On the use of iron oxide colloidal suspensions for improving water quality

Laburpena

Among all threats that impair water quality (acidification, salinization or contamination with xenobiotics), eutrophication has become one of the most common affecting lakes and reservoirs from all over the world. Enrichment of phosphorus (P) in water bodies, which is the growth-limiting nutrient of primary productivity in freshwater bodies, constitutes the main cause of eutrophication. Fertilizer run-off coming from agricultural areas and insufficient wastewater treatment are the main responsible for the increase of P inputs to the aquatic ecosystems leading to biodiversity loss, toxic cyanobacteria blooms and changes in the function of the system. The use of lake restoration techniques in order to meet the requirements set by environmental policies is increasing from the last 50 years. In Europe, The Water Framework Directive (WFD) has established as a legal requirement to achieve a "good status" in all water bodies by 2027. Recently, the approaches that aim at reducing P availability in the water column have been differentiated in three categories: reduction of external P inputs to the system, increasing P retention by the system and increasing P export from the system. Although a multiapproach is often required, the keystone of a lake restoration project is the reduction of external P load which could be accomplished by sewage diversion, natural/constructed wetlands or by P elimination plants (PEPs) situated at the main inflow of the lake. After a reduction of external P load, biological resilience and P release from sediment (internal P load) may be responsible for the delay in lake recovery. In such cases, the application of alum, iron (Fe) and calcium salts (P inactivation methods) is recommended to increase P adsorption capacity of the sediment. Nitrate addition, hypolimnetic aeration and sediment capping are other known techniques to reduce internal P load. Among the techniques intended to increase P export from the system can be highlighted hypolimnetic withdrawal, external elimination of P-enriched hypolimnetic waters and sediment dredging. The manipulation of the physical and also the biological structure of the aquatic ecosystem constitute other type of techniques to improve water quality. A direct consequence of eutrophication is the solubilisation of heavy metals due to the low redox potential present in the anoxic bottom waters of lakes and reservoirs during summer stratification. Natural waters with high concentration of soluble manganese (Mn) may be pumped to water distribution network generating operational problems as the obstruction of pipes and the impairment of water quality for human consumption if they are not properly treated. Current techniques to remove Mn from aqueous solution are based on Mn oxidation and precipitation by adding chemical substances that alter water quality. Nowadays there exists not a panacea in lake restoration since all the available restoration techniques present advantages and disadvantages depending on the characteristics of the system, but without doubt, a successful restoration strategy requires a substantial knowledge of the history lake and also the system ecology in order to anticipate to possible changes or responses. Due to the difficulties that present traditional P inactivation methods -redox sensitivity, biotoxicity in highly acidic or alkaline media, acidification of lake water due to floc formation, reduction of P adsorption capacity with ageing of the adsorbent- and given the unsuitability of traditional methods to remove Mn in water treatment plants, the development of new tools to combat eutrophication and to improve water quality is of special interest for water resources management. Magnetic nano and microparticles, mostly based on Fe, have attracted special attention in areas such as biomedicine, industry, environmental remediation and the ones that concern the present study: water treatment and lake restoration. The reason why their use has become widespread in the last decades for adsorbing pollutants from aqueous media stems from their unique properties: fast adsorption kinetic due to their high surface reactivity, high adsorption capacity due to their high surface area and their magnetism which makes them different to other adsorbents. The latest property implies that magnetic particles can be removed from the media by applying a magnetic separation gradient once the target contaminant is adsorbed, thus minimizing contact time with biota and enabling their recovery for further reuse. One of the main advantages of using magnetic particles as adsorbents is the possibility of surface functionalization in order to increase the affinity for the adsorbate, reduce the biotoxicity or modify particle density for specific purposes. A low-cost granulate material consisting of dried amorphous ferric oxy-hydroxide (CFH-12®, Kemira) has been recently proposed to adsorb P from wastewater with excellent results, being those low cost, high adsorption capacity and strong P-bonding with low desorption. However, those studies were only conducted at aerobic conditions and only for short exposure times (up to 2 weeks) for P adsorption. Further knowledge about long-term ageing effects or reductive effects on CFH-12® under anoxic conditions and their implications in P adsorption capacity is needed. In this PhD dissertation the main purpose was to assess the performance of three different adsorbents to remove/adsorb P from lake water and sediment in a context of lake restoration. Two of these compounds are magnetic as the case of commercial micronsized carbonyl Fe particles (grade HQ and CC, BASF) and hybrid chitosan magnetic particles synthesized in laboratory whereas the other one is non magnetic and corresponds to CFH-12®. Apart from the removal of P, the feasibility of using micronsized carbonyl Fe particles to remove Mn from the aqueous solution in water