Arsenic in Groundwater: A Slow-Motion Poisoning

The arsenic was always in the sand

Arsenic is the 20th most abundant element in Earth's crust, averaging about 1.5–2 mg/kg. In the Himalayan headwaters, weathering of sulfide minerals — chiefly arsenopyrite (FeAsS) and arsenic-bearing pyrite — releases arsenic into the rivers. As the Ganges, Brahmaputra and Meghna slow across the flat delta, they drop their sediment load. Crucially, the arsenic does not travel as a free ion. It travels adsorbed onto, or coprecipitated within, coatings of iron(III) oxyhydroxides — ferrihydrite (Fe₅HO₈·4H₂O), goethite (α-FeOOH) and lepidocrocite — that jacket every grain of quartz and mica.

These iron oxides are extraordinary arsenic sponges. A single gram of fresh ferrihydrite has a surface area near 200–600 m²/g studded with reactive ≡Fe–OH groups. Arsenate binds these sites as an inner-sphere bidentate complex, sharing oxygen atoms directly with iron. So long as the sediment stays oxidized, the arsenic is bound so tightly that pore water holds only a few µg/L. The poison is present but caged. What springs the cage is the disappearance of oxygen.

The trigger: bacteria breathing iron

Young deltaic sediments are packed with buried organic carbon — peat, leaf litter, and dissolved organic matter percolating down. Microbes consume it, and they do so by respiration: they oxidize carbon and dump the electrons onto whatever oxidant is available. There is a strict redox ladder, run in descending order of energy yield:

• Aerobic: CH₂O + O₂ → CO₂ + H₂O (O₂ used up first, within meters)
• Denitrification: 5CH₂O + 4NO₃⁻ + 4H⁺ → 5CO₂ + 2N₂ + 7H₂O
• Iron reduction: CH₂O + 7CO₂ + 4Fe(OH)₃ → 4Fe²⁺ + 8HCO₃⁻ + 3H₂O
• Sulfate reduction: 2CH₂O + SO₄²⁻ → H₂S + 2HCO₃⁻

The third rung is the killer. Once O₂ and NO₃⁻ are exhausted, dissimilatory iron-reducing bacteria (Geobacter, Shewanella) turn to the Fe(III) oxyhydroxides as their electron acceptor. As the ferric iron is reduced to soluble Fe²⁺, the mineral scaffold that held the arsenic literally dissolves out from under it. The arsenic is set free into the pore water. This reductive dissolution is now the consensus mechanism for the Bengal Basin and most Asian deltas — you can see its fingerprint in the water: high dissolved Fe²⁺, high bicarbonate from oxidized carbon, near-zero nitrate, and abundant dissolved arsenic all appearing together.

Which arsenic you get — and why As(III) is worse

Dissolved arsenic comes in two oxidation states, and the reducing conditions that release it also decide the ratio. Under the mildly reducing conditions of these aquifers (Eh roughly −100 to +100 mV, pH 6.5–7.5), arsenite, As(III), dominates, often 50–90% of the total. This matters enormously for three reasons rooted in acid-base chemistry.

• Charge. Arsenite exists as neutral arsenous acid, H₃AsO₃, up to its first pKa of ≈ 9.2. At drinking-water pH it carries no charge, so it is not electrostatically attracted to iron-oxide surfaces and sorbs weakly. Arsenate (H₃AsO₄, pKa₁ ≈ 2.2, pKa₂ ≈ 6.97, pKa₃ ≈ 11.5) is an anion — H₂AsO₄⁻ / HAsO₄²⁻ — at the same pH, and binds far more tightly.
• Mobility. Because it is uncharged and poorly sorbed, As(III) travels with the water. It is the more mobile, harder-to-capture species.
• Toxicity. Arsenite is about 60 times more toxic than arsenate. As(III) binds avidly to the sulfhydryl (–SH) groups of enzymes — a textbook soft-acid/soft-base interaction — shutting down pyruvate dehydrogenase and disrupting cellular respiration. Arsenate's danger is subtler: as a phosphate mimic (AsO₄³⁻ vs PO₄³⁻) it substitutes into ATP to form unstable adenosine-diphosphate-arsenate, uncoupling energy metabolism.

So the geochemistry loads the dice twice: reducing conditions release the arsenic and favor its most toxic, most mobile form.

The controlling variables: pH, competing ions, and phosphate

Whether arsenic stays in the rock or enters your glass is governed by a handful of levers, and small shifts flip the outcome.

