Arsenic in Ground Water

Thermodynamic data are critical as input to models that attempt to interpret the geochemistry of environmentally important elements such as arsenic. Unfortunately, the thermodynamic data for mineral phases of arsenic and their solubilities have been highly discrepant and inadequately evaluated. This paper presents the results of a simultaneous weighted least-squares multiple regression on more than 75 thermochemical measurements of elemental arsenic, arsenic oxides, arsenic sulfides, their aqueous hydrolysis, and a few related reactions. The best-fitted thermodynamic database is related to mineral stability relationships for native arsenic, claudetite, arsenolite, orpiment, and realgar with pε-pH diagrams and with known occurrences and mineral transformations in the environment to test the compatibility of thermodynamic measurements and calculations with observations in nature. The results provide a much more consistent framework for geochemical modeling and the interpretation of geochemical processes involving arsenic in the environment.

Many of the important chemical reactions controlling arsenic partitioning between solid and liquid phases in aquifers occur at particle-water interfaces. Several spectroscopic methods exist to monitor the electronic, vibrational, and other properties of atoms or molecules localized in the interfacial region. These methods provide information on valence, local coordination, protonation, and other properties that is difficult to obtain by other means. This chapter synthesizes recent infrared, x-ray photoelectron, and x-ray absorption spectroscopic studies of arsenic speciation in natural and synthetic solid phases. The local coordination of arsenic in sulfide minerals, in arsenate and arsenite precipitates, in secondary sulfates and carbonates, adsorbed on iron, manganese, and aluminium hydrous oxides, and adsorbed on aluminosilicate clay minerals is summarized. The chapter concludes with a discussion of the implications of these studies (conducted primarily in model systems) for arsenic speciation in aquifer sediments.

Adsorption is the predominate mechanism controlling transport of arsenic in many ground water systems. Hydrous oxides of iron, aluminum, and manganese, and clay minerals are commonly associated with aquifer solids and have been shown to be significant adsorbents of arsenic. The extent of arsenic adsorption is influenced by the chemistry of the aqueous phase including pH, arsenic speciation, and the presence and concentration of competing ions. Under moderately reducing conditions, trivalent arsenite is stable and adsorption increases with increasing pH. In an oxidizing environment, arsenate is stable and adsorption decreases with increasing pH. The presence of phosphate, sulfate, carbonate, silica, and other anions have been shown to decrease adsorption of arsenic to varying degrees. The effects of complex aqueous and solid phase chemistry on arsenic adsorption are best simulated using surface complexation models. Coupling of such models with hydrologic solute transport codes provide a powerful method for predicting the spatial and temporal distribution of arsenic in ground water.

The release of arsenic from geothermal systems into surface and ground waters compromises the use of these waters as drinking water resources. In surface waters, As contamination can also adversely affect aquatic ecosystems, accumulating in sediments and plants. This review examines the source of arsenic in geothermal areas, its transport and speciation in geothermal fluids and receiving waters, as well as the deposition and removal mechanisms occurring in both natural environments and waste or water treatment systems. The effect of microorganisms on As mobility, and the opportunities that exist for further research in this field, are discussed. The review focuses on two geothermally active regions which have been intensively studied: Yellowstone National Park in the USA, and the Taupo Volcanic Zone in New Zealand, and their associated catchments.

The U.S. Environmental Protection Agency recently established a new maximum contaminant level of 10 micrograms per liter for arsenic in drinking water in the United States. Ground water is the primary source of drinking water for half the population of the United States. Several national assessments have found that high arsenic concentrations (above 10 micrograms per liter) are widespread in drinking-water aquifers in the western United States, the Great Lakes region, and New England. Moderate to high concentrations were identified in ground water in parts of the central and southern United States. This chapter summarizes national trends in the use of ground water as drinking water, and national estimates of arsenic occurrence in potable ground water. The chapter also briefly describes several studies on arsenic in specific settings and water-use scenarios; these studies illustrate by example the potential power of a regional approach to understanding and managing arsenic in drinking water.

