QA/QC

Triangle NoTES — March 2016

Discussion of issues clients found with other sources for equipment and analysis during the period we were not an independent laboratory.

What has been so surprising to the people who, for whatever reason, had to use an alternative source for their analysis is the way things are done elsewhere.

The worst complaint was that the methods had been modified but that there had been no specific information provided for the modifications before the project was started. Once the results were submitted and questions were asked the modifications came to light. The problem was the modifications precluded the agency accepting the results thereby forcing a complete re-sampling of the source.

A close second was the dearth of information in the report, which consisted of essentially a table of results, a sparse narrative and a copy of the COC. This is a far cry from getting all of the data from area counts at the analyzer, response factors, dilution factors, sample calculations for every step and every piece of data for every sample should anyone wish to check the calculations. Over the years we have added more data to the report if even one regulatory agency requested it after receiving our report. We strive to have any data already in the report so if there is ever a question the answer is already available and documented. I know the program being use for calculations has been checked and the results verified by everyone here, a multitude of clients and regulators, and several auditors, but it never hurts to have just one more person follow up.

The one issue that completely surprised me was one I thought had been so completely discussed that everyone knew about it, which was making sure the return shipments were shipped legally and with as few problems as possible.

While it sounds a little counter-intuitive, if you meter 5 L of sample into a 6 L canister the volume of the sample is 6 L not 5 L. This is because the gas expands to fill the container albeit at a lower pressure. That is the reason we have the ability to determine the effect on the variables of pressure and volume when the temperature is constant. Using the formula P1V1 = P2V2 we can see how the pressure changes proportionally as the volume is changed. For example, a 3 L container with an absolute pressure of 760 mmHg will have an absolute pressure measurement of 380 mmHG if it is introduced into a 6 L container using this formula. The volume increased and the pressure in the container decreased. The inverse is also true, which means if the 6 L container had an absolute pressure of 760 the same sample in a 3 L container would have 1520 mmHg of pressure. The DOT defines volume of a gas sample as that of the container in which it is shipped as long as the absolute pressure of said gas is less than 15.22 psia per 49 CFR 173.306 (a)(4). Although most law enforcement agencies are not aware of the regulations or fines, the DOT has allowed anyone authorized to issue a travel citation to issue a citation for improper transportation of sample containers since 1988. This includes any state, county, city, or town police officer which may for any reason stop a vehicle for a travel violation. This includes personal and company vehicles in the potential for receiving a fine, but still not as likely to be fined as with a common carrier either.

 

Triangle Environmental Services, Inc.

Wstollings@aol.com

(919) 361-2890        (800) 367-4862            FAX: (919) 361-3474

Triangle NoTES - October 2011

With progress on the Method 25 audits hopefully resulting in commercial products in the next few months, I have information on the new audit sampling interface we developed and some insights on the extra QA/QC we perform and why we feel it is necessary.

First, an important last minute addition. I have recently been informed by UPS that any shipment charged with the special handling fee does not get a service failure guarantee. Therefore, any shipment not in a cardboard box essentially has no delivery guarantee from UPS. This is evidently a change made about the first of the year and would only be found if you claim a service failure and are told the shipment does not qualify due to the special handling. This is something that could be very important to know in our industry.

The new audit interface is essentially an extra sample console which is connected between the audit cylinder and the field sample console. Our version uses a critical orifice to maintain constant flow, which is only adjusted by adjusting the pressure regulator setting. There is a flow measurement going into the system and another coming out of the system to the vent. The outlet reading should decrease by the amount being diverted to the field sample console as the audit sample.

There will be two interfaces available for rental when audit cylinders are again available. If more are needed, we can at least double that number fairly quickly. As with all of the equipment we provide, the audit interface will be cleaned and analyzed to confirm cleanliness prior to shipment to the field, even though there should be little chance of persistent contamination by the audit gases.

To check the audit sample interface for correct function, I used fixed gases as the compounds of interest instead of the organic compounds used for actual audits. The fixed gases allowed for the interface and only the interface to be tested before any more detailed tests of audits or the collection thereof were performed. The error between the expected concentration of gas (+2% of reported by provider) and the analyzed concentration was <0.25%, which indicated not only the interface was working correctly , but also the variation of the gases when compared to each other were much less than the guaranteed 2% maximum. If anyone decides to construct their own interface, I would strongly suggest they perform a similar test using something like a methane standard, which can then be analyzed with an FID to determine proper function prior to being used for audit sample collection.

This ties into the next item I wanted to discuss, the QA/QC performed on the various pieces of sampling equipment and what it actually means. The information on the “Chain of Custody for Supplies” details the levels to which the equipment is checked under our SOPs, but that information is a little less specific in scope for brevity.

