeted HAP component. Such QA/QC standards may not have been required by the test, or the documentation was never generated as a cost saving measure, or no review of the QA/QC documentation was performed at the time and is now found to be inadequate and beyond redemption due to the time delay. In any event, the quality of such data becomes suspect.

CKRC's comments on the proposed HWC regulation included an analysis by Rigo & Rigo of the data the EPA had used in calculating the MACT standards. Their analysis of this data indicates that significant portions of the data were "made up" by the EPA statisticians. This included data for the best performing 12% of the existing sources in some catagories. Clearly, an egregious example of the abrogation of scientific ethics.

In observing the EPA's use and misuse of emissions data in making MACT floor determinations, it becomes grossly obvious that HAP emissions sources must conduct quality stack emissions sampling and analysis. For facilities that will fall under the proposed HWC regulation these analyses are necessary to determine the impact of the regulation on their facility and to build a database of appropriate information to enable informed responses to EPA's modifications or extensions of this proposed rule. For facilities that will fall under future MACT rules, these analyses are necessary to provide a database that is representative of their category, data that has been collected under the proper QA/QC procedures and is indicative of normal routine operations; not the "operating limit" conditions frequently required by permit authorities. Possessing quality HAPs data in advance of EPA's publishing a proposed MACT standard will allow a facility to make informed comments to the EPA before the regulatory comment period expires and the rule is promulgated.

GCI, in conjunction with Clean Air Engineering, based on our combined experience in performing HAPs emissions trial burn and BIF Certification of Compliance testing have formulated an emissions sampling and analysis plan that conforms to a test protocol currently accepted by the EPA. This is not a "standard" stack emissions test. It contains features that ensure the quality of the data, including a materials mass balance for the targeted metals and chlorine.

For processes utilizing combustion the minimum requirements for this testing are as follows:

1) Develop a test plan and a quality assurance project plan (QAPP) that meets the protocol accepted by the EPA.

2) Execute a HAPS stack test that includes three runs each for:

The "HCl" emissions testing should include ammonium analysis for determination of true HCl and should also include cation determination to check that particulate filter breakthrough has not occurred. Such a test can be performed in one day if properly planned and executed.

3) Process sampling and analysis should include collection of samples during each of the three runs for each input and output stream for the targeted metals and chlorine. Process samples should be analyzed for metals using a total digestion method that provides for true total metals results and more certain closure on the mass balance.

4) HAPs test oversight and supervision should include: Control room interaction/coordination; the collection of operational process data such as production rates, temperatures etc.; the proper training and oversight of process sampling including sample compositing, shipping, chain of custody; and a thorough oversight of the QA/QC elements of the sampling and analysis.

5) The HAPs test report should include the operational process information, process sampling location/methods and analyses, the stack emission sampling locations/methods and analyses and a system metals and chlorine mass balance.

Ultimately, every industrial process with significant HAP emissions will be required to comply to a MACT standard. Will your facility be fairly represented? Will the database that the EPA uses to determine the MACT floor levels be representative of the category? Without quality HAPs emissions data as outlined above, the answer to these questions will be unequivocally negative. EPA's past performance in data acquisition and management, as demonstrated by the proposed HWC MACT documentation, virtually ensures the MACT floor levels will poorly represent the category unless industry makes a concerted and continuous effort to supply appropriate, quality, emissions data.