Is your facility a major HAP source with combustion turbines that were added after 2003? If so, the US Environmental Protection Agency (EPA) has proposed that you conduct stack testing to demonstrate compliance with a formaldehyde standard, and for many turbines, also limit the duration of startup.

On April 12, 2019, the EPA proposed a rule to lift the stay on the National Emissions Standards for Hazardous Air Pollutants (NESHAP) 40 CFR 63 Subpart YYYY for two subcategories of stationary combustion turbines. Lifting this stay, which has been in effect since August 2004, would require operators of lean pre-mix and diffusion flame gas-fired combustion turbines at major sources constructed after January 14, 2003, to meet a 91 parts per billion by volume (ppbv) formaldehyde limit. The rule is expected to be finalized in 2020, at which time affected sources would become subject to the rule and have 180 days from the effective date to complete initial performance testing, which may be problematic for some facilities as it limits the time for contingencies. Although EPA is proposing that the health risk from the turbine source category is acceptable, and there have been no new technology developments, there are significant changes to the proposed rule:

  • Removing the stay for lean pre-mix and diffusion flame gas turbines
  • Proposing amendments to startup, shutdown, and malfunction (SSM) provisions, including limitations on startup duration for simple and combined cycle gas turbines
  • Implementing electronic reporting


NESHAPs are emissions standards set by EPA for categories of facilities that emit one or more of 187 listed toxic air pollutants. The standards for a particular source category at major HAP sources require the maximum degree of emission reduction that EPA determines to be achievable, which is defined as the Maximum Achievable Control Technology (MACT). EPA identified stationary combustion turbines as a category of sources that emit one or more of the listed air toxics. The final MACT standards for stationary combustion turbines were published on March 5, 2004 (69 Fed. Reg. 10512); however, they never went into effect for gas-fired combustion turbines.

In response to a petition filed by the Gas Turbine Association, EPA proposed to delist four subcategories from the stationary combustion turbines source category under Section 112(c)(9) of the CAA: (1) lean pre-mix gas-fired turbines, (2) diffusion flame gas-fired turbines, (3) emergency combustion turbines, and (4) stationary turbines located in the North Slope of Alaska. Pending the outcome of its proposed delisting, EPA stayed the rule as applied to lean pre-mix and diffusion flame gas turbines to avoid unnecessary compliance expenditures.

In 2007, the DC Circuit Court held that EPA had no authority to delist subcategories under Clean Air Act Section 112(c )(9)(B). According to the DC Circuit decision, delisting applies only to entire source categories. [See NRDC v. EPA, 489 F.3d 1364 (DC Cir. 2007)] Based on the proposed results of the residual risk analysis conducted as part of this current rulemaking, EPA has indicated that it does not have sufficient information to delist the entire source category at this time. Therefore, the lean pre-mix and diffusion flame gas turbine subcategories must remain subject to the rule, the stay will be terminated, and many combustion turbines at major HAP sources will now be subject to the emission limits. Based on a limited review of data, many, but not all combustion turbines constructed since 2003 will meet the formaldehyde standard.

SSM and Electronic Reporting

In addition to lifting the stay, EPA proposed revisions to the SSM provisions in order to ensure consistency with the 2008 DC Circuit decision Sierra Club v. EPA 551 F.3d 1019 (DC Cir. 2008) in which the court vacated the SSM exemption contained in 40 CFR 63.6(f)(1) and 40 CFR 63.6(h)(1) stating that emissions standards must be continuous in nature. For lean pre-mix and diffusion flame gas turbines subject to the numerical standard, a work practice is proposed for startup. EPA proposes that during turbine startup, owners and operators must minimize the turbine's time spent at idle or holding at low load. In addition, operators must minimize the startup time to a period needed for appropriate and safe loading of the turbine, not to exceed one hour for simple-cycle turbines and three hours for combined-cycle turbines. Because EPA does not have information to indicate that organic HAP emissions are higher during shutdown than during normal operation, it is not proposing a different standard during shutdown. Combustion turbines, other than these two types, do not have an hour limitation in the rule as currently proposed.

While many recent combined-cycle plants have quick-start options, depending upon the type of startup (hot, warm, or cold), a conventional combined-cycle turbine cold startup can last anywhere between three and six hours. Therefore, while a work practice standard is a good approach, the three-hour combined-cycle startup time limit may be too inflexible for some units.

In addition to the work practice and duration requirements, there are proposed changes to recordkeeping and reporting during SSM. These include maintaining records of startup conditions and excess emissions events. Owners and operators will be required to record the date, time, and duration of startup periods. For excess emissions events, owners/operators will be required to record actions taken to minimize excess emissions, to estimate the amount of emissions over the standard, and to describe the method used to estimate the excess emissions. Electronic reporting via the Compliance and Emissions Data Reporting Interface (CEDRI) is proposed for routine compliance reports and for annual emissions test reports.


Turbines that are subject to the stay could potentially be required to comply immediately with all applicable regulatory requirements of 40 CFR Part 63, Subpart YYYY, upon a final action to remove the stay. Required initial performance tests must be conducted (according to the rule change proposal) within 180 calendar days after the effective date of the final action to remove the stay. The 180-day timeframe may be problematic for facilities that installed a new gas turbine after 2003 without an oxidation catalyst. While many natural gas -fired, dry low NOx turbines can likely meet the formaldehyde limit without add-on controls, compliance with the formaldehyde limit for other gas-fired turbines (such as coke oven gas, refinery gas, etc.) without additional controls is largely unknown. The 180 days for initial performance testing could also prove problematic in that it does not allow for compliance contingencies, and many sources will conduct stack testing for formaldehyde for the first time. The timing required to demonstrate compliance with the proposed formaldehyde limit is a focus of industry comments on the April 2019 rule proposal. Industry groups have requested that EPA allow additional time for installing catalytic oxidation, as it may be required for some turbines that do not already employ this control equipment.


EPA estimates that 182 new lean pre-mix and diffusion flame gas-fired turbines were constructed or reconstructed between 2003 and 2017, with 39 of those becoming subject to the rule over the next three years. With the rule expected to be finalized in 2020, new (2003-onward) gas turbines at major sources will become subject to the rule and have 180 days from the effective date to complete initial performance testing.

A copy of the proposal and rule background can be found at the Federal Registry or EPA websites.