Pom Pom Ganguli welcomed the SRC members and the audience to the ninth SRC meeting of 1999. The topics listed below were discussed during the meeting.

1. New BACT determinations for the following categories of equipment: Circuit Board Photoresist Developer (new category), Spray Booth, Rubber Compounding – Roll Mill (new category), Lithographic or Offset Printing - Non-Heatset, Boiler, and Heater – Refinery.
2. A report on measuring low concentrations of NOx and VOC.
3. A review of boiler BACT/LAER.

No comments were received regarding the minutes from the August SRC meeting.

The technical presentation by Cannon Technology and BOC Gases scheduled for this day was rescheduled to October 12, 1999, at 10:00 am, in Conference Room GB.

Item: Staff should prepare a formal written report for the SRC (every six months) regarding the status of the BACT process. The report should include operating data for the new types of controls that are used to establish BACT limits.

Response: Staff is working on this item, and the status report should be available by the next meeting. So far, staff has not received any suggestions (in writing) from SRC members regarding the content of the report.

Item: Staff should post SRC meeting agendas on the BACT web page.

Response: Staff intends to post SRC meeting agendas and meeting minutes on the web page. This information should be available on the BACT web page by the next SRC meeting.

Item: Staff should include in the BACT docket information that indicates where other sources of BACT information can be found (i.e., some kind of cross-reference).

Response: Staff intends to include cross-referencing information in the BACT docket.

In July 1999, the Boiler Working Group presented the SRC with a report on the boiler BACT determination at Alta Dena Dairy. The report raised issues regarding the quality of data used by staff to support the BACT determination. Based on their review of the operating data, the Boiler Working Group recommended that the District suspend the BACT determination until more adequate data--that proves that the unit met the reliability criteria specified in the BACT Guidelines Part A document--is provided to the District. Staff subsequently analyzed the Boiler Working Group’s report and the operating data at Alta Dena, and the preliminary finding is that there is not enough data available to prove that the boiler operated continuously for at least six months at a minimum of 50% design capacity. The supplier of the control equipment at Alta Dena Dairy is currently trying to produce additional data relevant to the BACT determination and they have requested two weeks time to produce the additional data. Based on the vendor’s request, permit applications potentially subject to the 5 ppm NOx limit (based on the Alta Dena Dairy BACT determination) will be held for another 7-10 days after which a decision regarding boiler BACT will be made. Staff has also begun looking for alternative technologies (such as ultra low NOx burners and SCR) that could qualify as achieved in practice examples for BACT.

• If the currently available data does not support the 5 ppm BACT determination at Alta Dena Dairy, then staff should allow the permits (currently being held) to go through at the next level of BACT. (Greg Danenhauer)

• The holding of permits potentially subject to the 5 ppm BACT limit is causing construction delays and the Disney Company would like to see a speedy resolution of this issue. (Bob A.)

• I think the process for making BACT determinations needs to change and Sempra Energy supports efforts to standardize and streamline protocols. (Deanna Haines)

• How is the District handling BACT for other types of external combustion equipment (e.g., ovens and dryers)? (Matt Brueck)

• How is the District going to handle the ozone slip issue? (Hal Taback)

Knut Beruldsen presented a Section I BACT determination for a photoresist developer at Morton International. The equipment has been operating since August 1999 and is located in Tustin, California. The primary pollutant from this type of operation is VOC. Permit conditions limit the photoresist developer (organic) vapor pressure to 5 mm Hg at 68 degrees Fahrenheit, and the VOC emissions to 20 lbs/month. Due to the small quantity of VOC emissions, add-on controls are not required for this operation. For this case, the BACT requirement is use of a low vapor pressure photoresist developing material.

Knut Beruldsen and Marty Kay presented five Section I BACT determinations for spray booths located at several different places in the District. The VOC emissions from the spray booths are limited by permit conditions to levels below the emission levels where add-on controls have been achieved in practice. In addition to BACT requirements, the spray booths are subject to various Regulation XI rule requirements.

Knut Beruldsen presented a Section I BACT determination for a rubber roll mill operating in Valencia, California. The permit requires that no organic solvents be added to the equipment, and that the maximum operating temperature not exceed 200 degrees Fahrenheit. The primary pollutant from the roll mill is particulate emissions, and these emissions are vented to a baghouse.

