Editorials

The Chemical Enginner

Nitrification toxicity monitoring

Dr. Steve Callister and Prof. Tom Stephenson
on preventing wastewater treatment failure

A new technique for evaluating nitrification efficiency in activated sludge systems directly measures the greenhouse gas dinitrogen oxide (dinitrogen oxide) as an indicator of nitrification failure. The technology helps prevent ammonia pollution of the aqueous environment by providing an early warning when the wastewater treatment is breaking down and, by monitoring dinitrogen oxide levels, helps operators to reduce greenhouse gas emissions. The dinitrogen oxide monitor has been developed and patented by Water Innovate Ltd. under the name N-Tox®

Nitrification Inhibition

As the regulations on effluent quality get increasingly tight, many industrial and municipal wastewater treatment plants now remove nitrogen through nitrification and denitrification. Researchers at Cranfield University in the UK found that when the nitrification process fails, dinitrogen oxide is released. The rate of dinitrogen oxide production was linked to oxygen depletion and higher than normal ‘shock’ loadings of ammonia arriving at the treatment plant.  When oxygen is depleted, nitrification fails and dinitrogen oxide is produced.

Figure 1 – Activated sludge nitrogen cycle

Nitrification, the oxidation of ammonia to nitrite and nitrate, is a sensitive process undertaken by nitrifying bacteria which can be inhibited by high ammonia concentrations and chemical pollutants, which reduces the efficiency of the process or causes the treatment system to break down entirely.  It is not easy to predict when the nitrification process will be inhibited, but it results in ammonia leaking into the effluent, which in turn often causes a breach of the ammonia discharge consent.

The occurrence of dinitrogen oxide under normal stable process conditions in the activated sludge nitrogen cycle is shown in Figure 1. dinitrogen oxide is normally present as an obligate intermediary in the denitrification part of the cycle, but its presence is short-lived and at very low concentrations. However, if nitrification-inhibiting compounds are present, this causes  the production of dinitrogen oxide in the nitrification component of the cycle to increase.

Measuring the level of dinitrogen oxide off-gas is a useful, non-invasive way of monitoring the performance of the nitrification process and helping prevent ammonia leakage.

Techniques

As nitrification can be inhibited frequently for short periods of time, toxic shocks need to be stringently monitored. This has fuelled development of new technologies to monitoring changes in wastewaters.

In addition to the new non-invasive technique, methods for detecting nitrification failure include generalised toxicity tests, on-line ammonia probes and on-line respirometry systems. However, toxicity kits do not target nitrification, and on-line probes are susceptible to fouling, have to be frequently recalibrated, and do not give a rapid failure indication. On-line respirometry requires the operation of a mini biological treatment plant, with all the attendant maintenance problems.

Figure 2 – Pilot study results

Researchers using pilot-scale activated sludge plants in the pilot hall facility at Cranfield University, measuring the dinitrogen oxide concentration above the start of an activated sludge aeration lane, found that it takes at least one hydraulic retention time (HRT) of the aeration tank and final clarifier before nitrification fails. Figure 2 demonstrates that, as a result of a short aeration failure, the delay between the increased level of dinitrogen oxide being detected and the appearance of ammonia in the effluent means that operators could get the duration of one HRT cycle as advance warning of nitrification failure.

Once a failure is detected, there are various options to restore nitrification, including increasing aeration rates, bypassing influent to storage tanks, recycling final effluent to works inlet, or controlling the return of high ammonia sludge liquors to the works inlet.

Full Scale Test

Cranfield University and Severn Trent Water set out to examine whether the non-invasive dinitrogen oxide system could be applied to monitor dinitrogen oxide on a full-scale municipal activated sludge plant. A detailed study was carried out at Finham wastewater treatment works, UK.

Figure 3 – Activated sludge aeration lane at Finham

The effluent treatment system at Finham incorporates a single stage nitrification process, with both chemical oxygen demand (COD) and ammonia removal taking place in a single biological process with a total volume of 4500 m3. Aeration lanes are 30 m long, the first 5 m portion relating to the anoxic zone (see Figure 3). The activated sludge lanes have a total HRT of 12 hours corresponding to a daily average influent flow of 110,000 m3/24hours. The anoxic component of the HRT is 2 hours.

One of the activated sludge lanes was fitted with a prototype of the dinitrogen oxide off-gas monitoring system. The researchers also designed a sampling system to produce a lane profile of dinitrogen oxide levels to detect the production of off-gas and monitor potential toxic shock upsets. Each sampling point was analysed for 24 hours and then either moved across the lane or down the lane by 5 m. A floating hood, suspended over the mixed liquor and connected to sample line tubing, directed gases to a gas analyser that measured dinitrogen oxide in the range 0-1000 mg/l.

The gas analyser remained separated from the activated sludge at all times, drawing the off gas via a small pump inside the gas analyser, from the headspace via additional tubing. This avoided problems such as corrosion and sensor fouling.

The influent and effluent chemical oxygen demand (COD), ammonia (ammonia), nitrite, nitrate and pH levels were measured daily, as were levels of mixed liquor suspended solids (MLSS), online pH, temperature and dissolved oxygen.

The removal rates for the monitored activated sludge lane can be seen in Table 1. There were no toxic shock loadings during the monitoring period. The samples taken for effluent ammonia showed a range of 0.1–1.2 mg/l. If there had been a shock load, this would have caused a noticeable increase in the effluent ammonia concentration.

