Iron removal and pH stabilization of synthetic acid mine drainage (AMD) using peat soil and limestone.

A Successive Alkalinity Producing System (SAPS) was evaluated for the remediation of synthetic Acid Mine Drainage (AMD) by using an organic substrate of peat soil and limestone aggregate. The characterization of the AMD was conducted by the testing the AMD euent originated from active mining site at Pengkalan Hulu, Perak, Malaysia. The characterization of the peat soil and limestone revealed stipulated composition, carbon content in the peat soil and particle size distribution of the limestone. Synthetic AMD was prepared incorporating iron sulphate (FeSO 4 ) and sulphuric acid (H 2 SO 4 ). The synthetic AMD was acidic (pH < 4.0) and with 50, 75 and 100 ppm Iron (Fe) concentration. A laboratory scale physical model to simulate a tailing pond was developed. The synthetic AMD was introduced and passed through the ltration media. Subsequently, water samples were collected and analyzed to determine pH level. Additionally, the Fe concentration was analyzed using UV-Vis test at 6 to maximum 48 hours’ retention time. The residues of the peat soil and limestone from the physical model were further analyzed using SEM-EDX microscopic to determine the remaining iron content absorb or precipitate onto the peat soil and limestone. The analysis revealed that the innovative methodology has successfully reduced more than 85 % iron content and neutralized the pH only after 6 hours of retention time. These results proved the combination of peat soil and limestone potentially can be proposed as the alternate solution for treating the AMD euent from the mining site.


Introduction
Acid mine drainage (AMD) is an environmental pollutant that impairs water resources due to anthropogenic activities including agriculture plantation, mining and logging. Unfortunately, mining industry often be criticised as a major contributor of the AMD. Exploitation of valuable minerals due to huge demand has unsuspectedly urged uncontrolled mining activities. These uncontrolled and some illegal mining activities escalate serious environmental problems that degrade most of the mining land area, surface water quality (stream and river), and eventually affected the quality of underground water reservoir. Previous study showed mining as a major source of heavy metal contaminations in the water nationwide (Affandi and Ishak 2018). This problem occurred due to formation of the AMD and released into adjacent water bodies from mining sites without proper and effective treatments.
In general, the reactions of the AMD formation are best illustrated by examining the oxidation of pyrite (FeS 2 ). The oxidation of pyrite occurred when the mineral surface is exposed to an oxidant and water either in oxygenated or anoxic systems (Blowes et al. 2013). The presence of oxygen and water, pyrite oxidizes to form ferrous (Fe 2+ ), sulphate ion (SO 4 2 ) and hydrogen (H + ) ion (Equation (1) Then, the Fe 2+ ion produced reacts with oxygen (O 2 ) to form ferric ion (Fe 3+ ). If oxygen is low, the chemical reaction in Equation (2) will not occur until the pH reaches 8.5 because the conversion of ferrous to ferric is slow at pH values equal to ve and lower under abiotic conditions (Simate and Ndlovu 2014). This reaction is facilitated by the sulphur-oxidizing bacteria (Thiobacillus thiooxidans, Thiobacillus ferooxidans) as they utilize the produced energy from this reaction for their metabolism (RoyChowdhury et al. 2015).
Over centuries, researchers dedicated their time to overcome problems created by mining industry.
However, some of the treatment methods are too expensive to be adopted in mining operation due to high capital cost and operation cost (Johnson and Hallberg 2005). The conventional method by constructing series of retention pond and adding hydrated lime into the AMD was proven successful to reduce heavy metal concentration and increase pH to the almost neutral level. Unfortunately, the timely maintenance and operation costs are immense. Additionally, this traditional method caused excessive hydrated lime usage; mass production of sludge with low cost-bene t ratio.
Currently, the industries adopted two AMD treatment methods known as active and passive treatments. The latter is the interest of the research. A Successive Alkalinity Producing System (SAPS) is one of the most commonly used passive treatment systems for AMD treatment in recent years. It is a hybrid treatment system incorporating technical merits of anaerobic wetland and anoxic limestone drain (Yim et al. 2015). Three major system elements of this system are the drainage system, a limestone layer and organic layer (Zipper et al. 2011 In this research work, passive treatment system of SAPS was adopted and tested on a laboratory scale using peat soil as organic substrate. The objective of this research work is to quantify the effectiveness of the peat soil for AMD e uent treatment. This research looks further into the application of the organicbased material as a treatment media. The peat soil was introduced as an alternative to conventional treatment media such as mushroom spent, wood shaving and animal manure.