Groundwater is now extensively used for drinking water in Bangladesh and present estimates indicate that there are some 6–11 million tubewells in Bangladesh. It is now apparent that approximately 1/4 of these wells contain arsenic at concentrations exceeding the Bangladesh drinking water standard (50 μg L ⁻¹ ). As many as 35 million people may be drinking arsenic-affected groundwater. We discuss a national survey of groundwater quality in Bangladesh that attempted to map the distribution and nature of affected wells. Other solutes measured included Na, K, Ca, Mg, Fe, Mn, P and SO₄. The worst-affected part of Bangladesh lies in the south-east of the country where the sediments are of Holocene age and where concentrations of arsenic frequently exceeded 200 μg L ⁻¹ . Where sampled, deep groundwaters (>150 m) were only rarely affected as were shallower groundwaters from older sediments including the aquifers underlying the Barind and Madhupur Tracts. Seven groundwater samples from the capital city of Dhaka also suggest that the city is not affected. The arsenic is undoubtedly of natural origin and the problem arises even though the sediments do not contain abnormal quantities of total arsenic. There is no evidence to suggest that the dissolved arsenic is derived from the oxidation of pyrite as some have suggested. Rather it appears that the high concentrations reflect a combination of factors: (i) young sediments undergoing rapid change from an oxidizing to a reducing environment following sedimentburial; (ii) the release of arsenic by one or more mechanisms which are poorly understood at present but which probably involve the desorption and dissolution of arsenic from iron oxides which are quite abundant in many of the worst-affected sediments; (iii) the very low hydraulic gradients throughout much of Bangladesh mean that groundwater flow is very slow which, combined with the ‘young’ age of many of the sediments, means that the natural flushing of the shallow aquifer will be slow allowing any released arsenic to accumulate. The rapid rate of deposition of sediments in Bangladesh and the Bengal Basin means that the chance of a well intercepting arsenic-rich water is likely to be relatively high compared with smaller deltas and other alluvial environments where the sedimentation rate is much lower.

Arsenic concentrations up to 12,000 μg/l have been measured in ground water from a sandstone aquifer in the Fox River valley in eastern Wisconsin, USA. In addition to a sulfide-bearing secondary cement horizon (SCH), which is present at the top of the aquifer, sulfide mineralization is also present throughout the aquifer. Within the SCH, arsenic occurs in pyrite and marcasite, and in iron hydroxides, but not as a separate arsenopyrite phase. Geologic, hydrogeologic, and geochemical data were used to characterize the arsenic source and the predominant geochemical process that controls its release to ground water. Several lines of evidence suggest that oxidation of sulfides is the cause of high (>100 μg/l) concentrations of arsenic in ground water, including 1) the presence of the arsenic-bearing sulfides in the aquifer; 2) water chemistry data that show a positive correlation between arsenic, iron, and sulfate and negative correlation between arsenic and pH; and 3) similar sulfur isotopic signatures in sulfides of the SCH and dissolved sulfate in ground water. We propose that atmospheric oxygen, introduced to the SCH through well boreholes, provides an oxidant to the system. This hypothesis is supported by the occurrence of high arsenic concentrations where water levels within the well intersect the SCH. However, the data do not unequivocally show sulfide oxidation to be the cause of the moderate (10−100 μg/l) and low (<10 μg/l) arsenic concentrations measured in ground water in the study area. The variability in thickness of the SCH and the concentration of arsenic within the sulfides, as well as the local availability of oxygen to the SCH, likely contribute to the spatial variability of ground water arsenic concentrations.

Arsenic levels exceeding 10 μg/L are present in hundreds of private supply wells distributed over ten counties in eastern and southeastern Michigan. Most of these wells are completed in the Mississippian Marshall Sandstone, the principal bedrock aquifer in the region, or in Pleistocene glacial or Pennsylvanian bedrock aquifers. About 70% of ground water samples taken from more than 100 wells, have arsenic contents ≥10 μg/L with a maximum value of 220 μg/L. Water samples and continuous cores were taken from two test wells. Arsenic content of core samples ranges from <5 to more than 300 ppm, with the highest values found for pyritic black shales. Authigenic cements in the Marshall Sandstone include patchy authigenic pyrite that locally contains arsenic-rich (up to 8.5 wt. % As) domains. Bulk arsenic contents of pyrite-bearing intervals, sampled in well cuttings, are a high as 1020 ppm. Arsenic-rich pyrite is likely the ultimate source of arsenic in eastern and southeastern Michigan ground water, but evidence for pyrite oxidation at depth in bedrock aquifers is generally lacking. Pyrite oxidation may occur or have occurred in tills derived from the Marshall Sandstone and Coldwater Shale, which were found to contain arsenic-rich (up to at least 0.7 wt. % As) iron oxyhydroxides. Plausible mechanisms for widespread arsenic mobilization in eastern and southeastern Michigan ground water include weathering of pyrite in tills, reductive dissolution of iron oxyhydroxides in tills, and potentially, pyrite oxidation in bedrock aquifers, due to drawdown in wells or lowering of water-table levels in response to Pleistocene glaciation.

Chemical data from more than 400 ground-water sites in the Middle Rio Grande Basin of central New Mexico indicate that arsenic concentrations exceed the U.S. Environmental Protection Agency drinking-water standard of 10 micrograms per liter across broad areas of the Santa Fe Group aquifer system, which is currently the almost exclusive source of drinking-water supply for residents of the basin. Identification of sources of arsenic to ground water of the basin is complicated by multiple sources of ground-water recharge that differ substantially in chemical composition. Establishment of a clear hydrologic framework for the basin was useful in interpreting the significance of patterns in arsenic concentration. This investigation indicates that there are two main sources of high-arsenic water to the Middle Rio Grande Basin. One primary source is related to silicic volcanism in the Jemez Mountains to the north, where dilute recharge water likely flows through rocks that have been altered by contact with geothermal fluids. The other primary source is mineralized water of deep origin that mixes with shallower ground water in several locations around the basin, particularly along major structural features. Ground water that has not been affected by either of these two high-arsenic sources generally has low arsenic concentrations. In some areas of the basin, values of pH exceeding about 8.5 appear to contribute to elevated arsenic concentrations through desorption of arsenic from metal oxides.

