Cementing the PC MACT

On September 9, 2010, the U.S. EPA promulgated the final amendments to the National Emissions Standard for Hazardous Air Pollutants from Portland Cement Manufacturing Industry (40 CFR 63 Subpart LLL), also known as the PC MACT.  EPA’s self-proclaimed “historic” rulemaking includes some important implications for cement manufacturers and other sources that may be subject to revised MACT rulemaking in the future.  Among the landmark issues that EPA promulgated were:  1) finalizing a more rigorous methodology for setting MACT floor levels that will require significant investment in control technology by industry; 2) incorporating unprecedented levels of continuous emissions monitoring for compliance assurance; and 3) a new regulatory schema for regulating Startup, Shutdown, and Malfunction (SSM) emissions.  Simultaneously, EPA finalized aggressive new limits for PM, NOx, and SO2 in a revision of the New Source Performance Standard for Portland Cement Plants under 40 CFR 60 Subpart F which is not addressed in this article.

MACT Floor Values

In calculating these emission limits, EPA considered numerous court cases that have fundamentally changed the methodology for calculating MACT floors.  EPA has generally expressed that the approach closely adheres to the tenets ascribed by the courts in various rulings and a plain reading of the Clean Air Act.However, regulated sources may consider this approach as an impractical or unachievable application of the law and the related court rulings.  Many of EPA’s assumptions are based on the Brick MACT case [Sierra Club v. EPA, D.C. Cir. 2007] which concluded:

  • Sources with low HAP emissions, due to low levels of HAP in their raw materials, can be considered best performers in establishing the MACT floor
  • Floors for existing sources must reflect the average emission limitation achieved by the best performing 12 percent of existing sources, not levels EPA considers to be achievable by all sources
  • EPA cannot set floors of “no control”;  EPA must set floors for all HAP, including those not controlled by at-the-stack control devices
  • EPA cannot ignore non-technology factors that reduce HAP emissions
  • EPA can consider the variability in the dataset; however, EPA can only use the variability of the “best performers” in calculating the floor.  Limiting the dataset will result in a smaller standard deviation and will lower the MACT floors.

This calculation methodology can also affect the floor values for other industries to be regulated in the future such as the proposed Boiler MACT.  (For a discussion on the potential effect of this methodology on other industries, see Portland Cement MACT Proposal Potentially Affects Other Industries, published in Environmental Quarterly, July 2009.)  In the final rule, the outcome has effectively codified the approach that some have dubbed the “Franken MACT” approach – where the final limits are the best of best on a pollutant-specific basis and not on a best source achieved basis.  The result, in effect, provides final emission limits that no single U.S. cement kiln in existence can meet without additional controls.

Table 1 summarizes the final MACT limits for existing and new cement kilns, and compares them with the limits in the proposed rule.  Table 1 shows that the mercury standards for new and existing sources are higher than the standards in the proposed rule due to the additional data that EPA received after proposing the rule and EPA’s reconsideration of how to account for variability.  The PM and THC emissions limits were revised downward and upward, respectively from the proposed rule.  A key reason for this reduction in emission limits is EPA’s correction of its variability equation rather than a change in the dataset.  Additionally, EPA has required CEMS for PM in the final rule and has therefore added a 30-day averaging period to the limit that resulted in a change of the standard.
Table1_MACT Kiln Limits

In spite of receiving numerous comments from industry and other interested stakeholders requesting a separate category for raw materials with high mercury content, EPA has not included a separate category for such materials in the final rule.  EPA reasoned that there are only two quarries at cement plants in the U.S. where the mercury content of the limestone is well above the general population of cement kilns in the U.S. EPA has also noted that the two kilns that use the limestone from these quarries emit more than 25 percent of the mercury emissions from this source category.  As such, EPA has not defined a separate subcategory for the two facilities.  EPA stated in the preamble, however, that if these two kilns achieve significant near-term reductions by the compliance date, EPA could extend the compliance deadline for them.  However, they also acknowledge that given the current state of mercury control that these kilns are not expected to achieve the final MACT limitations.  

Another significant issue that has been disregarded by EPA is the option of utilizing a risk-based standard for HCl.  EPA has agreed that a risk-based standard for HCl would have been much higher than the proposed standard.  However, EPA noted that emissions of other non-HAP pollutants will also be reduced if a lower HCl standard is implemented and, on this basis, it rejects a higher standard for HCl.  That is, EPA relied significantly on collateral benefits of non-HAP pollutant reductions to support the final HCl limits.  Although EPA criticized the cement industry’s HCl risk assessment, it supported the use of a risk based HCl limit for other MACT categories.

