Nitrogen Testing and Alkalinity Control
Updated: May 16
In this post I focus on aspects of laboratory testing and plant operation that relate to nitrification, a topic I am endlessly fascinated with! Though nitrification is a wastewater process in place at many treatment plants around the world, which means there are many operators quite expert in this process, I still consider nitrification to be more on the complex side and for some, therefore, not well understood. In addition, the testing required to do a nitrogen mass balance can be cumbersome due to the number of test procedures, concentration range options for each nitrogen form (explained below), and potential interferences with the testing (also explained below).
The majority of my laboratroy work done in the field is performed using Hach’s TNT (test-in-a-tube) chemistries (field-friendly chemistries!) and that includes the full range of nitrogen components illustrated in the graphic below.
One of the problems I have, particularly when testing at an industrial watewater plant, is to figure out the dilution factor I need to use for the total Kjeldahl nitrogen (TKN) test. Hach's test for TKN is an involved, multistep procedure that requires 30 minutes of digestion and spans the narrow concentration range from 1 to 16 mg/L N. (Refer to Hach's TKN test procedure at the bottom of this post.) I can't think of a single wastewater plant I've done testing at with an influent TKN concentration of 16 mg/L N or less (see M&E influent nitrogen table 3-15 below). Given the narrow TKN test range, I always have to test diluted samples. If I get the dilution wrong, and end up with a test result over-range, I have to do another, larger dilution, and test again, which dramatically increases the time spent testing. Trying to find the right dilution ratio is not as easy as you might think because you want to keep the ratio as low as possible to improve accuracy in the test result. So, much of my first day onsite is spent doing the critical task of range finding for different parameters (COD, chloride, phosphorus, sulfate, etc.).
Recently, while spending 10 days onsite doing testing at the wastewater plant shown below, I was given a comprehensive data set produced by the operations and laboratory staff at this well-run, well-maintained, municipal, conventional activated sludge plant in New York. This excellent data set spans several years, which allowed me to do a detailed analysis of numerous aspects of the plant's operation. As a side note, there wasn't a single piece of trash on the grounds of this wastewater plant. You might think this is a trivial observation but I assure you it is not. Little details like this matter and they inform you as to just how much the staff care about their plant.
The graph below is a histogram of influent TKN data from this municipal wastewater plant. The mean TKN is 41.3 mg/L N and 98.1% of the data falls in the range of 16 to the maximum data set value of 58.3 mg/L N. Only 1.9% of the influent TKN values are less than 16 mg/L N!! Dividing 58.3 (the dividend), by 16 (the divisor), produces a value of 3.6 (the quotient), directing me to a 4 to 1 dilution ratio as my starting point for the TKN test. In this situation I had key data available from the customer, which, unfortunately, is often not the case.
Nitrogen Concentration Ranges and Interferences
Each form of nitrogen testing using Hach's TNT reagents has a range of concentrations as well as chemicals and compounds that can interfere with the testing as illustrated in the graphic below. The chemical oxygen demand (COD) and chloride concentration can be an important factor influencing the nitrogen test, particularly in industrial wastewater systems. Municipal plants use 5-day biochemical oxygen demand (BOD) to determine the organic load entering their biologcal system whereas industrial plants use COD. As a result, I typically have to measure the COD (wide range of Hach TNT reagents available) and chloride concentrations (Hach test strips) before I begin doing nitrogen testing. The COD test using Hach's TNT chemistry requires two hours of digestion, testing that must take place before beginning nitrogen testing.
From the graphics above you can see the only way to avoid doing TKN testing is to test for all other nitrogen forms and then subtracting as shown in the equation below. I take this approach often becasue I can usually expect the influent nitrate and nitrite concentrations (15 minute tests) to be low and Hach's total nitrogen test goes as high as 100 mg/L N, increasing the likelihood I'll measure the correct total nitrogen concentration without having to worry about using a diluted sample. (See Hach's total nitrogen test procedures at the bottom of this post.)
Let Me Contradict Myself!
If I talk too much, or write too much as I'm doing here, I often find I have to contradict myself. I used to be so embarrassed when this happened but after all these years I'm okay with it. So let me go ahead and contradict myself now. In the graphic below you can see I had an industrial sample with very high COD and nitrogen concentrations. Where do I begin in trying to determine the required dilution ratio? For this testing I did not try to measure the TKN concentration, using the approach just described above, instead. And, fortunately, I was "vaguely" forewarned by the customer I had a "high-strength" sample.