Migration of leachate from a municipal landfill in Saco, Maine has resulted in arsenic concentrations in ground water as high as 647 μg/L. Laboratory experimental data indicate the primary source of arsenic to be reductive dissolution of arsenic-enriched iron oxyhydroxides in the aquifer by organic carbon in landfill leachate. A core from an uncontaminated part of the aquifer yielded no dissolved iron or arsenic when leached with oxic ground water. Eluent ground water spiked with organic carbon in order to create reducing conditions mobilized both ferrous iron and arsenite from this core. The landfill was capped in early 1998 to eliminate the source of leachate. Cores from the contaminated portion of the aquifer were collected and leached with uncontaminated ground water in the laboratory to simulate natural remediation conditions. Data from these experiments show that significant concentrations of labile organic carbon have accumulated on aquifer solids, causing significant biological oxygen demand. In laboratory leaching experiments of the most contaminated core, the organic carbon caused complete consumption of the influent dissolved oxygen (6 mg/L) for 220 pore volumes. Arsenic leaching from contaminated cores rapidly decreased in concentration initially in response to flushing with uncontaminated ground water. Subsequent leaching produced more gradual decreases in dissolved arsenic concentrations, controlled by a combination of reductive dissolution of arsenic-enriched iron oxyhydroxides and adsorption/desorption. In leachate from the most contaminatedcore, arsenic concentrations exceeded the new United States Environmental Protection Agency drinking-water standard of 10 μg/L for more than 200 pore volumes. A geochemical model simulated the concentration of selected constituents as uncontaminated ground water eluted through contaminated aquifer solids. Concentrations of dissolved oxygen, arsenic, and iron, in leachate from one core were used to calibrate the model. This model was validated by successfully simulating constituent concentrations in leachate from cores collected from other contaminated areas of this aquifer.

The cyclic injection of oxygenated water in an aquifer may induce in situ iron removal from groundwater. During injection of aerated water, sorbed ferrous iron is displaced by cations, oxidized in the pore space, and precipitated as ferric iron oxyhydroxide. During pumping, ferrous iron is sorbed from groundwater on the exchange and sorption sites, and the breakthrough of dissolved iron is retarded. Other trace elements such as arsenic may be eliminated jointly with iron by sorption or co-precipitation.

The volume of iron-free groundwater that can be pumped per volume of injected, aerated water defines the efficiency of the process. The efficiency is determined by the ratio of the retardations of oxygen during injection and of iron during pumping. This chapter shows how these retardations can be calculated forgiven water qualities and aquifer compositions.

The first seven cycles of an in situ iron removal project in The Netherlands were simulated with the hydrogeochemical transport model PHREEQC (version 2). The concentration changes of CH ₄ , NH ₄ ⁺, Mn ²⁺ , Fe ²⁺ , PO ₄ sk ^3- and As are discussed in detail. Arsenic shows concentration jumps in pumped groundwater which are related to oxidation/reduction and sorption/desorption reactions resulting from the water quality variations.

In situ removal of arsenic from ground water used for water supply has been accomplished in circum-neutral ground water containing high dissolved iron concentrations. In contrast, the ground water at our study site is alkaline, contains measurable dissolved oxygen and little dissolvediron. Because the dissolved iron concentration is low in the basalt aquifer, the iron oxide content of the aquifer would not increase with successive pumping cycles unless iron is added to the injected water. Additionally, the high pH limits adsorption onto iron oxide present in the aquifer. Having the ability to lower arsenic concentrations in high-pH, oxic ground water could have wide application because similar high arsenic ground water is present in many parts of the world.

Laboratory and field results show that the basalt has limited capacity for adsorption of As(V), presumably by naturally occurring hydrous ferric oxide (HFO). However, addition of HFO can significantly increase As(V) adsorption. Lowering the pH combined with increasing the iron oxide content in the basalt aquifer reduces arsenic concentrations in produced ground water. Arsenic removal was very effective in laboratory experiments and during early part of the initial push-pull experiment. Moderate arsenic removal after increasing the iron oxide content of the basalt aquifer and lowering the pH during the cross-flow experiments was limited to about 50% at best. This relatively low removal may be due to a variety or combination of factors, including local hydraulics of the aquifer, chemical reaction kinetics, and changing aqueous chemistry caused by the lowered pH.