We have performed a series of QA/QC analyses on the equipment from the very beginning because of my experiences with possible equipment contamination with the old Research Triangle Laboratories (RTL) procedures. The traps were cleaned as part of the recovery of samples as specified in the method. That was a source of potential problems in my mind because there were a

lot of different sources being tested and there seemed to be at least a possibility of cross contamination, especially from the additional trap used as part of the recovery process. This was evident by the problems with audit pass rates at the time, which were the only means to indicate such issues. Here at TES, we moved to a separate cleaning and analysis for each and every trap, which clearly seemed to help to improve the pass rate of the audits at the time. We used a methanol and water injection through the trap followed by baking in an oven to a temperature 100 degrees C

higher than the temperature at which the traps are recovered for a period of at least eight hours. This is followed by an analysis of the trap at a temperature 50 degrees C higher than our standard recovery temperature using the same type of catalytic FID used for the analysis of the samples. This helps to ensure there is no cross contamination between trap fractions, as the recovery of samples should never approach the temperature levels of the cleaning ovens nor even the QA analysis. The use of the same type analytical system follows the method requirements and helps to ensure the consistent analysis even though a TCD or NDIR detector should provide similar results as the method analyzer specified. The method allowance of <10 ppmC is mirrored in the SOP, but the actual level of cleanliness achieved is usually 60 - 70% lower than that. There are some cases where the sources are so heavily laden with organics that even with the improved clean up and higher analysis temperatures multiple cycles are required to meet even the method requirements. For those cases, we have retained the higher limit for consistancy in the SOP. In reviewing the data generated by our QA/QC analysis of the rental equipment we purchased when RTL was shut down, a surprising number of traps, which were supposedly ready for sampling, did not meet our or even the method requirements without additional cleaning.

The SCAQMD method of cleaning traps uses a much higher temperature than we use. There is no temperature specified, but the trap is to be made “cherry red” by direct heating. This poses a potential problem according to our experiences and the expert information we have sought on the subject. It seems that if stainless steel is heated to a temperature in excess of 400o C and cooled, flaws can begin to develop in the structure of the steel where carbon migration forms veins of weakness. These veins are extended and expanded each time the temperature cycles above and then returns below that 400o C level. When these veins eventually fail the trap becomes permiable to gases, including ambient air and the samples within the traps. Thus a failure would cause the loss of the sample or at the very least some level of contamination. Metallurgy is not my field of expertise, so I rely upon those who are experts in the fields for this advice. However, our first hand experiences with stainless steel failures does support this warning. As a precaution, all trap ovens of any type are kept well below 400o C to lessen any potential impacts to the integrity of the trap structure.

The field sample units, both heated and unheated, are each analyzed using the trap QA analyzer. The associated adapters used on sampling equipment are also individually analyzed prior to use. The actual measured concentration of each unit is generally well below the allowance, but again there are sources which are nearly pure organic compounds which would make lowering the limit a little more difficult and a little more expensive. None of the sample units purchased from RTL passed the requirements of our SOP, including those which had been on the shelf ready to be used for sampling. I firmly believe this was one of the problems with the method at that time. Not only were the sample units assumed to be clean after being processed, there was a deviation from the method, in which the temperature measured in the filter box was the average of the gas stream and the filter box, instead of only the gas stream measurement as specified in the method.

The sample tanks have two requirements for cleanliness. The first requirement is there be no detectable NMOC, which translates to ~2 ppmC for a sample using ~3X the MDL. The second requirement is the remaining carbon, either CO2, CO, or CH4 is less than 10 ppmC combined. This is because the tanks are cleaned via a mass dilution method and with very high concentration sources there still may be a little of these gases left after a series of cleaning. At this level, when the tank is evacuated for sampling, the remaining concentration should be taken below the detection limit for those compounds for the sample analysis. There is no allowance for NMOC because it can generally be removed through the normal cleaning process. If the tank cannot be cleaned using two of the cleaning processes, it will require a very intensive cleaning that usually is effective. The only way to know if the tank is clean is to analyze each one, because lingering contamination can be introduced from any level of source.

The tanks purchased from RTL were cleaned via mass dilution while being heated in an oven similar to a vacuum oven in usage. The tanks were again assumed to be clean after such a process, but again when we analyzed the individual tanks supposedly ready for sampling there were several which failed our SOP requirements until additional cleaning was performed. While the majority of tanks may be cleaned by any process, unless you verify they are clean by individual analysis, you are taking a risk. If you have the one sample tank where the contamination actually makes a difference between a pass and a failure, it does not help to consider the odds were in your favor for nothing to have been wrong.

If the range for the low level Method 25 audit approaches 50 ppmC as was seen in some of the previous program’s audits, the allowable contamination may make the difference between passing or failing that level of audit unless the range is widened considerably. For example, the window for passage of an audit in the 50 ppmC range was + 10 ppmC, which is easily exceeded by a slightly dirty tank, trap, and/or sample console. In fact, we have seen high level audits fail the + 20% range because the sample console was assumed to be clean after processing and was later found to have still been contaminated.

So, while it is true we could reduce our individual certification analyses and thereby reduce the costs of the project, it is also true that you get what you pay for ... and you cannot assume you will get what you did not pay for. The major cost differential between assuming cleanliness and being able to certify cleanliness is not as great as the major cost differential between a good set of data and a questionable set of data. The insurance of certification helps to prevent problems across the board.

Wayne Stollings

Triangle Environmental Services, Inc.

Wstollings@aol.com