Effect of Arsenic on the Denitrification Process in the Presence of Naturally-Produced Volatile Fatty Acids and Arsenic Removal by New Zealand Iron Sand (NZIS)
Thesis DisciplineCivil Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
This thesis is comprised of two phases; the first phase concerns the effect of arsenic on the denitrification process in the presence of naturally-produced volatile fatty acids (VFAs); while the second phase evaluates the arsenic removal efficiency of New Zealand Iron Sand (NZIS) by adsorption.
To accomplish the first phase of the study, VFAs were first produced naturally in an acid-phase anaerobic digester by using commercially-available soy flour. Secondly, a denitrifying biomass was cultivated in a sequencing batch reactor (SBR) using domestic wastewater as a feed solution. Finally, a series of biological denitrification batch tests were conducted in the presence of different concentrations of arsenic and nitrate.
As mentioned, the VFAs were generated from an anaerobic digester using 40 g/L soy solution as a synthetic feed. The digester was operated at a solids retention time (SRT) and hydraulic retention time (HRT) of 10 days. The pH of the digester was measured to be 4.7 to 4.9 while the mean temperature was 31 ± 4 °C; however, both these parameters were not controlled. In the effluent of the digester, a mean VFA concentration of 5,997 ± 538 mg/L as acetic acid was achieved with acid speciation results of acetic (33 %), propionic (29 %), butyric (21 %), iso-valeric (5%) and n-valeric acid (12 %). The specific VFA production rate was estimated to be 0.028 mg VFA as acetic acid/mg VSS per day. The effluent sCOD was measured to be 14,800 mg/L (27.9 % of the total COD), as compared to 9,450 mg/L (16.8 % of total COD) in the influent of the digester. Thus, the COD solubilization increased by 11.1 % during digestion yielding a specific COD solubilization rate of 0.025 mg sCOD/mg VSS per day. The extent of the digestion process converting the substrate from particulate to soluble form was also evaluated via the specific TOC solubilization rate (0.008 mg TOC/mg VSS per day), and VSS reduction percentage (17.7 ± 1.8 %).
A denitrifying biomass was developed successfully in an SBR fed with domestic sewage (100 % denitrification was achieved for the influent concentration of sCOD = 285 ± 45 mg/L and NH₄⁺-N = 32.5 ± 3.5 mg/L). A mean mixed liquor suspended solids (MLSS) of 3,007 ± 724 mg/L and a mean SRT of 20.7 ± 4.4 days were measured during the period of the research. The settleability of the SBR sludge was excellent evidenced by a low sludge volume index (SVI) measured to be between 50-120 mL/g (with a mean value of 87 ± 33 mL/g) resulting in a very low effluent solids concentration (in many cases less than 20 mg/L).
Several preliminary tests were conducted to estimate the right dosage of VFAs (digester effluent), nitrates and arsenic to be added and to confirm the occurrence of denitrification in an appropriate time frame of 4-6 h. From these tests, an optimum C/N ratio was observed to be somewhere between 2 to 4, somewhat higher than all the theoretical C/N ratios required for a complete denitrification using the four major VFAs identified in the digester effluent. During the denitrification batch tests, it was also observed that some NO₃⁻- N was removed instantaneously by reacting with As (III) (As₂O₃); while an increase in alkalinity of around 5.60 mg as CaCO₃ produced per mg NO₃⁻- N reduction was also observed. This latter number was very close to the theoretical value of alkalinity production (i.e. 5.41 mg as CaCO₃ per mg NO₃⁻- N).
The effect of arsenic on the denitrification process was evaluated by observing the specific denitrification rate in series of denitrification batch tests (with different concentrations of arsenic). Results from the denitrification batch tests showed that there was a clear effect for both As (III) and As (V) on denitrification. In particular, the specific denitrification rate fell from 0.37 to 0.01 g NO₃⁻- N /g VSS per day as the concentration of As (III) increased from 0 to 50 mg/L. In contrast, there was comparatively less effect for As (V); i.e. only a 37 % decrease in the specific denitrification rate (from 0.34 g NO₃⁻- N /g VSS per day to 0.23 g NO₃⁻- N /g VSS per day) when the initial arsenic concentration increased from 0 to a very high level of 2,000 mg/L. The effects of both the As (III) and As (V) forms of inorganic arsenic on the denitrification rate were further quantified by constructing exponential equation models. It was suspected that the effect of As (III) on denitrification was more substantial than the effect of As (V) because of the former’s toxicity to microbes.
Finally, the fate of arsenic was tracked by examining bacterial uptake. During the normal denitrification batch tests (i.e. designed for evaluation of the effect of arsenic on denitrification), no significant arsenic removal was observed. However, additional batch tests with a comparatively low concentration of biomass revealed that the denitrifying biomass removed 1.35 µg As (III) /g dry biomass and 2.10 µg As (V) /g dry biomass.
In the second phase of this research, a series of arsenic adsorption batch tests as well as a column test were performed to examine the arsenic (As (III) and As (V)) removal efficiency of NZIS from an arsenic-contaminated water. The kinetics and isotherms for adsorption were analysed in addition to studying the effect of pH during the batch tests. Breakthrough characteristics for both As (III) and As (V) were studied to appraise the effectiveness of NZIS treating an arsenic contaminated water.
Batch tests were performed with different concentrations of arsenic as well as at different pH conditions. A maximum adsorption of As (III) of approximately 90 % occurred at a pH of 7.5, while the As (V) adsorption reached its maximum value of 97.6 % at a very low pH value of 3. Both Langmuir and Freundlich Models were tested and found to fit with R² values of more than 0.92 in all cases. From the Langmuir adsorption model, the maximum adsorption capacity of NZIS for As (III) was estimated to be 1,250 µg/g, significantly higher (about three times) than for As (V) of 500 µg/g. In column tests, arsenic-contaminated water with total As concentration of 400 µg/L (in either form of As) were treated and a pore volume (PV) of 700 and 300 yielded a total arsenic level less than the WHO guideline value of 10 µg/L for As (III) and As (V) respectively; while, the breakthrough occurred after a throughput of approximately 3,000 PV of As (III) and 2,700 PV of As (V) with an average flow rate of approximately 1.0 mL/min.