AMD characterization
Six main rivers were identi ed and affected by these pro table mining activities. The AMD e uent samples were collected and analysed and the sampling locations are summarized in Table 1. . Both in-situ test and ex-situ test were conducted for analysis purposes. The water samples were kept in 1-L polyethylene bottles, soaked into 10 % of nitric acid overnight and rinsed with deionized water. This method ensures clean and uncontaminated samples. The samples that need to be taken to laboratory have been preserved to make sure the samples are unchanged from its original state. The samples were immediately acidi ed by adding HNO 3 to ensure pH lower than two (Affandi and Ishak 2018).
In-situ test was conducted on site to identify pH value, temperature and dissolved oxygen (DO). The reading of pH value and temperature were recorded using YSI Multiparameter. Subsequently, ex-situ tests were conducted in the laboratory for further analysis to determine heavy metal concentrations using Perkin Elmer ICP-OES. The heavy metals that have been detected by ICP-OES are Cadmium (Cd), Copper The limestone sample has been collected at quarry production stockpile via grab sampling method. About 1200 gram of limestone aggregate was used for size distribution analysis. Eight different size fractions of sieve were selected which are 20, 14, 10, 6.3, 4.75, 2.8, 2.3 mm and pan. Sieve results are reported as the differential weight percentage retained on each sieve as well as percentage of cumulative weight less than the sieve size.
Synthetic AMD composition The system consists of peat soil as organic substrate and limestone aggregate to treat the synthetic AMD with low pH level (below 4) and high iron concentration (50 to 100 ppm). The simulated AMD used in this study containing high acidity (pH 3 to 4), high concentration of Fe and sulphate. Distilled water was used due to its purity and less contaminant compared to tap water. The distilled water was collected in a 60 L container. The H 2 SO 4 was pipetted approximately 2.3 ml and dissolved into 60 L container.
Then, to prepare concentration of Fe as heavy metal in the synthetic AMD, FeSO 4 powder was incorporated accordingly. The consumption of FeSO 4 is shown in Table 2. The FeSO 4 was weighed and quantitatively transferred into the 60 L acidic solution and stirred thoroughly to make up the solution with respectively contained 50, 75 and 100 ppm of Fe concentration. Lab scale con guration The experiment was setup using transparent polypropylene tanks. The synthetic AMD was prepared in 60 L solution tank to supply the synthetic AMD solution to SAPS tank (Tank A and Tank B) by gravitational ow as shown in Fig. 1. The dimension of the SAPS tanks are 45 cm length, 32 cm width and 27 cm depth. The limestone aggregate was packed at the bottom of the SAPS tanks and covered the e uent pipe (Fig. 2). Subsequently, peat soil was placed over the limestone. As suggested by Zipper et al. 2011, no compact and physical disturbance on the organic layer to prevent creation of zones of preferential vertical ow, that will 'short circuit' the system and decrease treatment effectiveness. Total thicknesses for both media is 15 cm height. The synthetic AMD height was set constant at 5 cm throughout this study and was gravitational supplied from solution tank to SAPS tanks through in uent pipe at the top of the tank. Then the synthetic AMD was own downward through the peat soil before contact with limestone before being discharged thru e uent pipe.

