Introduction

USEPA wet chemical methods that are approved in 40 CFR Part 136 are for the most part actually determinations and not analysis (CO Ingamells and Francis Pitard “Applied Geochemical Analysis” page 191). By this I mean that they are trying to identify traces of analyte within the matrix with almost no preliminary separation. The methods are written and include, not only the reagents necessary to create a measurable final product, but also complex reagents and buffer reagents formulated to adjust the pH and prevent interferences from occurring. It is impossible for one method or one set of conditions to work on all arrays. Since wastewater is so complex, wastewater methods must allow users to modify the methodology to work with their samples.

Rapid chemical methods, with as few steps as possible, are the only way semi-skilled or unskilled analysts can analyze thousands of samples. A quick method works fine on the array it was tested on, assuming it is followed exactly.

So why do we have method flexibility? We have it because we asked for it. The ability to modify a method has been built into the QA/QC section of various methods and has been under the ‘heading’ of performance based for years. Later, the flexibility of the method was added to the 40 CFR as Part 136.6. Reasons for modifying a method range from changing from one instrument to another to removing interference from the matrix.

Rapid chemical methods and environmental tests

Most countries have adopted rapid chemical methods for use in environmental regulations. The logic is that there is a maximum allowable limit for a toxin and we will measure it. Industry and government agencies invest large sums of money based on the numerical results of these methods. Because the methods are promulgated, the results, whether correct or incorrect, determine compliance. Failure to follow the SOP will put you out of compliance simply for not following the SOP word for word.

I have written numerous articles on total cyanide by distillation. It is possible to test a sample for total cyanide, get a positive result, be fined for releasing cyanide, and the cyanide not be there. A simple modification of the approved method reports the correct result, however, previously the modification of the method was considered a violation even if the modification produced the correct result.

In many cases, following the SOP is fine, however, without modifying the methods to handle as much interference as possible, you’ll never know. Total cyanide is interesting because distillation generates the cyanide. Spike recoveries and duplicates will be acceptable, even though the cyanide isn’t actually there.

Letters of Approval and the Alternative Test Procedure

So why add method flexibility to the CFR? There are Alternative Testing Procedures (ATPs) that need to be properly evaluated before these procedures can be promulgated. A method that simply modifies an approved method is not an ATP.

EPA’s ATP Coordinator was overwhelmed with requests for approval of discrete test methods. Previously, this so-called approval occurred as powder sachet and flow injection approvals instead of segmented flow. Part 136.6 allows users and method generators to change methodology without EPA review or rule making, as long as the minimum criteria are met.

If you look closely at the wording of these “approval” letters, they are not approvals at all, but an acknowledgment by the EPA that the method in question is essentially equivalent to the approved one. The EPA decided that the ATP coordinator’s time was better spent seeking approval of new alternative test procedures than providing instrument manufacturers with letters to use as marketing material. The deal breaker was pretty much the discrete analyzer, which is essentially a new way of adding reagents to samples and doesn’t really change the methodology at all.

Another reason for the flexibility of the method is the rigidity of the SOP. The EPA’s previous methods were written under the assumption that chemists would use them. Over the years, as we all know, they became used less by chemists and more by lawyers. The mindset of following the SOP word for word developed and reached a tipping point. As we try to measure lower levels, we need to be able to adapt our current methods to allow us to see these lower levels. Not allowing a 25 milliliter purge volume because the method only says five milliliters, or not allowing capillary columns because the method requires a packed column is silly and slows progress. There is not enough time, and certainly not enough of us, to revalidate every method for every foreseeable change. We must allow experienced users to adapt existing methods to serve modern purposes.

The inability to alter a method hinders progress

There are thousands and thousands of potential arrays, and in them an infinite supply of potential things that can cause a method to not work. The method developer tries to compensate for everything, but this is not possible. What do we do with emulsions, foaming or analyte suppression? What do we do when we know that the approved procedure generates biases and we are reporting erroneous results? If our purpose is to protect the environment, are we really protecting it by generating incorrect data? What is fraud, generating wrong data by following the method, or not following the method and getting the correct result? Is it okay to modify an extraction to overcome the interference and then not run the whole process on each sample? Is it okay to mark the data as failed in QA, but accept it because it passed an LCS?

