Design and testing of a permeable reactive barrier system for treatment of sulphate rich waters at the former steep rock iron mine, Atikokan, Ontario

Abstract

Acid mine drainage (AMD) is one of the greatest challenges facing the mining industry globally. Methods for addressing this issue have been widely studied; however, few studies have addressed sites with a less common water quality problem resulting from AMD: neutral pH, metal-poor, and sulphate-rich water. The Steep Rock Iron Mine site in Atikokan, Ontario is utilized as a case study where AMD-affected waters have acidity neutralized by carbonate rocks, and metals precipitate out of solution as the pH rises. This process alleviates major environmental hazards associated with acidic waters and toxic metal concentrations; however, sulphate is not removed and presents toxic conditions for aquatic fauna. These sites are also a risk to human health, and can potentially contaminate drinking water supplies. Funding for the remediation of abandoned mine sites is limited, and innovative solutions utilizing passive treatment mechanisms are needed in order to deliver efficient and effective remediation. The goals of this project were: 1) to assess the capability of a permeable reactive barrier (PRB) system to remediate sulphate-rich, pH neutral, metal-poor water, 2) to assess nutrient balance within the system to ensure the availability of nutrients is not a rate-limiting factor for sulphate reduction, and 3) to improve reactive substrate selection procedures by determining which assessment tools are most useful in selecting substrates effective at stimulating sulphate reducing bacteria (SRB). Candidate reactive substrates including cow, horse, poultry, rabbit, and sheep manures, as well as leaf compost and hay, were assessed according to their concentrations of vital nutrients for SRB, including carbon, nitrogen, and phosphorous. Additionally, their relative degradability was tested via a procedure known as easily available substances (EAS). This testing determined how readily a given substrate could be broken down by bacteria, as well as the change in concentration of desired nutrients in the substrate before and after EAS testing, which gives an indication as to the availability of those specific nutrients. Plant and manure substrates were tested, with one of each type used in each reactive mixture. Based on this testing, poultry and sheep manures were selected as the most likely manure substrates to provide effective nutrition for SRB. In contrast, there was no significant difference found between hay and leaf compost. Poultry consistently performed the best in each test, with a C:N ratio of 11, a C:N:P ratio of 1772:160:1, and an EAS mass loss of 71%. Eight flow-through reactors were constructed and operated for a period of 23 weeks. Six of these reactors contained organic materials to stimulate SRB, while two were controls. Of the six reactors using organic materials, three mixtures were used, each containing a different combination of the four substrates. One control reactor assessed the impact of zero-valent iron which was also added to all of the organic reactors, while an additional control simulated the natural environment and contained only creek sediment and silica sand. Reactors 3 and 7 were the most effective at sustaining high rates of sulphate removal, with >80% sulphate removal maintained for the first 14 weeks. These reactors utilized a mixture of poultry manure and hay, validating the measures which indicated poultry manure as the most effective manure-based substrate. However, poultry manure was also used in reactors 1 and 5 in combination with leaf compost, and were not as effective for sulphate removal. These results indicate that hay was a more effective substrate than leaf compost. Comparing this finding against the original substrate testing presents two differences between hay and leaf compost; the C:N:P ratio and the availability of phosphorus in EAS testing both had stronger results for hay. This result indicates that phosphorus is a critical nutrient for SRB, and that tests considering phosphorus should be an integral part of reactive substrate selection procedures in systems attempting to stimulate SRB. The control reactors found that the addition of only zero valent iron did not have a significant impact on sulphate removal, as performance in this reactor was similar to the natural aquifer conditions control reactor. Eh/pH conditions supported the activity of SRB; but did not support the stability of sulphide produced by SRB, and it is unclear of SRB were in fact active within the flow-through reactors. Significantly reduced sulphate concentrations in reactor effluent initially appeared to indicate the sulphate reduction was occurred as intended. However, in post-experiment analysis there was no evidence of iron sulphide formation that would confirm sulphide production by SRB, and Eh/pH conditions were not supportive of sulphide stability. Furthermore, saturation index calculations using PHREEQC determined that iron sulphides were highly under saturated in effluent waters. In contrast, sulphate minerals including barite, gypsum, and jarosite were slightly oversaturated, and present a viable sink for the sulphate removed from solution. Following experiment completion it was found that the reactors with the greatest sulphate removal also had the most significant declines in nutrient concentration, with 52-64% C, 45-58% N, and 24-62% P losses in reactors 3 and 7. This is strong evidence of a bacterially driven process for sulphate mineral precipitation. The reduced availability of these nutrients may have played a role in the decline of sulphate removal over time.