I'm involved with a facility that discharges cooling tower water and has a permit with a total chlorine limit of 33 ug/L  (essentially non-detect).  The facility operators analyze for this parameter themselves  using a field portable filter photometer and the DPD method (Hach, but essentially Standard Methods 4500-Cl G).  I am beginning to conclude that the current test methods lack the sensitivity to get any reliable numbers down to this level.  

I performed a method detection limit (MDL) study last week and found that our water (at that time) had an average concentration of 56 ug/L, a std dev of 23.6 ug/L (n=7 replicates).  Crunching the numbers using the EPA's procedure (40 CFR 136 App B), this is an MDL of around 75 ug/L.

So here's my question for the group:

What are my options for decreasing the MDL for our measurements?  I suspect many of our numbers have been interpreted as false alarms.

Is there a more sensitive method available?  How easy is it to implement?  Can I modify the existing method to lower the MDL?

Thought?

Thanks much,

Matt

Thanks for the reply.

The detection limit estimate was one part of my method troubleshooting.  I also did a set of standard additions of a low range chlorine standard to wastewater and reagent grade DDI water (and got satisfactory recoveries).

Standard deviations of the spiked method blank set of samples (the lowest spike level was 250 ug/L) were much lower than my wastewater sample measurements, and if we used these numbers to derive the MDL, it would have been 34 ug/L - much lower.  I am hesitant, however, to use the data from the spiked method blank to derive the MDL, because our real world sample exhibited a much larger spread of measurements (and because the 250 ug/L number wasn't really at the low end of the measurement range - I had some difficulties preparing standard solutions at lower concentrations).

Are there reasons that I should be using the spiked method blanks to derive the MDL instead of our sample?  I do not have a ton of experience with 40 CFR 136 App B.

Lowering the detection limits for a spectrophotometric method is an interesting challenge.  Be aware that when you lower the detection limit, you will also lower the upper range.  The most commonly used method is to increase the path length of the cell that you are using.  This will allow a lower color concentration to exhibit a higher absorbance.  You can see this effect yourself by putting a drop of food coloring in a 1000 mL flask of water.  If you get a drop of that solution, you will be hard pressed to see any color in it.  But if you look through the side of the flask, you will be able to see the color.  The increase in path length reaches practical limitations to its implementation.  Most specs will accommodate up to a 10 cm cell. The next way to do it is to increase the concentration of the color developing reagent that you add.  This will intensify the resulting color that you get from the completed reaction.  This is not a 100% perfect solution because the intensity increase is not always a 1:1 ratio with reagent concentration.  Hach has a low level chlorine kit that implements both of these steps.  One of their portable meters has a flow through cell attachment.  They also have liquid versions of the DPD powders.  They are more concentrated and because they are a liquid, there is a better mixing of the reagent with the sample.  This lowers noise from optical diffraction of the sample.  It can see down to 15 ug/L of chlorine in the sample.  I don't work for Hach, nor do I sell for them,  but I do have this method in use and it works for us.

To answer your question about low limits, I would like to quote a paragraph from an NPDES permit issues by Missouri Department of Natural Resources. The permit limit is 17ug/L Daily Max and 8 ug/l Monthly Max.

This effluent limit is below the minimum quantification level (ML) of the most common and practical EPA approved CLTRC methods. The department has determined the current acceptable ML for total residual chlorine to be 0.13 mg/L when using the DPD Colorimetric Method #4500-CL G. from Standard Methods for the Examination of Waters and Wastewater. The permittee will conduct analyses in accordance with this method, or equivalent, and report actual analytical values. Measured values greater than or equal to the minimum quantification level of 0.13 mg/L will be considered violations of the permit and values less than the minimum quantification level of 0.13 mg/L will be considered to be in compliance with the permit limitation. The minimum quantification level does not authorize the discharge of chlorine in excess of the effluent limits stated in the permit.

... looks like Hach method 10014 is also known as method 8370.  

If we adopted this method for TRC, it'd be quite a bit of upfront money to jettison the portable colorimeter/photometer and upgrade to a photospectrometer with a flow thru cell.  

... not what I call the most "common and practical" of the EPA approved methods (a reference to BarfingCat's permit language).  ...especially for non-wastewater types.

I could really live with the 0.13 mg/L TRC number in BarfingCat's Missouri permit.  If were up to me, a quick copy-and-paste to amend our permit would solve this problem.

Thanks folks.

GThorney:  If you find out, let me know how Hach's ultra low level TRC method performs for you in Phoenix.

-Matt