Monitoring Requirements

Undoubtedly, complying with the numerical emission limits in the final PC MACT will be challenging for the cement industry.  However, the emission limits themselves may not be the most challenging part of the rule when compared with the monitoring requirements.  While the 30-day averaging period that is utilized by EPA offsets the stringency of the numerical limits to some extent, a plethora of monitoring devices is required by the rule; some of which have not been demonstrated in continuous use on cement kilns in the U.S.  For instance, EPA requires mercury CEMS on kilns while virtually all of the mercury CEMS in the U.S. that are used for compliance demonstration purposes are installed at power plants.  A summary of monitoring requirements is provided below.Kilns and clinker coolers subject to PM emission limits must measure the concentration of PM, the flow rate to convert the PM concentration to mass, and the clinker production or the kiln feed rate to normalize the pollutant mass for comparison to the standard.  As such, EPA requires facilities that are subject to this limit to install PM CEMS with the ability to collect data at least every 15 minutes.  The 15-minute measurements will then be rolled into daily and 30-day averages.  In addition to the CEMS, EPA requires installation of a flow meter to measure the exhaust flow rate and hourly recording of the feed or clinker production rate with +/- five percent accuracy.  The monitoring requirements for facilities subject to mercury limits are similar to PM, in that the rule requires Hg CEMS or a sorbent trap-based integrated monitoring system as well as continuous measurement of the exhaust gas flow rate.Other pollutants that must be monitored with CEMS include THC and HCl.  All sources that are subject to the THC emission limit must use a THC CEMS to monitor the concentration of the pollutant.  However, sources that are equipped with an alkali bypass stack can use the results of the initial or following performance tests to demonstrate compliance.  All sources that utilize the Organic HAP (OHAP) emission limit alternative to total THC must also monitor THC.  Sources subject to HCl limits are required to install HCl CEMS unless they install a wet scrubber.  However, in that case, they must also install an additional continuous parameter monitoring system (CPMS) to measure the parameters of the wet scrubbers.With the exception of THC monitoring, a facility can petition EPA to perform alternate monitoring for the pollutants subject to emission limits.  The alternative cannot have different averaging periods than the rule unless the facility can justify that the alternate monitoring method is at least as accurate or stringent.  Should the facility prefer alternate monitoring, the application must be submitted and approved before the notification of performance test.  However, until the administrator approves the application, the facility must comply with the monitoring requirements of the rule.  Facilities must also prepare a site-specific monitoring plan that addresses operation and maintenance procedures, accuracy tests, and data quality assurance procedures for the monitoring device.  The monitoring plan must be kept on site and will only be submitted to the EPA upon request.

Startup, Shutdown, and Malfunction Limits

Another unique aspect of the final PC MACT rule is EPA’s approach to startup, shutdown, and malfunction (SSM) emissions.  The D.C. Circuit Court ruled in 2008 that EPA cannot exempt periods of startup, shutdown, and malfunction from MACT requirements and EPA is now requiring sources to be in compliance with emissions limits at all times.  In this rulemaking, EPA has recognized that applying the same MACT floor limits for normal operation to SSM periods is impractical.  EPA noted the following reasons to explain why the same limits cannot be applied to SSM periods:

  • The rate of production during the periods of startup and shutdown is zero or virtually zero which rules out the possibility of normalization by feed or production rates.  Even when this rate is non-zero, EPA finds it difficult to directly correlate the HAP concentration with the feed rate or clinker production rate.
  • Accurate flow rate measurement, necessary to convert concentration to mass, is impossible during SSM periods. All parameters affecting the flow rate are constantly changing during these periods since the kiln is not operating in a steady state, by definition.
  • Averaging out the concentration spikes over time by using a 30-day averaging period is challenging.  A rotary kiln can operate continuously for an extended period of time with no need for maintenance and a 30-day startup/shutdown dataset could take a few years to develop.

To address the above-mentioned items, EPA established standards on a concentration basis.  An averaging period of seven days was established to facilitate data collection.  EPA recognizes the fact that startup and shutdown events do not take seven days and as a result, compliance with the standard cannot be determined based on one startup or shutdown event.  However, EPA believes that this shorter averaging period will allow the facility to collect the data, calculate the average, and determine compliance in “certainly less than a year.”  EPA’s approach to shutdown is exactly the same as startup as EPA considers shutdown operation to be “in many ways a mirror image of startup.”  The limits for startup/shutdown have been summarized in Table 2.