The best answer would be that all of our methods work perfectly on every sample. We can ensure this by allowing experienced chemists to modify methods while ensuring basic performance criteria are met. If a modification, such as changing a complex reagent, results in each sample getting results closer to the true value, isn’t that better than marking 10% of them? Time to put some chemistry back into chemical analysis. We are not implying a flagrant and random disregard for the SOP. We are saying that the data should matter more than the SOP. It is more difficult to modify and document an existing method to overcome interference than it is to simply follow the SOP. The criteria are not easy and assume that interferences are identified, characterized and analysed.

The purpose of method flexibility

The purpose of method flexibility is to improve quality control. The purpose is to obtain better data. The purpose is to have no data qualifiers. The purpose is to allow you to overcome the interference. The end result should be the correct result. A method cannot work on all arrays. It’s just not possible. The flexibility of the method is improvement.

These are reasons to modify the methods. We can improve accuracy and precision, we can get better method detection limits (MDLs), we can compensate for interferences, we can increase throughput, we can use “green reagents” and we can cover the analytical range that is really needed. These things can be done, but only by modifying existing methods. There is no time or resources to revalidate all methods after making small changes. If we change a buffer reagent, for example, why not just show that the new buffer works the same or better than the old one?

The criteria for a modification of the 40CFR part 136.6 method is equivalence of results. In other words, do you get the same result before and after a modification? The point, however, is that these equivalent results are obtained on sample data without interference. In a way we are comparing apples and oranges. We can’t just be creating more apples. We want our new method to work better than the old one.

equivalence test

To demonstrate equivalence, or to illustrate what is different, the method builder prepares a two-column comparison between the approved method and the new one. For most items like scope, detector, etc. the language will be the same. For method flexibility, differences will be throughput, instrument, column, sample volumes, and some of the reagents as long as criteria 136.6 are met.

The most important aspect of method modification is that quality control criteria are met. The QC must meet or exceed the criteria given in the approved method. If you get certain MDLs and response factors with a cycle time of 30 minutes, do you get similar data when the time is shortened? If you reduce a volume from 500 ml to 50 ml, is the precision the same?

Method flexibility gives us, for the first time in a long time, the opportunity to make those changes to methods that we know need to be changed. We can modify the pH, or complex reagents, or buffers, and even some reagents in order to obtain better data. We can automate a manual method without getting a letter of permission. 40 CFR part 136.6 should be seen as a good thing.

Conclusion

In trace analysis, the analyte is present in very small amounts relative to the rest of the sample matrix. It is literally looking for the needle in a haystack. A quick chemical method finds the needle without removing the hay. Lowering the CDM is akin to finding even more hay for that needle. At the same time, we want to find it faster and with the least expense. Of course, to use this analogy with accuracy and precision, or range, we end up looking for multiple needles in the same haystack.

(Our calibration standards or QC samples, prepared in pure solvent, look for a little of something in a lot of nothing, whereas our samples contain a little of something in a lot of something else.)

Method flexibility is necessary because method developers cannot anticipate all possible interferences or anticipate all matrices in which the method might be applied. Blindly following a method that is not working strictly to stay “in compliance” defeats the purpose of the measurement, i.e. getting the correct result, not just following the method. Remember, not all instruments are the same. A method developed on one instrument may require slight modifications to work on another. This is especially true when the analytical technique is changed. The flexibility of the method should allow us to switch to less toxic reagents and still measure the same final product. So we could actually do our monitoring with “green” chemistry instead of using toxic chemicals to monitor toxic chemicals.

I think the flexibility of the 40 CFR Part 136.6 method is the light at the end of the tunnel. Finally, we are not required to follow rigid SOPs. You can modify your instruments and methodology to obtain accurate and reliable results in a cost-effective manner. It’s not about shortcuts. It’s about being scientific, once again. Analysis is all about the data. It has to be defensible, but it also has to be correct. I realize that the commercial lab is hard-pressed with time and resources to make their own modifications. Therefore, the burden of proof falls on the manufacturer of the instrument. It is in the manufacturer’s best interest to ensure that the instrumentation and methodology works for its intended purpose. The manufacturer’s best interest is to work closely with the EPA to verify that modifications to our method are acceptable. The manufacturer will collect the data and provide all necessary information that proves, not acceptance, but CFR equivalency.