treatment was also studied. To accomplish that, firstly the influence of pH on Mn removal efficiency by carbonyl Fe particles (grade CC) in batch conditions was studied. In these experiments a strong influence of pH on Mn removal efficiency was obtained, being the removal efficiency extremely high at pH higher than 9 (>98%) as a result of chemisorption of Mn on magnetic particles surface but considerably lower (<30%) when decreasing pH value due to the presence of Mn as soluble cation. Secondly, the effect of other parameters such as contact time, adsorbent concentration and surface coating with Mn oxides [MnOx(s)] on Mn removal efficiency was evaluated. Results demonstrated that MnOx(s) precipitate on magnetic particles surface in less than 1 hour at pH 9 and that Mn adsorption efficiency clearly increases when increasing adsorbent concentration (from 1 g/l to 2 g/l) and when coating magnetic particles with precipitated MnOx(s) which have high affinity for soluble Mn. The maximum adsorption capacity of MnOx(s)-coated magnetic particles was not reached since even working at very high Mn concentrations, adsorption sites were not saturated due to a multi layer adsorption process in which adsorption surface was continuously regenerated by overlapping MnOx(s) layers. The desorption of precipitated MnOx(s) from magnetic particles surface was tested as well as and the possibility of reusing the adsorbent. Results showed a low desorption rate (<10%) in a wide range of pH (6-9) as well as a high Mn removal efficiency (>99%) of reused magnetic particles, highlighting the stability of MnOx(s) precipitated on magnetic particles and the possibility of reusing them without losing adsorption capacity. As a result of previous studies that highlighted the good performance of carbonyl Fe particles to remove P under batch and flow conditions -high P adsorption capacity (18.83 mg/g), fast adsorption kinetic (less than 2 h) and pH-independency of P removal efficiency (>80% in pH range 6-9)- it was set up a core incubation experiment with sediment and water from a brackishwater lagoon (Laguna Honda, Almería, Spain). Cores were incubated in oxic and anoxic conditions with the purpose of assessing the effect of adding carbonyl Fe particles (grade HQ) for 24 h on P efflux and sedimentary P pools. To do so, magnetic particles were added over the surface water of the cores and after 24 h they were removed by applying a magnetic separation gradient. Results showed a complete suppression of P efflux in the most unfavourable conditions that are usually present in hypertrophic lakes (anoxic conditions). Likewise, the application of magnetic particles resulted in an increase of P export from the system since mobile P (PMobile) was reduced 22-25 % in anoxic and 12-16 % in oxic conditions. In another experiment, the effect of adding carbonyl Fe particles (grade HQ) on P cycle, but also on other nutrients such as carbon (C), reactive silicate (Si) and nitrogen (N) as well as other parameters related to water quality, was evaluated. For that, an experiment with large enclosures containing 6600 cm3 of sediment and 40 litres of water from Laguna Honda was set up. Magnetic particles were added to the enclosures in two different ways (1) above the surface water or (2) above the surface sediment. After a contact time of 24 h they were removed from the enclosures by a specially designed magnetic rake. The enclosures were incubated for a 70-days oxic period and a 5-days anoxic period. It was found that the addition of magnetic particles caused a reduction of more than 68 % of dissolved inorganic P (DIP) in both oxic and anoxic period as well as a reduction of 50% of Si and a 15 % of dissolved organic C (DOC), indicating that both Si and DOC interfered in P adsorption capacity by magnetic particles. Sedimentary P pools as the case of P bound to humic acids (P→NaOH, Humic), NaOH extractable-P (P→NaOH) and easily degradable organic P (Org-PLabile) were reduced by 11-21%, 15% and 12 %, respectively by the addition of magnetic particles. It was concluded that adding magnetic particles over the surface water was more recommendable due to a higher recovery efficiency of magnetic particles with the magnet (90%) against the 30% obtained when adding them over the surface sediment. In the cores as well as in the enclosures experiments above mentioned it was evidenced the high density of carbonyl Fe particles which tend to penetrate deeper in sediment layers. Deeper penetration of the adsorbent within the sediment implies less availability for P adsorption (since the main target is PMobile that concentrates in the first layers of sediment and decreases with depth), higher resuspension and less recovery efficiency of magnetic particles when removing them by applying a magnet. To overcome these drawbacks, composite chitosan + magnetite particles were synthesized. Resulting hybrid microparticles exhibited enough P adsorption capacity (4.84 mg/g) to be used in hypertrophic lakes and they were demonstrated to be excellent candidates to extract PMobile from the upper sediment layers because they sediment slower than traditional carbonyl Fe particles. Finally, it was carried out a core incubation experiment with sediment and water from three Danish lakes in order to evaluate CFH-12® and the well-known freshly formed Fe(OH)3 floc as P inactivation agents for lake restoration. The experiment was divided into an oxic and anoxic period in order to see effects of reducing conditions on the performance of the two compounds as well as the effect of ageing on P adsorption capacity. From results, it was stated that CFH-12® is a promising lake restoration tool which does not lead to a pH drop of lake water during addition and shows, with a Fe: PMobile binding molar ratio of 7.6-8:1, a significant reduction of DIP efflux. Moreover, CFH-12® showed, in contrast to freshly formed Fe(OH)3 floc, no changes in its P binding capability with ageing and low redox sensitivity.