Tabel2_MACT Kiln Limits

Due to the issues above, concentrations of PM and Hg can be directly compared to the standard, as measured with CEMS, during startup.  The THC standard during normal operation is already concentration-based and needs no conversion.  Since the oxygen content often fluctuates notably in the stack gas during startup and shutdown periods, the THC standard during these periods does not include oxygen correction (as required under normal operations).  The standard for HCl during startup/shutdown periods is the same as normal operation for the sources equipped with CEMS, with the removal of the oxygen correction factor and the different averaging period of seven days.  However, sources that control emissions with wet scrubbers may use stack testing and parametric monitoring.  EPA set the HCl standard of zero in such cases stating that there are no parameters to be monitored because emissions of HCl are not expected from the cleaner fuels that are utilized during startup.  

EPA took a different approach to setting the standards for dioxin/furan.  For dioxin/furan, the standards are either based on concentration or a combination of concentration and inlet temperature to the PM control device (See the notes in Table 1).  The temperature measurements at the inlet of the particulate matter control device (PMCD) can be increased by 10 percent during startup/shutdown.  

The MACT rule does not include specific malfunction limits as EPA found setting limits for these unplanned events impossible.  EPA recognized in the preamble that malfunction events are not actual operating modes and therefore, EPA cannot set limits for them due to their sudden and unexpected nature.  As such, EPA added an “affirmative defense” to the MACT that requires the facility to demonstrate the event actually meets the definition of malfunction as described in 40 CFR 63.21.  Penalties will apply if the facility cannot prove that the malfunction was not caused by poor maintenance or careless operation.

The definitions for startup and shutdown have not been well described in the rule, potentially creating confusion.  Nevertheless, EPA offered the following definition of startup in the preamble of the final rule, “Startup is the period of time between when fuel is first introduced into a cement kiln that is not firing fuel, and when the kiln temperatures are within normal operating limits, the kiln is using its normal operating fuel, and the kiln is producing clinker.”  This definition raises a few questions.  First, the terms “normal operating fuel” and “normal operating limits” are undefined.  Second, EPA states in the preamble that kilns will probably be in compliance during startup/shutdown periods because cleaner fuels are burned and feed rates are lower during periods.  However, based on the definition, the kiln can burn its normal operating fuel and still be in startup since it has not met the temperature or production requirements.  Shutdown is defined as “the period of time between when kiln raw material feed is shutoff and gas flow through the kiln ceases.”  Therefore, until the material feed is completely stopped and the kiln ID fan is turned off, the kiln is not considered to have completed a shutdown period.

Determining initial compliance with the SSM requirements of the new PC MACT may be difficult.  Most facilities do not have any test data on startup or shutdown because of the previous exemptions, nor do they have the ability to control emissions very well in these phases of operation.  Collecting such data to calculate the 7-day average concentration for all subject pollutants could take longer than a year.  This leaves a short time to review and implement all steps necessary for compliance before the compliance date, which is only three years from the date of publication.  It is also important to note that while it was EPA’s intent that the same limits apply during SSM, the 7-day averaging period that is applied to startup/shutdowns is more stringent than the 30-day averaging period applied to other pollutants during normal operation.  This small discrepancy could make the startup/shutdown limits subject to scrutiny and future public commenting.  Lastly, add the complications of applying and managing CEMS to meet two sets of emissions limits (normal and startup/shutdown) provides for a very complicated regime for a cement kiln operator to continuously comply.


Indeed, it is understandable that EPA would describe the final PC MACT rule as “historic.”  Given the scope of the final rule, litigation of its lawfulness is a near certainty.  For those source categories in the process of or expected to go through a MACT revision, understanding EPA’s approaches to the PC MACT rule is a must to prepare for these new realities.  Given the nature of cement kilns and their emissions, each compliance assurance approach will be unique to a kiln.  Although the final chapter on the PC MACT rule has not yet been written, the compliance date is known.  As a result, sources affected by the PC MACT rule should begin efforts to comply with the rule.  In addition to efforts to comply with aspects of the final rule (not addressed in this article) that are slated to become effective on or about November 8, 2010, these efforts may include collecting kiln specific emissions data (especially on startup/shutdown operations) while also assessing variability of emissions, developing plans to install continuous monitoring, and beginning to assess various control strategies.

140 CFR 63.2 defines Malfunction as any sudden, infrequent, and not reasonably preventable failure of air pollution control and monitoring equipment, process equipment, or a process to operate in a normal or usual manner which causes, or has the potential to cause, the emission limitations in an applicable standard to be exceeded.  Failures that are caused in part by poor maintenance or careless operation are not malfunctions.