Arriving at the 2,202 mg/L N total nitrogen concentration ended up taking eight iterations (seven failed attempts) and ultimately required a 25 to 1 dilution ratio to get the result. Current pricing (as of May 2022) from Hach for their high range Total Nitrogen (20 to 100 mg/L N) TNT 828 reagent set is $108.25 for 25 tests, or $4.33/test. If you figure about 45 minutes in total to: 1) prepare the dilution, 2) heat (digest) the sample for 30 minutes in a digital reactor, 3) cool the sample for another 10 minutes, and 4) finally measure the concentration in a spectrophotometer, to complete the eight iterations took me six hours at a reagent cost of $34.60 just to determine the total nitrogen concentration in this single sample. I'm not including the nitrate, nitrite, ammonia, chloride, and COD tests that were also done, nor am I including my labor cost.
Getting the COD result (2-hr digestion required) was actually easy because I used Hach's UHR+ TNT 824 reagent which gives COD concentrations in the range of 5,000 to 60,000 mg/L. With industrial samples I always run one COD at this ultra high range and another sample using the TNT 823 reagent with a concentration range of 250 to 15,000 mg/L.
As you can see in the equation below, the math from this nitrogen testing worked out perfectly! Yeah, right, you are thinking and so was I when this testing was done. I never get perfect balance when doing the complete nitrogen analysis. I was surprised when I got these results and knew immediately I couldn't wait to share them with you because this will never happen again!
Before moving on, I want to point out the potential interference issue in doing the nitrogen testing. In an earlier graphic I showed that a COD concentration of 2,500 mg/L and above will interfere with the total nitrogen test in the 20 to 100 mg/L N range. The careful reader will have noted the 25 to 1 dilution required to finally obtain a nitrogen value also resulted in a diluted COD value of 1,100 mg/L (27,509 mg/L/25), eliminating COD as a source of interference in the vial. The undiluted wastsewater sample also had a chloride concentration of 554 mg/L, eliminating chloride as another potential source of interference. Interfering ions, chemicals, and compounds need to be considered when doing this testing to be certain you have a vaild result.
Alkalinity Control to Optimize Nitrification
The nitrogen and alkalinity data used in this post are from a municipal wastewater treatment plant with a substantial industrial contribution representing as much as 50% or more of the influent flow and loading. Because of the large industrial contribution, this plant needs to add a chemical product to the aeration influent for alkalinity addition to optimize nitrification. Over the years various chemicals have been used to increase alkalinity, including sodium hydroxide, magnesium hydroxide, and calcium carbonate.
An important point I am going to make in the discussion that follows is this: When you need to feed a chemical to increase alkalinity to optimize nitrogen oxidation, your control parameter for alkalinity dosing should be the final effluent alkalinity concentration, not pH alone. I fully understand how easy it is to have continuous pH monitoring in the effluent. All that is needed is a simple, relatively inexpensive pH probe that requires low maintenance (cleaning and calibration) in contrast to more expensive and complex online alkalinity measurement. The solution is to measure the effluent alkalinity concentration at least three times per week using either grab or composite samples. Alkalinity testing is easy and fast, taking just five minutes to complete. Of course, I use Hach's TNT reagent set for this test (see Hach's total alkalinity procedure at the bottom of this post).
Influent wastewater characteristics for several key parameters are tabulated below.
Effluent values for several key parameters are tabulated below.
In the next table we can see treatment plant performance as determined by the percent reduction in BOD, TSS, TKN, ammonia, and alkalinity. BOD and TSS removal rates are excellent at 88.1% and 91.1%, respectively. TKN and ammonia removal rates are not nearly as high, coming in at 43.9% and 44.4%, respectively.
You expect a reduction in alkalinity in a wastewater plant that nitrifies. And you can anticipate that some alkalinity will be recovered if denitrification occurs. For me, it was the extent of the alkalinity reduction that was so surprising, with an average reduction between the influent and final effluent of 46.5 percent.
“During the conversion of ammonia to nitrate, mineral acidity is produced. If insufficient alkalinity is present, the system’s pH will drop and nitrification may be inhibited. Approximately 7.1 mg alkalinity is required for each mg of ammonia nitrified. Almost half of the alkalinity, however, may be recovered through denitrification. To maintain the system’s pH, it is recommended that a residual alkalinity of 50 to 100 mg/L be provided. Water Environment Federation. “Operation of Municipal Wastewater Treatment Plants.” (Manual of Practice—MOP 11. 5th ed. Volume 2. Alexandria, Virginia: Water Environment Federation, 1996.)
In the histogram below you can see this wastewater plant fell below 50 mg/L of effluent alkalinity, over a three year period, just 5.6% of the time. They operated in the “recommended” alkalinity range of 50 to 100 mg/L 32.2% of the time. But, and this is a big "BUT," alkalinity in excess of 100 mg/L as CaCO3 occurred 62.1% of the time, with a mean alkalinity concentration of 128.7 mg/L as CaCO3, representing a significant opportunity for cost savings if greater focus and tighter control over the effluent alkalinity concentration had been maintained. This plant uses their effluent pH as the control parameter to determine the alkalinity dosage rate into the aeration system. Using pH alone will work when feeding sodium hydroxide but pH by itself is an inefficient and unreliable control method when feeding either magnesium hydroxide or calcium carbonate. You must check effluent alkalinity and use that result to set the chemical feed rate.
