1. How can I measure PM10 and PM2.5 emissions from my stack?
There are several approaches in common use for measuring particulate matter (PM), particulate matter smaller than 10 microns (PM10) and particulate matter smaller than 2.5 microns (PM2.5) from industrial sources. The choice of method depends on several factors:

• Reason for the measurements – compliance with regulatory standard, design information, control equipment guarantee, etc.
• Source gas characteristics – temperature, particulate concentration, moisture content, etc.
  
• Source construction – size of stack, size and geometry of test ports.

The standard reference method for measuring PM is EPA Method 5 of 40 CFR 60, Appendix A. This method is suitable for most industrial sources, and provides a measure of the total amount of filterable solid particulate matter in the source. The method defines the particulate by the temperature of the filter that is used to capture the particles. This temperature is normally 248F, but can vary depending on the type of source and regulatory requirements.

EPA Method 201A of 40 CFR 51, Appendix M is the standard method for measuring PM10. This method uses an in-stack cyclone that separates the PM10 from the total PM. There are three significant limitations to this method:

1. It normally requires a stack diameter of at least 18 inches (approximately).
2. It normally requires a sampling port diameter of 6 inches.
3. The gas cannot contain entrained water droplets.

Usually, if any of these criteria are not met, one of two alternatives are used in lieu of Method 201A:

1. Method 5 is used to measure total PM, and all of the particulate is considered to be PM10. This potentially overestimates the PM10 level. This approach is normally acceptable from a regulatory standpoint, but may not be acceptable for design or other diagnostic purposes.
2. Method 5 is used to measure total PM10, and the filter is analyzed (usually microscopically) to determine the particle size distribution of the particulate sample. The total PM result is then prorated using the PM10 fraction from the particle size distribution. This approach is usually more difficult to get accepted by regulators, but is often more valuable for engineering purposes.

Both Method 5 and Method 201A measure solid, filterable particulate matter. Other particulates may exist in an aerosol state and go undetected with these methods. If this form of particulate is also of interest, then EPA Method 202 (40 CFR 51, Appendix M) can be combined with either method to provide a separate measurement of "condensable" particulate matter (CPM). CPM is usually considered to be less than 10 microns, and is often included with PM10 compliance determinations.

At the present time, there is no USEPA promulgated or proposed reference method for measuring PM2.5. Nor is there a generally recognized "off-the-shelf" procedure that has gone through enough industry evaluation to be considered validated. One of the chief problems with PM2.5 measurement is the fact that PM2.5 is believed to be composed predominantly of aerosols and other non-solid chemical species that are not necessarily a particulate at stack conditions of temperature and pressure.

There are two lines of approach currently being pursued for PM2.5 measurements. One approach is to use an adaptation of Method 201A and incorporate a second cyclone to differentiate between PM10 and PM2.5. This approach has been evaluated by the Portland Cement Association (PCA) and is described in detail in the 1996 PCA R&D Publication No. 2081. The limitation of this approach is the fact that, as mentioned above, most of the PM2.5 may not be in solid form and thus would not be collected using this method.

A second approach is being evaluated that tries to address the issues of aerosol particulate and atmospheric formation of PM2.5. This approach involves extracting the stack gas and allowing it to slowly cool in a settling chamber. This emulates what may actually happen as the stack gas exits to the atmosphere. The cooled gases are then sampled using an approach similar to the PCA modified 201A method.
This approach is still at the research stage and is not available for commercial applications.

It is possible that valuable information concerning PM2.5 could be gained by collecting particulate using Method 5 and determining the particle size distribution microscopically. This would provide an estimate at least of the solid portion of the PM2.5. In this case, Method 202 should be incorporated to determine the CPM. However, the fraction of the CPM that may be attributable to PM2.5 cannot be determined.

2. What advantages are there in using 3-dimensionsal flow measurements in certifying my Part 75 CEMS?

With the introduction of Methods 2F, G and H, we can now provide more accurate flow data. In the past, there have often been descrepencies between heat input calculations and measured volumetric flow rates. The has been caused primarily due to the inherent bias of the direct pitot tube measerements.

The pitot tube does not allow for the pitch and yaw components of velocity in a stack or duct. More importantly, the S-type pitot tube introduces a positive bias when there is cyclonic flow. Refering to an EPA document, you can see that there is a positive bias from 0 to approximately 20°. A cyclonic angle of up to 20° is allowed by the EPA because of this positive bias.

In using the 3D flow procedures, we can get an accurate flow that is corrected for both yaw and pitch effects to the velocity. Using these corrected flows we can now have accurate mass emission calculations.

The other correction involved is the Wall Adjustment Factor (WAF). This procedure allows for the measurement of degredation of velocity along the walls of a stack or duct. This allowance can be from 1% to 3% of total flow, depending on the materials of the stack or duct.

