                            EPA and Ecology comment
                           Emerald-Kalama Delisting
                           Proposed Cobalt analysis


 The proposed method seems to have a brief summary procedure of the extraction procedure. The SW-846 Methods 1311 has considerably greater detail, along with a number of very important technical considerations, that EPA and Ecology believe are needed to ensure exraction is appropriately and consistently accomplished. We recommend that the draft procedure remove those elements related to the extraction procedure, and instead state that extraction will be carried out according to SW-846 Method 1311. EPA and Ecology remain supportive of the balance of the procedure for analysis of the extract, subject to the additional comments below.

      Analysis starts with TCLP (Toxicity Characteristic Leaching Procedure) extract preparation from the SOMAT sludge.  Extract preparation (TCLP extraction) must strictly follow the SW-846 Method 1311 (https://www.epa.gov/sites/default/files/2015-12/documents/1311.pdf). Keep in mind that quantitative transfer of all cobalt from the sludge into solution is not objective of the procedure. Important context:
      
 Volatiles are not of interest, cobalt is the only analyte of interest, follow suitable parts of the Method 1311 accordingly
 Pay attention to the selection of the extraction fluid (acetate/acetic acid buffer or acetic acid solution) based on sludge properties
 According to robustness study presented in the Method 1311, control of the pH of the extraction fluid is one of the most sensitive parameters of extraction. Make sure pH measurements conform with the method requirements.
      


 Under "Reagents, solutions, solvents," we recommend that the procedure include a Materials and Equipment list for the SOP (balances, pipettes, spectrophotometer, etc).

Materials and Equipment

pH meter (as per Method 1311 or better, capable of +- 0.05 pH unit accuracy, currently udsed: Beckman Φ-3)

Analytical balance, 0.0001 g resolution, with current certificate of calibration (currently used Mettler Toledo AB204-S)

Spectrophotometer with a monochromator capable of absorbance determination in the visible range, currently used: VWR UV-1600PC spectrophotometer
Rotary agitation apparatus as described in Method 1311

Filter paper: any filter paper for chemistry applications with ash content < 0.06% (ash content of 0.01 preferred) currently used: Whatman 42 Ashless

pH meter calibration solutions (pH 4.00 , 7.00, 10.00)

3 mL disposable syringes

0.45 μm PTFE syringe filters






 Step 1 reading " ". Unless there is something unique about Millipore DI water, might want to say "Millipore DI water or equivalent." Further, the method should include a specification something like the DI water must have >18 mega ohm resistance reading so that the quality of water used is consistent.


.. high-quality DI water (resistivity at 25 °C not less than 18 MΩ::cm = 1.8x10[5] Ω::m measured at the source, resistivity will decrease rapidly upon absorption of atmospheric carbon dioxide; in practice Millipore DI water or equivalent) stored in dedicated poly bottles..




 Step 1 in the calibration procedure mentions use of Erlenmeyer flasks. Erlenmeyer flasks are not appropriate for making standards, or for any other quantitative purpose. Appropriate volumetric flasks should be specified. This comment also applies to the Procedure Step 1.


Changes will be made from "Erlenmeyer flasks" to: "suitable volume, class A volumetric flasks".








 The calibration procedure Step 7 discusses filter paper. What type of filter paper is to be used?  Please add to the requested equipment list.

Filter paper: any filter paper for chemistry applications with ash content < 0.06% (ash content of 0.01 % or lower preferred) currently used: Whatman 42 Ashless paper.

 The calibration procedure Step 8 discusses specific light wavelengths. Please clarify if is 456 nm the analytical wavelength and 500 nm the background. Describe how these wavelengths are used to generate absorbance results.

Both wavelengths are analytical: 456 nm is the absorption maximum. Both wavelengths are used to create two calibrations. Results from two calibrations are averaged, so far agreement between values obtained from the two calibrations has been excellent. Averaging two values guards against human error (entering incorrect absorbance for one wavelength would generate discrepancy that would be easily noticeable). It serves also as a partial test of the identity of the spectrum of the analyte  -  reproducible discrepancies would signify interference that was not previously present (e.g., other species absorbing in this region). Method can be easily modified to include only measurement at 456 nm, reducing somewhat time and amount of work needed per analysis, but also removing benefits described above.


 The calibration procedure Step 11 discusses calculation of calibration lines. A criterion for correlation coefficient (e.g., >0.99) or some measure of quality of calibration to be met would help maintain consistency for accuracy.  Also, a frequency might be needed to state such as daily before samples. EPA notes that Spectrophotometers are calibrated each day of use as general lab practice. Maybe just testing a fresh mid-level standard to confirm the curve or instrument performance is still good like a calibration check.  Would also need criteria for such a check such as within 90-110% of expected level otherwise re-calibrate. Finally, a second source calibration verification standard is typically required. Should be listed in reagents list and the analysis described.