SAPS operation
Six experimental set up were conducted. Tank A and Tank B were running simultaneously for each run with same in uent Fe concentration and pH level. The treated solution was collected at e uent pipe within 6 hours' interval of retention time. During e uent water sample collection, pH and temperature were measured in-situ and 15 ml of the sample was collected to measure the Fe concentration by using Ultraviolet-Visible (UV-Vis) spectrophotometry. The actual pH level and Fe concentration during experiment is shown Table 3. Generally, the pH levels for the collected sample were between 6 to 8 except for pH reading for rst sampling at Station 5 with pH equal to 4.70 due to discharge of AMD from several mine sites into the streams which eventually connected to Kuak River. Another reason that related to this nding is s due to logging and agriculture activities along the river. The ground that contains sulphide mineral was exposed, oxidised and formed AMD.
Most of the heavy metal concentration were found exceeding the allowable limit by the Mineral Development ( Heavy metal concentration for the third sampling is higher and mostly exceeded the limit. The highest Fe concentration reading was recorded is 188.5870 ppm (S3-3; Station 3, third sampling), As is 1.13 ppm (S6-3), Zn is 22.668ppm (S2-3), Pb is 114.871 ppm (S4-3), Cu is 6.107ppm (S5-3) and Cd is 2.492ppm (S5-3).
During the third sampling day, it was raining the night before, which caused higher heavy metal concentration as compared to the rst and second sampling. This proved that conventional method of the AMD treatment via retention pond and hydrated lime dosing are not effective during raining day. Over ow of the AMD at the discharge point or lack of e ciency of the retention pond in term of design volume to trap heavy metal caused the problems. By adopting this conventional method, the AMD is treated by the settlement of heavy metal inside the retention pond. In this case, during the raining day, the owrate of the AMD to be treated will increase and the heavy metals are not having enough time to settle hence, it will discharge out as untreated e uent with high concentration of heavy metals.

Peat Soil Characterization
The carbon content of the peat soil from CHNS analysis is show in Table 4. The peat soil sample can be classi ed as low carbon content compare to other organic substrate. Previous study was reported with 25.6 % carbon content by using mushroom compost (Yim et al. 2015). The carbon content recorded from EDX analysis is 6.24 % (Fig. 3). Major element in the peat soil is oxygen (O) 50.01 % followed by silica (Si) 25.18 %.
The higher carbon content of organic substrate, the higher the effectiveness of sulphate reducing medium to oxidize sulphide minerals via precipitation reaction. Carbon exist to provide bacteria to utilize dissolved oxygen in the AMD as energy source.
Limestone characterization Figure 4 shows the image of the limestone at 100 and 500 times of magni cation via SEM, respectively. The close up indicates the existing of the porous structures on the limestone surface. The porosity of the limestone is vital where precipitation of heavy metal such as iron hydroxide (Fe(OH) 3 ) expected to occur with aid of the porous structure.
The EDX analysis identi ed multiple mineral elements in limestone sample as the result shows in Fig. 5. Ca is the most abundant element (70.51 %) followed by O (26.31 %). Chemical composition of the limestone from XRF analysis in summarised in Table 5. The highest compound contained in the limestone is a CaO which is 98.55 %. Other compounds are too marginal less than 1 %. Total 100 Figure 6 shows cumulative weight passing vs size from the particle size distribution analysis. It was noted than D80 was at size 18 mm and D50 was at size 14 mm, respectively. Only 20 % from the limestone sample is more than 18-20 mm and 50 % of the total limestone sample is more than 14 mm.
Therefore, most of the limestone aggregates are within 10 to 20 mm. Limestone size range in a passive treatment study is not speci c, it is totally depending on tailing pond size or pilot column set up. Previous studies reported various size ranges. For instance, Zipper et al. 2011, in his paper stated that limestone size in actual design of SAPS that has been used was 100 mm to 152 mm. In contrast, Yim et al. 2015 via column set up that almost similar to this study adopted course gravel limestone particle size between 20 to 30 mm.