The benefits can be seen in two ways. One is that we can see a reduction in emissions above 10%. The second view is that industry has always been overreporting and the new methods finally give accurate numbers.

Depending on the view taken, a company can maximize the monitary rewards. If a unit has been permitted or is permitted at a historically based emissions limit, a permit modification to the new (actual) reduced rates can open up emission credits that can be a source of income, either in sales or internal use.

Using sulfur dioxide as an example, the current market value is near $150 per ton/yr. From current projects, we have seen emissions at 20000 ton/yr. With a 10% savings due to the remeasured flow and economic gain of $300,000/year could be obtained.

The outstanding issues are to determine how to legally obtain the credits as well as find the market to sell the shares in. The arguements to overcome will be whether these emissions "reductions" are real or imaginary. Have they really been created or have corporations been just over reporting.

3. In total hydrocarbon testing, what does the term relative response factor mean, and how is it used?

In general, the response of a hydrogen flame ionization detector (FID) is proportional to the number of carbon atoms (specifically carbon-hydrogen bonds) in the compound being analyzed. However, the FID is more sensitive to aliphatic hydrocarbons (e.g., paraffins) than any other class of organic compounds. Compounds containing oxygen, such as alcohols, aldehydes and esters give a somewhat lower response than that observed for hydrocarbons. Similarly, nitrogen-containing compounds (e.g., amines, amides and nitriles) and halogenated compounds also show lower relative responses as compared to hydrocarbons. Materials containing no hydrogen (e.g., carbon tetrachloride) generally give the lowest responses.

As the ratio of carbon atoms to non-carbon atoms (e.g., oxygen, nitrogen, halogens) increases, the effect of the non-carbon atoms upon the FID response diminishes. At a carbon to non-carbon ratio of about 5:1, the effect diminishes to such an extent that the response is similar to that of the corresponding hydrocarbon.

4. How do I measure formaldehyde emissions from my gas turbine?

Numerous methods exist for determining formaldehyde in gas streams, including colorimetric analysis, chromatography (gas, liquid and ion), and spectrometric. However, only a few are considered accurate and specific enough to demonstrate compliance with federal or state air requirements.

The USEPA generally accepts three methods for determining formaldehyde emissions:

• EPA Method 0011 of SW-846
  
• California Air Resource Board (CARB) Method 430
  
• EPA Method 320 of 40 CFR 63
  

EPA Method 0011 and CARB Method 430 are very similar methods. Both use a liquid solution of dinitrophenylhydrazine (DNPH) to absorb formaldehyde from the gas stream and convert it to a stable chemical derivative that can be quantified using high performance liquid chromatography (HPLC). These methods utilize traditional "wet" stack sampling procedures to collect an integrated sample that is then analyzed at an off-site laboratory. This approach yields a "snapshot" measurement of average formaldehyde concentration over a specific interval of time.

EPA Method 320 is a generic method using Fourier transform infrared spectroscopy (FTIR). This technology uses sophisticated computer analysis of infrared spectra of the gas stream to determine its composition. The analysis is direct and on-site, and provides near real-time and semi-continuous measurements of the gas composition.

Table 1 summarizes some of the pros and cons of each of the methods. The most relevant problem with the DNPH methods is their lack of reliability in gas streams containing significant amounts of NO2. The Gas Technology Institute - GTI (formerly the Gas Research Institute - GRI) has done research that suggests that NO2 scavenges the DNPH from the sample solution, thus making it unavailable to collect the formaldehyde and producing a negative bias in the results. The GTI internet web page contains several publications regarding this research (www.gti.org).

Much of the GTI research into this phenomenon was done on IC engines, which generally produce higher concentrations of NO2 than do gas turbines. Their data suggests that the interference may be insignificant if the NO2 concentration is below about 50 ppm. Therefore, it may not be an issue with gas turbines.

There is an alternative DNPH method developed by Ashland Chemical, Inc. using solid sorbent tubes. This method has been reported to have several inherent advantages over the liquid DNPH methods, including less interferences, easier method operation (thus less expensive and more versatile), and more repeatable results.

A study done in 1995 by CARB compared the Ashland method with CARB 430 on IC engine exhaust. This study showed no correlation between the formaldehyde measured by the Ashland method with the NO2 concentration in the gas (CARB 430, on the other hand, did show a negative correlation, suggesting potential interference). Additional information concerning this study can be obtained from Cindy Castronovo of CARB.

The USEPA has not endorsed the solid DNPH method. However, approval may be possible on site-specific cases if the agency can be convinced that acceptable alternatives do not exist. It would probably be a good idea (or even a requirement) to collect side-by-side samples using either CARB 430 or EPA Method 320 for at least some representative sources.