• pH. Iron-oxide surfaces have a point of zero charge near pH 7–8. Above it the surface turns negative and repels arsenate; rising pH also swaps HAsO₄²⁻ for a more weakly held distribution. High-pH aquifers (the arid Southwest, parts of Argentina and Chile) mobilize arsenic by a different route — alkaline desorption — rather than reductive dissolution.
• Phosphate and bicarbonate. Phosphate (PO₄³⁻) is arsenate's chemical twin and competes directly for the same ≡Fe–OH sites — a common-ion-style competition. Fertilizer-derived phosphate and high bicarbonate from organic-carbon oxidation both bump arsenic off the mineral surface, raising dissolved concentrations.
• Redox potential (Eh). The master switch. This is why arsenic behavior is best read off a Pourbaix (Eh–pH) diagram: it maps exactly where H₃AsO₃, H₂AsO₄⁻ and the insoluble sulfide phases are stable.
• Sulfide. Push the system fully anoxic into sulfate reduction and generated H₂S can precipitate arsenic as insoluble sulfides such as orpiment (As₂S₃) — a natural sink. This is why the deepest and the shallowest wells can be safe while the middle is lethal.

The scale of the poisoning

The numbers are staggering. When millions of hand-pumped tube wells were installed across Bangladesh from the 1970s — a public-health triumph meant to escape cholera and dysentery in surface water — no one tested for arsenic. By the late 1990s the consequences were undeniable. Roughly 1 in 5 deaths in the most affected districts is now attributed to arsenic. Global estimates put 140 million people above the 10 µg/L guideline, with tens of millions drinking water above 50 µg/L — the old, laxer standard.

The exposure is chronic and cumulative. The classic signs of arsenicosis — hyperpigmentation, palmar-plantar keratosis — take 5–10 years to appear. The cancers (skin, bladder, lung, kidney) take 20+ years and are the leading cause of arsenic-related death. Arsenic is a confirmed Group 1 human carcinogen (IARC). Critically, there is no chelation therapy that reverses the cancer risk; the only real intervention is to stop drinking the water. Testing is done by hydride-generation atomic absorption spectroscopy or ICP-MS in the lab, and by simpler field kits based on the Gutzeit reaction, in which arsenic is reduced to arsine gas (AsH₃) that stains mercuric-bromide paper.

Beating the chemistry: how you actually remove it

The good news is that the same iron chemistry that released the arsenic can be used to recapture it. Nearly every practical treatment works by putting fresh Fe(III) oxide back into the water and letting arsenic sorb onto it — then filtering out the solids.

• Oxidize first. Because As(III) sorbs poorly, the first step is to convert it to As(V), which binds strongly: 2H₃AsO₃ + O₂ → 2H₂AsO₄⁻ + 2H⁺. Aeration alone is slow; chlorine, permanganate (KMnO₄) or even solar UV (the SORAS method) speed it up. This single redox step can raise removal efficiency from ~30% to over 90%.
• Coagulation–coprecipitation. Adding ferric chloride, FeCl₃ + 3H₂O → Fe(OH)₃(s) + 3H⁺ + 3Cl⁻, generates a flocculent Fe(OH)₃ precipitate that scavenges arsenate as an inner-sphere complex; the floc is then settled and filtered.
• Adsorptive media. Fixed beds of granular ferric hydroxide, activated alumina, or iron-oxide-coated sand pull arsenate out as water passes through.
• The simplest fix. Community-scale iron filters and household SONO filters (a composite iron matrix) exploit exactly this coprecipitation. And the cheapest solution of all is often just to drill deeper or shallower — into older, oxidized Pleistocene sands or protected zones where arsenic never mobilized — or to switch to rainwater harvesting.

Arsenite As(III) versus arsenate As(V): the two faces of dissolved arsenic

Property	Arsenite — As(III)	Arsenate — As(V)
Dominant species at pH 7	H₃AsO₃ (neutral)	H₂AsO₄⁻ / HAsO₄²⁻
Oxidation state	+3	+5
First pKa	≈ 9.2	≈ 2.2 (pKa₂ ≈ 6.97)
Charge at pH 7	0 (uncharged)	−1 to −2
Sorption to Fe/Al oxides	Weak below pH 8	Strong (electrostatic + inner-sphere)
Toxicity / mobility	~60× more toxic, far more mobile	Less toxic, more easily removed