Though variability in the effluent alkalinity concentration is evident in the graphs above, the pH variation through this wastewater treatment plant is small, as shown in the two graphs and summary table below.
pH Is a Poor Predictor of Alkalinity
It is somewhat counterintuitive but the "tight" pH ranges across the wastewater system actually provide little insight into how the alkalinity concentration is changing relative to changes in pH. A more formal statistical review follows using fitted line plots created with the Minitab statistical analysis program (which I highly recommend). The three plots are: 1) Influent Alkalinity vs. Influent pH, 2) Primary Effluent Alkalinity vs. Primary Effluent pH, and 3) Final Effluent Alkalinity vs. Final Effluent pH. The plots fit the log of alkanity to match the logarithmic scale of pH.
The fitted line plots tell (and show) us how much the pH concentration (hydrogen ion concentration or, really, the hdyrogen ion activity) explains variation in the alkalinity concentration. We can simply observe this in each plot but it is helpful to look at the R-Sq value Minitab produces for this analysis. In the first plot, it's easy to see at a glance there is no relationship between influent alkalinity and influent pH and the R-Sq value of just 0.2% verifies how poorly the model fits the regression line. It is also interesting to note the slope of the regression line is nearly horizontal and this slope is going to change (unpredictably) in all three plots. Minitab states that "R-Sq is the percentage of variation in the response (alkalinity) that is explained by the model. The higher the R-Sq value, the better the model fits your data. R-Sq is always between 0% and 100%."
In the second plot of primary effluent alkalinity vs. primary effluent pH we see a small improvement in the R-Sq value which has increased to 10.7% but is still too low to provide a good explanation of how the primary effluent alkalinity varies with primary effluent pH. Also note how the slope of the regression line is negative telling us that as the pH increases the alkalinity is decreasing (inverse correlation) which we would not expect. Keep in mind that alkalinity addition occurs after the primary effluent so what we are seeing is the random and variable influence of primary clarification combined with recyle streams. Here again, the regression model does not fit the data well enough to use pH as a control or predictor of alkalinity.
In the third and final plot of final effluent alkalinity vs. final effluent pH, we again see an improvement in the R-Sq value which has now increased to 37.7%. But, the R-Sq value is still too low to provide a good explanation of how the final effluent alkalinity varies with final effluent pH. Also note how the slope of the regression line is positive, telling us that as the pH increases the alkalinity is increasing (direct correlation) which is what we would expect. Here we can see the influence of alkalinity addition into the activated sludge system. We are likely also seeing the unquantified, and likely highly variable, influence of alkalinity being replenished when denitrification in the secondary clarifiers occurs. Of course, let's not forget the variable nitrification rate occuring in the aeration tanks as well. Unfortunately, though a clearer (more logical?) picture emerges with the effluent data, the model still does not fit the data well enough to use effluent pH as a control or predictor of effluent alkalinity.
If you have gotten this far, I applaud you and thank you! It has been a long, complicated, and tortuous journey. So let me go ahead and wrap things up.
I've spent a lot of time trying to make the case that when you need to add alkalinity to your biological system you should focus on measuring alkalinity in your effluent. There is just too much going on as influent wastewater moves through the plant. There are a variety of physical, chemical, and biological processes and activities taking place. You have flow, concentration, and temperature changes, recycle loads, solids settling, air flow, variable waste activated and return activated sludge rates and concentrations, variable nitrification rates, the list goes on and on! When you need to nitrify and you need to add alkalinity to optimize nitrification, you must pay attention to your effluent alkalinity concentration rather than your effluent hdyrogen ion activity (pH!). If you are one of those many wastewater plants that nitrify and have sufficient alkalinity, all of this is a nonissue and likely something you give little thought to. If so, consider yourself lucky and have a nice day my friend! Thank you.
Hach Total Nitrogen TNT 828 ultra high range 20 to 100 mg/L N.
Hach Total Nitrogen TNT 827 high range 5 to 40 mg/L N.
Hach Total Nitrogen TNT 826 low range 1 to 16 mg/L N.
Hach Simplified s-TKN TNT 880 range 1 to 16 mg/L N.
Hach's Total Alkalinity TNT 870 range 25 to 400 mg/L CaCO3.
Sedlak, R. (Ed.). (1991). Phosphorus and Nitrogen Removal from Municipal Wastewater, Principles and Practices (2nd ed.). New York, NY: CRC Press.