Ammonia MDL Study: Gas Diffusion & 40 CFR 136 Appendix B

An ammonia MDL study is how a wastewater laboratory determines the lowest measured ammonia concentration it can report with 99% confidence that the result is distinguishable from a method blank. For Clean Water Act work, that procedure is Appendix B to Part 136. The study is run on the laboratory’s own instrument, in that laboratory. It is not a brochure sensitivity claim, and it is not a drinking-water requirement under 40 CFR Part 141.

Manufacturer sensitivity claims, including a “single-digit ppb” capability, describe what the detector can reach under controlled conditions. They are not the laboratory’s method detection limit. The study also picks up standard preparation, reagents, and the laboratory environment. That is why two laboratories with the same gas-diffusion analyzer can calculate different MDLs.

Establishing the Ammonia Method Detection Limit

Table 2 of Timberline Method Ammonia-001 gives two calibration ranges, as ammonia-N. The standard 4-inch membrane is 0.050–500 mg/L. The high-sensitivity 10-inch membrane is 0.020–2 mg/L. The high-sensitivity range starts lower. It does not extend the top of the standard range. Neither figure is the laboratory’s MDL.

A single fixed method constant also does not transfer from one facility to another. Appendix B is the procedure for calculating the laboratory-specific value: the minimum measured concentration that can be reported with 99% confidence that it is distinguishable from method blank results.

Triggers for a New MDL Study

An MDL study is not finished once, at commissioning. Section 12.1.1 of Timberline Method Ammonia-001 states when a new study is required.

The study is required at initial start-up, before sample analysis. It is required again after maintenance or repair of the membrane, the conductivity detector, or another internal analytical component. Those parts are where the signal is produced. Changing them changes the baseline the study measured.

Consumable supplies and reagents do not, by themselves, trigger a new study. Replacing pump tubing or making fresh buffer and caustic calls for a new calibration. It does not, under section 12.1.1, call for a new MDL.

This study is also not the Initial Demonstration of Capability. The MDL study estimates the lowest reportable concentration. The IDC covers the wider precision and accuracy checks. Those belong in Validating Ammonia Analyzer Methods for Regulatory Compliance.

Preparing the Spike for Gas Diffusion Analysis

The spike level has to sit near the limit the study is trying to measure. Section 12 of Timberline Method Ammonia-001 calls for seven replicates prepared at three to five times the estimated MDL. A spike that is far above that estimate makes the standard deviation reflect preparation error more than baseline noise. A spike the instrument cannot see does not produce a usable study.

The same section says those replicates are prepared in “the Synthetic” matrix, and the sentence in the method extract breaks off there. Do not fill that name in from this article. Use the matrix named in the copy of the method the laboratory is following. If the permit names a matrix, confirm it with the primacy agency. A study run only in very clean reagent water can look tighter than the same study in the wastewater the permit covers.

The replicates still have to be prepared as ammonia-N, with the dilution controlled. Stock preparation and the as-N conversion are in How to Prepare Ammonia Calibration Standards.

Executing the Seven-Replicate Study

On a gas-diffusion analyzer, the peristaltic pump draws each replicate and mixes it with caustic. Timberline Method Ammonia-001 uses a working caustic that is typically 5% sodium hydroxide, strong enough to take the stream to about pH 11 or higher. At that pH, ammonium in the spike is converted to dissolved ammonia gas.

The stream flows along the outside of a tubular hydrophobic membrane. Liquid and ions stay on that side. Ammonia gas crosses the membrane into the absorbing buffer, about 250 ppm boric acid near pH 6.5. The gas converts back to ammonium, and the conductivity of the buffer rises in proportion to the ammonium that crossed. The detector records that rise for each of the seven replicates.

That spiked study is not the instrument detection limit. Section 12.2 calculates an IDL from three reagent blanks, as three times their standard deviation. The batch reagent-blank rule is a separate decision and is covered in When an Ammonia Reagent Blank Means the Run Stops.

Calculations under 40 CFR 136 Appendix B

After the seven replicates are read as concentrations, section 12 says to calculate the average and the standard deviation, then set the MDL equal to Student’s t at the 99th percentile times that standard deviation. The method points that step at the promulgated text of 40 CFR 136 Appendix B.

In the older seven-replicate procedure, the t value for seven results (six degrees of freedom) is 3.143. That figure belongs to that procedure, as written in the 2011 method. It is not a standing instruction to ignore a later Appendix B.

Appendix B is currently Revision 2, “Definition and Procedure for the Determination of the Method Detection Limit.” Revision 2 still uses 99% confidence against the method blank, and it also accounts for method blanks when those blanks show a signal, including in ongoing data collection. A primacy agency can require that revision. Do not treat the 2011 seven-replicate recipe as if it replaces the Appendix B text the agency enforces. Read that text and follow it.

40 CFR 136.3 identifies the approved test procedures. It is not the MDL calculation. Which ammonia methods are listed for Clean Water Act reporting is covered in 40 CFR 136 Table IB for Ammonia.

Reporting Limits: MDL, ML, and LOQ

The study produces more than one number. They are not interchangeable.

The calculated MDL. This is the result of the Appendix B study the agency requires. It is the lowest measured concentration the laboratory can distinguish from a method blank at 99% confidence. It is not, by itself, a precise quantitative result.

The minimum level (ML). Timberline Method Ammonia-001 sets ML as MDL × 3.18. That multiplier is this method’s formula for the ML. It is not the limit of quantitation, and it is not a claim about precision at that concentration.

The limit of quantitation (LOQ). LOQ is a separate calculation, often set by a state quality system or by TNI. Do not report the ML as the LOQ. The procedure for LOQ is in Defining Limits: How to Determine LOQ for Timberline Ammonia Analyzers.

Published detection levels from other methods are a fourth number. NEMI’s summary of Method 350.1 lists a detection level of 0.010 mg/L for automated phenate colorimetry. That is 350.1’s published level, for a method that needs a clear sample. It is not this laboratory’s MDL. EPA Method 350.1 remains the phenate reference. Gas diffusion keeps color and particulates off the detector, which is why many wastewater labs consider it. The membrane still fouls, and it does not make the MDL a catalog number.

Regulatory Constraints for Wastewater Compliance

Timberline Method Ammonia-001, gas diffusion with conductivity detection, is EPA-approved for ammonia. That approval does not extend to nitrate, nitrite, or total inorganic nitrogen.

A zinc reduction cartridge can reduce nitrate and nitrite to ammonium so the same detector can support a total inorganic nitrogen reading. The instrument does not speciate nitrate from nitrite. That use is for process monitoring and internal plant data unless the laboratory has a documented state variance or an approved alternate test procedure. Do not report nitrate, nitrite, or TIN on a Discharge Monitoring Report on the strength of the ammonia approval.

The MDL record shows what this laboratory could distinguish from a blank under the conditions of the study. It does not, by itself, approve a method, create an alternate test procedure, or set a nationwide detection limit.

Technical Support and Instrument Selection

Wastewater laboratories comparing options for turbid or colored effluent can review the ammonia analyzer as the instrument a study like this is run on. The study still has to be performed in the laboratory. A product range is not a substitute for it.

Timberline training and support can help a laboratory set up the initial study or the study that follows a membrane or detector repair. The agency’s current Appendix B still controls the calculation.

After the MDL is on file, each reportable batch still has to show that the result holds up in the actual sample. That check is Ammonia Matrix Spike Recovery.

References