Knut Beruldsen presented a Section I BACT determination for a lithographic printing operation operating in Sun Valley, California. The primary pollutant from the equipment is VOC emissions. Permit conditions limit the VOC usage to 667 lbs/month, and the fountain solution VOC content to < 8%, by volume. The BACT requirement for this operation is compliance with the permit limits and use of Regulation XI compliant solvents and inks.

Marty Kay presented one Section II and three Section I BACT determinations for boilers operating in the District and in Corcoran, California. The NOx emissions from the boilers are controlled by SCR systems or ultra low-NOx burners. The NOx emissions from the SCR systems are limited by permit conditions to 9 ppm at 3% oxygen, except during start-up. In addition, the 9 ppm NOx limit does not apply until the stack temperature exceeds 510 to 575 degrees Fahrenheit.

Note that newer SCR catalysts are effective at much lower temperatures. Permit conditions limit the ammonia slip from each of the boilers to 20 ppm at 3% oxygen. The permit limits for CO emissions vary from 100 to 400 ppm at 3% oxygen. The boilers are equipped with NOx CEMS to ensure that they continuously meet the NOx permit limits. The boiler operating at UCI Medical Center is equipped with an ultra low-NOx burner. The NOx and CO emissions are limited by permit conditions to 9 and 50 ppm, respectively, at 3% oxygen. The boiler is currently operating at around 25% capacity. The boiler is equipped with a NOx CEMS. The boiler at Corcoran is equipped with an ultra low-NOx burner. The NOx and CO emissions are limited by permit conditions to 12 and 30 ppm, respectively, at 3% oxygen. The boiler was source tested on 3/11/97 using EPA Methods 7E and 10 and compliance with the permit limits was demonstrated. The boiler is listed in both the CARB and USEPA BACT Clearinghouses. The boiler at Corcoran is noteworthy in that it is much smaller than the other three boilers included in this discussion. The maximum heat input to the boiler is 8.1 MMbtu/hour. All four of the boilers have been operating for several years and compliance information (including variances and/or violation notices) is included in the BACT determinations.

Marty Kay presented a Section I BACT determination for a refinery heater operating in Wilmington, California. The heater is fired with pressure swing absorption off gas (low Btu value) & refinery gas. The NOx emissions from the heater are controlled by SCR and the permit limit is 7 ppm at 3% oxygen, except during startup and shutdown and when the exhaust temperature is less than 600 degrees Fahrenheit. The permit limit for ammonia slip is 20 ppm at 3% oxygen. Compliance with permit limits has been demonstrated by source test and CEMS. The heater has been operating since 1994 and compliance information is included in the BACT determination.

• Water based photoresist developers are available. Why was use of a solvent based photoresist developer deemed BACT? (Katy Wolf)

• The BACT determination lists 99% overall control efficiency for the baghouse. Was this equipment source tested? (Karl Lany)

• Is the 667 lbs/month, the ink limitation, or does it include the cleanup solvent? If it includes the cleanup solvent, what percentage of the VOC emissions are due to use of these materials? (Katy Wolf)

• Are the boilers at Darling International and Kal Kan Foods base loaded or variable loaded? (Deanna Haines)

• An Alzeta burner was installed at a Boeing facility in Downey over 5 years ago, and I heard that there were some problems associated with that burner. Could you comment on this? Also, what’s the difference between that installation and what you are producing now? (Deanna Haines)

• Regarding the source test on the boiler at Corcoran, what were the loads during the three runs? (Deanna Haines)

John Higuchi presented technical information on low concentration measurements. His discussion primarily focused on the following three issues:

1. How well can we measure NOx and VOC emissions below 5 ppmv?
2. What needs to happen to make < 5 ppmv measurements of NOx and VOC more predictable and reliable?
3. How low can we go in concentration for measurements of NOx and VOC?

NOx emissions can be measured to plus or minus 1 ppm using enhanced Method 100.1, and VOC emissions can be measured to plus or minus 1 ppm using new Method 25.3. Method 25.3 is a modified version of Method 25.1. Method 25.1 has a significant positive bias when CO₂ and moisture levels are high and VOC levels are below 50 ppmv. Method 25.3 eliminates the bias and improves detection limits. Instead of using a stainless steel trap at dry ice temperature, Method 25.3 uses a small glass impinger containing hydrocarbon free water followed by a Summa-Polished^® stainless steel canister, to eliminate the CO₂ bias and the background from the steel traps, and to accommodate lower VOC concentrations, respectively.

