Design, Construction, and Methodology:
After several design iterations and modifications (some of which occurred mid-construction), the final system evolved into what you see in the pictures below. The system is composed of four 1' tall PVC tubes (inner diameter, 4”), each with a different inoculation strategy and carbon media composition (measured by volume):
1: Woodchips (95%), active pond sediment (5%)
2: Woodchips (75%), biochar (20%), active pond sediment (5%)
3: Woodchips (95%), sterilized pond sediment (5%)
4: Woodchips (75%), biochar (20%), sterilized pond sediment (5%)
In addition to an 11” repository of media, ½” layers of steel wool at top and bottom of tubes allowed for even flow rates and limitation of media escape. Microbially active pond scum was collected from the VT Duck Pond and, in the case of sterilization, autoclaved to limit inoculation. Tubes were water-sealed; a peristaltic pump provided flow from the bottom of each tube to an effluent collection pipe at the system’s top. Water to be denitrified was collected and filtered from Spring 3 located within the StREAM Lab. The system was flushed once with DI water and twice with spring water at a high flow rate before testing began.
The pump ran continuously for four and a half days at a flow rate of approximately 4 mL/min to cycle through about nine residence times, each about 10 hours long. Samples were collected every four hours for the first two residence times and then every six hours for the third residence time. For the remainder of the experiment, collection only occurred every twelve hours. At the specified times to take samples, the tubing from each pipe was placed in a clean beaker to collect enough effluent to filter and analyze for nitrate, phosphate, and dissolved organic carbon.
After several design iterations and modifications (some of which occurred mid-construction), the final system evolved into what you see in the pictures below. The system is composed of four 1' tall PVC tubes (inner diameter, 4”), each with a different inoculation strategy and carbon media composition (measured by volume):
1: Woodchips (95%), active pond sediment (5%)
2: Woodchips (75%), biochar (20%), active pond sediment (5%)
3: Woodchips (95%), sterilized pond sediment (5%)
4: Woodchips (75%), biochar (20%), sterilized pond sediment (5%)
In addition to an 11” repository of media, ½” layers of steel wool at top and bottom of tubes allowed for even flow rates and limitation of media escape. Microbially active pond scum was collected from the VT Duck Pond and, in the case of sterilization, autoclaved to limit inoculation. Tubes were water-sealed; a peristaltic pump provided flow from the bottom of each tube to an effluent collection pipe at the system’s top. Water to be denitrified was collected and filtered from Spring 3 located within the StREAM Lab. The system was flushed once with DI water and twice with spring water at a high flow rate before testing began.
The pump ran continuously for four and a half days at a flow rate of approximately 4 mL/min to cycle through about nine residence times, each about 10 hours long. Samples were collected every four hours for the first two residence times and then every six hours for the third residence time. For the remainder of the experiment, collection only occurred every twelve hours. At the specified times to take samples, the tubing from each pipe was placed in a clean beaker to collect enough effluent to filter and analyze for nitrate, phosphate, and dissolved organic carbon.
Results:
After five days of testing (including 5 sample collections at 4:30 a.m.), we turned off the pump and sent our samples in to be analyzed. We were hoping to see a nice curve showing denitrification rates over a period of about ten residence times in the DNBR. The figure below represents the denitrification rate over the time of the experiment. The figure demonstrates that inoculation of the system with biologically active pond sediment decreases the time required to reach equilibrium nitrate removal rates. This can be attributed to the pre-existing denitrifying bacteria in the pond sediment. Also, the addition of biochar appears to increase the time required to reach equilibrium of nitrate removal rates. This could possibly be due to inefficiency as a carbon source or from competing biological processes. All four systems reach an equilibrium rate relatively close to each other, but the first system with only woodchips and inoculated pond sediment seems to settle at the most efficient nitrate removal rate.
After five days of testing (including 5 sample collections at 4:30 a.m.), we turned off the pump and sent our samples in to be analyzed. We were hoping to see a nice curve showing denitrification rates over a period of about ten residence times in the DNBR. The figure below represents the denitrification rate over the time of the experiment. The figure demonstrates that inoculation of the system with biologically active pond sediment decreases the time required to reach equilibrium nitrate removal rates. This can be attributed to the pre-existing denitrifying bacteria in the pond sediment. Also, the addition of biochar appears to increase the time required to reach equilibrium of nitrate removal rates. This could possibly be due to inefficiency as a carbon source or from competing biological processes. All four systems reach an equilibrium rate relatively close to each other, but the first system with only woodchips and inoculated pond sediment seems to settle at the most efficient nitrate removal rate.
The phosphorus removal rate over a period of 108 hours is shown in the figure below. They systems with biochar in addition to woodchips do show a slight improvement in removal efficiency. The 96 hour data point of pipe 1 shows an increase in phosphate and would need to be investigated whether the discrepancy is due to incorrect testing or if there truly was an increase in phosphate levels, which could have been a result from build up in the reactor and then a flush.
Finally, dissolved organic carbon (DOC) levels were monitored over the testing period and are seen in the figure below. There is a spike in DOC concentration in all four of the effluents because of the carbon media present, however it decreases significantly over the 108 hour time period. It would be interesting to see a longer testing period to see if DOC continues to decrease or levels off around 7 mg/L. Also, the pipes containing biochar showed an increase in DOC levels in effluents. The variance in carbon media suggests that biochar may leech more carbon than woodchips. Again, more testing should be conducted.