Measure of calibration quality. For each non-blank color standard  i calculate actual cobalt mass (μg) delivered during standard preparation (from quantity of diluted cobalt standard, and known cobalt concentration, related to the primary standard cobalt solution): miactual. Using calibration relationship parameters (currently intercept and slope for linear least square fit) calculate cobalt mass for each standard based on observed absorbance: mipredicted. Calculate relative residual εi for each standard εi=miactual- mipredictedmiactual, and its RMS value: RMS(ε)=i=1Nεi2N . Recent observed RMSε was 2.5%. 

Proposed calibration quality measure: RMS(ε) for each calibration used will not exceed 4.0 %.


Calibration check: standard that delivers 2.5 to 3.0 μg of cobalt (relatable to the primary cobalt standard) prepared and measured daily nearly simultaneously with the sample. Re-calibration will be required, if calculated value deviates more than +-7.0 % from the actual value.

Primary cobalt standards:
 Cobalt Reference Standard Solution 1000 +- 10 mg/L, Hatch, Cat. Number 2150342 (currently used).
 Cobalt Standard for AAS, TraceCERT(R), Supelco, Aldrich catalogue number 05202-250ML (to be acquired).





 As a general comment on the Procedure itself, what standards are being used for QC?  With what frequency are they being run?  What are the acceptance criteria? In general, EPA and Ecology look to the method description to include appropriate laboratory QA/QC procedures and criteria. These are essential to validate the resultant data.


Quality control measures as described in (7) will be incorporated into the method. In addition, full calibration will be performed every six weeks. 


 Procedure Step 5. How exact are these times? Does color development continue or plateau? Is the color maintained over time or does it peak and decline? Since this process is being done manually, and presumably standards and samples are being created as a batch then measured as a batch, some description of timing controls may need to be included.

The reaction times for color development are based on the literature describing properties of the reagent used and its reaction with cobalt ions, and analytical methods developed around it (Kyoji Tôei  and  Shoji Motomizu, Properties and Uses of the Colorimetric Reagents 2-Nitroso-5-dimethylaminophenol and 2-Nitroso-5-diethylaminophenol for Cobalt, The Analyst, 1976, 101,497-511, DOI https://doi.org/10.1039/AN9760100497, and references therein). Formation of the complex requires proper pH range and sufficient time. This time was investigated by the authors of the original publications, additional time is added as a precaution in the current method.  Once formed and extracted, the complex is stable in the chlorinated solvents (references cited above). This was confirmed by our practice: an extract of the complex that measured 0.620 absorbance at 456 nm and 0.499 absorbance at 500 nm at preparation, measured 0.624 absorbance at 456 nm, and 0.501 absorbance at 500 nm after five days (increase could have been due to solvent evaporation loss).

 Procedure Step 9. Specific concentrations should be listed which would trigger reanalysis (points below lowest non-zero standard?) Detection limits should be determined and compared to level of interest to document that the corresponding data are suitable for their intended use.

Limit of quantitation, LOQ, is perhaps of greater interest. Proposed definition: 

LOQ=10Sba

Where Sb is standard error of the intercept of the calibration line at 456 nm, and a is the slope of the calibration line at the same wavelength. Values for the current calibration (a, Sb obtained using LINEST function of Excel) give LOQ of 0.51 μg. Calculated cobalt content of 0.50 μg will require re-analysis and increased sample load. Sample load is variable, for the maximum load of 30 mL, LOQ of 0.51 μg corresponds to concentration of 0.017  μg/mL. 


 Procedure Step 11. Please provide equations used for calculation.  Does the instrument perform any calculations or are all calculations done offline?


Calculate mass of cobalt contained in each analyzed sample from calibration for each wavelength:
                            m456=A456-b456a456, μg
                            m500=A500-b500a500, μg
where:
A456	observed absorbance for the extract prepared from the sample (at 456 nm)
b456	intercept of the calibration line at 456 nm
a456	slope of the calibration line at 456 nm (unit of 1/μg)

A500	observed absorbance for the extract prepared from the sample (at 500 nm)
b500	intercept of the calibration line at 500 nm
a500	slope of the calibration line at 500 nm (unit of 1/μg)

Next, calculate the average:

                              mav=m456+m5002, μg

Cobalt concentration in the sample is:

                               cCo=mavMs, μg/g
                              c'Co=mavVs, μg/mL

where:
Ms	extract sample mass, g
Vs	extract sample volume, mL

Measurement of extract density will be added to the procedure to facilitate conversion between units.

All calculations are done offline.