A test method for low NOx streams that may be considered is modified Method 7.1. In the absence of particulate nitrate interference the detection limit is 1 ppm. District staff is currently looking at this method for measuring the NOx emissions from a LTO system.

The Bay Area AQMD, CARB, and EPA have test methods based on continuous hydrocarbon monitors. These monitors provide continuous readout and have detection limits (measured as C1) below 1 ppm, depending on the detector, compound, and pre-concentration. However, for many applications (especially combustion sources) variable responses to different species by these detectors affects the uncertainties in the measurement of emissions by these devices. In addition, some of these devices typically detect ethane and methane, both of which are not VOCs, and these need to be accounted for separately.

Gas chromatographic methods can be applied but efforts to take a gas chromatograph out in the field can run into many problems. It takes a lot of work, but ppb detection levels are achievable.

Emerging technologies include: hot FTIR (which takes a sample right out of the stack hotlined to the FTIR), open path FTIR (which doesn’t even have a sampling system), and open path UV-DOAS. These technologies although commercially available are still basically in the experimental stage and they are not typically used in source tests. The detection limits for the open path technologies depend on the path length and may be from ppb to ppm levels.

John Higuchi provided information regarding the relative accuracy of NOx measurements based on actual operating data, and the data indicate that when NOx emissions are measured at less than 5 ppm there is very close agreement (less than 1 ppm) between the reference method and the emissions measured by the facility’s CEMS. Based on RECLAIM experience, testing done on the SCONOx installation in Vernon CA, NO and NO2 spiking with recoveries, and RATA information obtained over the past few years, the District feels that the level of uncertainty regarding NOx measurements below 5 ppm is plus or minus 1 ppm.

The District performed an extensive validation study to determine VOC measurement uncertainty using EPA Method 301. The pooled standard deviation for the unspiked samples was 0.5 ppm and there was a 8% statistically significant low bias. The EPA has tentatively approved that method for low VOC with a bias adjustment factor. The District expects that the interlaboratory variance will be larger; therefore, the level of uncertainty for VOC measurements below 5 ppm has been raised from 0.5 to 1 ppm.

There are several factors that affect measurement uncertainty. Analytically (at least for NOx) you can see down to ppb to ppm. The primary factor that affects measurement uncertainty is sample transport and conditioning. Also, physical conditions (such as, temperature and pressure) and process characteristics can affect measurement uncertainty.

District staff agrees with industry representatives and other governmental agencies that there needs to be a reexamination of the NOx reference method with low concentration in mind. Special considerations are needed for NO2 converter efficiency, NO2 losses, stratification, low level bias, interferences, calibration standards, and analyzer stability. It would also be nice if there was one national standard for low NOx measurement.

For low VOC measurement the following enhancements are needed: interlaboratory testing to tie down interlaboratory uncertainty, and guidance regarding the appropriate number of tests.

District staff also feels that we need to reexamine the CEMS itself with low concentration in mind. Consideration needs to be given to stratification requirements, sample probe configuration, periodic testing of sampling system for NO2 losses, and low level bias testing.

The end user (typically enforcement) defines what is acceptable data with respect to accuracy, precision, reliability, and timeliness. However, as emission limits get lower and lower reasonable tradeoffs are needed, such as requiring additional tests or greater uncertainties. Additional costs will be necessary in order to obtain good test information at low concentrations. Some of the limiting factors include (1) technology limitations (e.g., inherent lower limits of detection, ability to mitigate interferences, and conditioning samples without losing pollutant of interest), (2) cost issues (e.g., equipment and testing costs could rise exponentially to reduce absolute uncertainty, and method development and validation requires significant resources), and (3) it takes a significant amount of time develop a new test method. It took District staff 2 years to develop Method 25.3 to where we could submit it to EPA for approval. At this time (using the technology we have) the District feels that we probably can’t go much lower as far as uncertainty goes without some additional work.

• What method does the District use to measure ozone slip? (Audience)

• How does the uncertainty issue affect the implementation of BACT? (Jerry Kraim)

The District intends to present the results of the boiler BACT analysis at Alta Dena Dairy at the next SRC meeting. At the same meeting, staff will provide updated BACT guidance for small boilers and other external combustion devices. Staff also intends to post information on the BACT web page regarding this issue.

There was no further discussion and the meeting was closed. The next meeting is scheduled for October 21, 1999, at the same time. Note that meeting information is available on the AQMD web page.