Table 1 – Operating performance parameters at Finham

MLSS

Influent

Effluent

ammonia range

COD range

ammonia range

COD range

mg/l

mg/l

mg/l

mg/l

mg/l


2034 – 2879


35.2 – 78.5


571.5 – 687.5


0.1 – 1.2


37 – 67

Dinitrogen oxide off-gas measurements were averaged at each 5 m interval, and results from three monitoring points (left, middle and right side of aeration lane) were averaged at each interval down the lane for the corresponding 24 hour period. The lane profile (see Figure 4) shows an initial increase in the average dinitrogen oxide concentrations followed by a decrease as the sampling system moved down the lane.

At 0 m, a rise in the average dinitrogen oxide off-gas peaked at 7.2 mg/l (7 hours), followed by a decrease and then a second increase at 20 hours. At 5 m, average dinitrogen oxide concentrations declined and were found to be less than 0.9 mg/l up to 46 hours into the trial, with low levels maintained throughout the rest of the experiment apart from a 2.1 mg/l peak (84 hours), a peak of 1.8 mg/l (107 hours), and a minor increase to 0.9 mg/l (138 hours).

The results suggest that the initial increase of dinitrogen oxide concentration was a consequence of the denitrification process. As stated earlier, dinitrogen oxide is an obligate intermediate in this process. When the monitoring system was moved down the activated sludge lane it recorded a decrease in the dinitrogen oxide off-gas level, but only to background emission levels.

The pilot scale experiments at Cranfield had used high ammonia loadings of 2.44 g/l ammonia-N combined with oxygen deprivation to yield a maximum dinitrogen oxide concentration of 18.7 mg/l as a result of nitrification failure. A second experiment with an increased shock load (7.32 g/l ammonia-N) produced a significantly higher maximum concentration of dinitrogen oxide off-gas, but even the lower peak at 18.7 mg/l dinitrogen oxide was well in excess of the background levels monitored at Finham.

These pilot scale experiments had shown an initial increase in the dinitrogen oxide concentration followed by a time lag before a sharp increase of ammonia in the effluent. It is this period, before ammonia appears in the effluent, which is used to indicate failure of the nitrification process, thus providing an early warning for plant managers and operators.

Figure 4 – Dinitrogen oxide aeration lane profile results at Finham

 

 

The monitoring of dinitrogen oxide at Finham produced small variability in the dinitrogen oxide background emissions of 0 – 2 mg/l. Hence, a dinitrogen oxide peak would be measured in the event of a toxic shock load to the activated sludge plant.

If a toxic shock was to occur at Finham, then the dinitrogen oxide peak would have been expected from the activated sludge lane followed by a time lag of approximately one HRT (10 hours) of ammonia in the effluent. Hence, staff at Finham would have up to 10 hours to take remedial action to rectify the problem and prevent ammonia discharge.

Market Potential

The key markets for this non-invasive technology are municipal sewage and industrial effluent treatment. Many treatment works are consented for ammonia at < 5 mg/l. Like Finham, these plants rely on bacteria to remove ammonia by converting it to nitrate, and the use of the technique to measure dinitrogen oxide peaks could provide valuable time to attend to process problems.

dinitrogen oxide peak warning is even more critical in activated sludge plants treating some industrial effluents, such as landfill leachate or pharmaceutical wastewaters, because their high ammonia levels make the consequences of nitrification failure much more serious.

The technology is also used to quantify dinitrogen oxide greenhouse gas emissions from wastewater treatment plants - an area of increasing concern for plant operators and regulators, as the global warming potential of dinitrogen oxide is almost 300 times that of carbon dioxide. According to the US Environmental Protection Agency, wastewater treatment accounts for up to 4 % of dinitrogen oxide emissions and tightening of nitrogen discharge consents means more nitrification and denitrification will be needed. The new monitoring technique can help plant operators control processes effectively to significantly reduce a predicted increase in dinitrogen oxide emissions.

A dinitrogen oxide monitor based on the research carried out by Cranfield has been developed and patented by Water Innovate Ltd. under the name N-Tox®. The design comprises an integral sample pump, gas conditioning device, non-dispersive IR gas analyser, auto-calibration system and data logging unit, housed within an IP65 enclosure. The instrument transmits 4-20 mA in proportion to dinitrogen oxide level and has various configurable alarms for setting plant failure warnings.

Acknowledgements

The experimental work was supervised by Elise Cartmell and carried out by Mark Butler of Cranfield University, and funded by Severn Trent Water through an industrial CASE studentship and the UK Engineering and Physical Sciences Research Council.

Further Information

Burgess, J E, Colliver, B B, Stuetz, R M, and Stephenson, T (2002) Dinitrogen oxide production by a mixed culture of nitrifying bacteria during ammonia shock load and aeration failure. J. Ind. Microbiol. Biotechnol., 29, 309 – 313.

Burgess, J E, Stuetz, R M, Morton, S, and Stephenson, T (2002) Dinitrogen oxide detection for process failure early warning systems. Water Sci. Technol., 45 (4), 247 – 254.

Butler, M D, Cartmell, E, Stokes, L, and Stephenson, T (2005) Non-invasive monitoring for early warning of nitrification failure. Proceedings of WEFTEC 2005, Water Environment Federation, Alexandria, USA.

Download PDF

 

Back to top