Temporal changes of pH level and Fe concentration
In all experiments, regardless of Tank A or Tank B, the pH level and Fe concentration of e uents transformed from strong acidic and high Fe concentration to neutral and low Fe concentration solution, respectively. There are two major factors contributed to the changes; alkalinity generation by sulphate reduction bacteria (SRB) activity in organic substrate and dissolution of limestone (CaCO 3 ). According to Kastyuchik et al. 2016, Ca is the most important acid neutralizing element. Meanwhile, Fe removal mechanism could be via adsorption on to the surface of organic substrate, precipitate as iron sulphide (FeS) and precipitate as iron hydroxide (Fe(OH) 3 ). Both organic substrate and limestone layer play an important role in this system to increase pH level and to remove Fe concentration of the synthetic AMD.
After 6 hours of retention time, pH level in all experiments increased drastically as shown in Fig. 7. Considering during this period, it is likely that alkalinity is generated by SRB activity. Most of the SRB grow optimally between pH levels of 6 to 8. Beyond this pH range, there is a decline in microbial sulphate reduction rate (Ayangbenro et al. 2018). Reaction between SO 4 2 and organic matter produces bicarbonate (HCO 3 ). However, the pH level for Experiment 1, 2 and 6 were dropped at 12 hours onwards while the pH level in Tank B for the experiments were maintained at higher level than. The result proved that su cient amount of the organic compound could allowed SRB activity to generate alkalinity and maintain the pH at neutral level for long period. According to the study conducted by Yim et al. 2015, alkalinity generation rate progressively increase with the height of the organic substrate layer not the volume of the substrate. The researcher also suggests that acid neutralizing capacity must be evenly distributed along the height of the organic substrate material.
Other than that, primary factor governing alkalinity generation is the rate at which limestone dissolved, which is effected by solution chemistry (Zipper et al. 2011). Considering this reaction happen at nal stage of the SAPS prior discharged out the synthetic AMD solution as e uent. Limestone is expected to dissolve most rapidly during the rst few hours of AMD contact as CaCO 3 dissolved more rapidly at lower pH and happened within rst 6 hours of retention time. The pH level of Tank A in Experiment 1, Experiment 2 and Experiment 6 gradually declined after 6 hours suggesting that the synthetic AMD in contact with the limestone became saturated with dissolved Ca 2+ and HCO 3 and the rate of limestone dissolution reduced considerably. During this period, the SRB activity in Tank A do not take place to generate alkalinity where the organic functionality is expected to degrade over time, due to microbial biodegradation (Zipper et al. 2011). Besides, decreases of pH level in Tank A is due to precipitation of Fe(OH) 3 by reacting with oxygen in limestone layer where the reaction produced H + ion.
Removal of Fe concentration associated with alkalinity generation from the SRB activity and limestone dissolution. The main Fe removal mechanism via adsorption on the surface of organic substrate, precipitation of Fe sulphides (FeS) due to sulphate reduction reaction and Fe(OH) 3 where the formation process is occurs rapidly when pH equal to 4 or higher  and was released into aqueous solution (Clyde et al. 2016). When AMD introduced and in contact with an organic substrate (peat soil), dissolved metals precipitate as metal hydroxides by alkalinity produced from the organics matter. The reaction of Fe(OH) 3 reduction to Fe 2+ process is shown in Equation (5) below where electron (e ) represents as electron donor, which is generally supplied by an organic substrate (Johnson and Hallberg 2005).
Fe(OH)3 + 3H+ + e → Fe2+ + 3H2O (5) Other reason of Fe concentration increment is due to availability of the organic surface for absorption mechanism of the Fe. Since the peat soil depth for Tank A in Experiment 6 was only 5 cm and entertained high Fe concentration of synthetic AMD in uent, the Fe removal e ciency would gradually decrease over time because of the available surface for absorption on the peat soil becomes smaller and limited.
Besides, increased of Fe concentration and fail to perform to remove Fe over long-term due to exhaustion of the substrate required for sustaining the SRB.

Conclusion
The characteristic of peat soil and limestone play an important role on pH neutralization and Fe removal. The porosity structures of peat soil give an advantage for Fe absorption. Besides, presence of SRB in the peat soil give an aid on pH neutralization other than limestone dissolution.
The peat soil thickness was identi ed as one of the major parameter for the SAPS sustainability.
Su cient volume of the peat soil could increase longevity of SRB activity during SAPS treatment.
The lab scale study of SAPS for the treatment of synthetic AMD using peat soil and limestone successfully demonstrated that the system was capable to renovated the pH level to neutral level and reduce Fe concentration. However, the study of retention time need to be explore more in detail for future study to determine the optimum hydraulic retention time for the SAPS.