Due to the nature in which FTIR works, interference from other gas species is not a problem with this method. The most significant disadvantage with this method is cost, due to the requirement for more specialized instrumentation and highly trained personnel to conduct the analysis. However, this cost is somewhat offset by the larger and more useful set of data that the method gives.

Through recent conversations with USEPA personnel (Terry Harrison, Sims Roy, both of EPA OAQPS), Clean Air has learned that the soon-to-be proposed gas turbine MACT rule will likely endorse both CARB Method 430 and EPA Method 320 for formaldehyde measurements. The EPA has, however, voiced their concern over the potential NO2 interference with the CARB method. For this reason, there may be some language in the rule that more favorably recommends the use of the FTIR method.

In conclusion, the choice of method probably depends more on the importance of the data being widely accepted than it does on specific technical merits of the individual methods. Clean Air recommends that, although the liquid DNPH methods may provide suitable formaldehyde data for the low-NOx environments from gas turbines, the FTIR method is a better choice based on its higher degree of acceptance within the EPA and state agencies.

Clean Air also recommends that consideration be given to concurrent collection of solid-DNPH samples. This data may be used later for validation (or invalidation) of the solid-DNPH method as a low-cost alternative to the FTIR approach.

5. What is the proper test method to demonstrate organic vapor destruction efficiency compliance for my thermal oxidizer?

The USEPA requires that Method 25 be used for most destruction efficiency compliance tests. However, there are some situations that will allow Method 25A to be used instead. Sufficient justification for using Method 25A in lieu of Method 25 includes any of the following reasons:

1. the expected outlet concentration is less than 50 ppmv as carbon,
   
2. the VOCs are known to consist of only carbon and hydrogen (a.k.a. hydrocarbons),
   
3. the product of the percentages of carbon dioxide and water vapor exceed approximately 100.
   

6. What is the minimum detectable limit for particulate matter measured using EPA Method 5?

There is some concern in industry of whether the traditional measurement approach for particulate matter, EPA Method 5, is sensitive enough to demonstrate compliance with some of the newer regulatory standards being developed from the 1990 CAAA. Clean Air conducted a laboratory study to determine the minimum quantification limit of EPA Method 5 with respect to the operations of sample recovery and analysis. A synopsis of the work is outlined below.

Background
One of the results of the 1990 Clean Air Act Amendments has been the lowering of particulate matter (PM) standards for many sources. For example, the proposed MACT standards for hazardous waste incinerators (HWIs) would set a PM limit of 69 mg/dscm for these facilities. This is significantly lower than the current HWI particulate limit of 180 mg/dscm. Similarly, the MACT standards which were promulgated earlier this year for municipal solid waste combustors set a PM limit of 24 mg/dscm for new facilities, as compared to the previous NSPS limit of 34 mg/dscm.
Even lower limits are being imposed at the state level in many facility operating permits. For example, most natural gas-fired combustion turbines have PM limits less than 10 mg/dscm. Some units have even been permitted as low as 1 mg/dscm (0.0004 gr/dscf).

Scope of Work
Clean Air conducted a laboratory study to determine the minimum quantification limit of the sample recovery and analysis procedures of EPA Method 5. The study consisted of the preparation, recovery and analysis of multiple spiked and unspiked EPA Method 5 sampling trains. The work was done according to standard EPA methodology contained in 40 CFR 136, Appendix B, "Definition and Procedure for the Determination of the Method Detection Limit." The gravimetric results from the analyses were used to calculate the detection limit.

Results
The results show that the expected minimum quantification level of EPA Method 5 is 4.7 mg/dscm (0.002 gr/dscf) for a typical one-hour test (30 dscf sample). These results suggest that EPA Method 5 should be able to adequately test the particulate compliance level using a one to two-hour sampling period for nearly all facilities. Closer evaluation of the sampling duration should be considered, however, if the applicable limit is less than 10 mg/dscm, or if absolute quantification of the emissions is required at levels less than about 20 mg/dscm.

7. I have a source in which the PM10 emissions that were measured using EPA Method 201A are higher than the total particulate emissions which were measured using EPA Method 5. How can this be?

Logically, one would expect PM10 emissions, which are a subset of total particulate emissions, to be the smaller of the two. However, this is often not the case when using two different test methods to make the two measurements.

Reasons for the discrepancies include: moisture condensation in the 201A cyclone; degradation of cyclone o-rings, resulting in contamination of sample; acid-gas artifact formation; and, probably most believable, the fallibility of trying to compare apples to oranges with the two methods. The fact is, the temperature of the filter defines what is and what is not particulate matter, and thus one would not (and should not) expect an in-stack filtration method (201A) to give comparable results to an out-of-stack filtration method (M5), unless the stack temperature is around 250‹F.