What's in my water
Arsenic in drinking water — testing, limits and what actually removes it
Arsenic is a naturally occurring metalloid that dissolves out of aquifer sediments into groundwater, where it is colourless, tasteless and odourless at every concentration relevant to health.
Limits and standards
BIS publishes two figures for many parameters: an acceptable limit, and a permissible limit that applies only where no alternate source of water exists. The second is a fallback for communities without a choice — not a target to design a purchase around.
| Standard | Limit | What it means |
|---|---|---|
| BIS IS 10500:2012 — acceptable limit | 0.01 mg/L | Tightened from the earlier 0.05 mg/L. Some older reports and test kits still work to the old figure — check which one you are being shown. |
| BIS IS 10500:2012 — permissible in absence of alternate source | 0.05 mg/L | — |
| WHO guideline value | 0.01 mg/L | WHO designates this a provisional value, set by the practical limit of measurement and treatment rather than by an established no-effect level. |
Where it comes from
Aquifer geology
Arsenic-bearing sediments release arsenic into groundwater under reducing (low-oxygen) conditions. This is a natural process, not pollution.
Depth
Concentration varies sharply with depth in the same location. Two borewells 100 m apart, drilled to different depths, can differ by an order of magnitude.
Irrigation drawdown
Heavy extraction changes aquifer chemistry over years, and arsenic levels in a well are not fixed for life.
Industrial discharge
A localised contributor near smelting, glass and some pesticide-handling operations, but far smaller in India than the geogenic source.
Why it matters
- WHO classifies long-term arsenic exposure through drinking water as associated with skin lesions, and with cancers of the skin, bladder and lung.
- The exposure is cumulative and slow. There is no acute warning and no symptom in the first years, which is exactly why testing rather than observation is the only workable approach.
- Cooking concentrates it. Boiling rice or dal in arsenic-bearing water and letting it absorb the water increases dietary intake rather than reducing it.
Note
Can you detect it yourself?
This is the section most purifier buying decisions skip. A handheld TDS meter is the instrument almost every Indian household is shown, and for most contaminants that matter it reads exactly the same whether the contaminant is present or not.
| Question | Answer |
|---|---|
| Can you sense it? | None. Arsenic is colourless, odourless and tasteless in water at every concentration that matters. Anyone claiming to identify it by taste is mistaken. |
| Does a TDS meter show it? | Nothing. At the BIS limit of 0.01 mg/L, arsenic contributes 0.01 ppm to a TDS reading — inside the rounding error of any handheld meter. A borewell at ten times the limit reads the same as one with none. |
| Field / home test | Mercuric-bromide field kits give a rough band (typically <10, 10–50, >50 ppb) and are widely used for screening in affected districts. They are useful for triage, not for confirming that water is safe — false negatives near the limit are common. |
| Laboratory test to ask for | Arsenic by AAS-HG or ICP-MS — Hydride generation atomic absorption spectrometry, or ICP-MS for lower detection limits (indicative ₹600–₹1,800) |
When to pay for a laboratory test
- You are on any groundwater source in West Bengal, Bihar, Uttar Pradesh, Assam, Jharkhand, Manipur or Chhattisgarh — the belts where CGWB has documented arsenic occurrence.
- A new borewell, or an existing one deepened. Arsenic varies with depth and the old result does not carry over.
- Anyone in the household has unexplained skin pigmentation changes or thickening on palms and soles, and the supply is groundwater. Test the water and speak to a doctor — do not wait for one to prompt the other.
- You installed an RO unit for arsenic and have never verified the output. Rated reduction and delivered reduction are different numbers.
What removes it
Effectiveness ratings below describe the technology working correctly, within its service life, on water or air that suits it. “Depends” means there is a condition in the caveat that decides the outcome — read it before buying.
- Reverse osmosis (RO)Depends
As(V) >95%; As(III) typically 50–70%
Highly effective against As(V), which carries a charge the membrane rejects. Much weaker against uncharged As(III).
Caveat: This is the single most important caveat on this page. Reducing groundwater — precisely the water that has an arsenic problem — is dominated by As(III). Pre-oxidation is not optional.
- RO with pre-oxidationEffective
>95% total arsenic
Chlorination or another oxidation step converts As(III) to As(V) before the membrane, after which RO performs as rated.
Caveat: Requires a carbon stage after oxidation to remove the residual oxidant before the membrane.
- Activated aluminaEffective
>90% with correct pH control
A well-established arsenic adsorbent, used in community-scale plants across affected districts.
Caveat: Performance is pH-sensitive, best around pH 5.5–6.0, and media must be replaced on a schedule, not on taste.
- Iron-oxide / iron-hydroxide adsorbent mediaEffective
Granular ferric hydroxide and similar media adsorb both arsenic species well and are less pH-sensitive than alumina.
- Coagulation with iron saltsPartial
Effective at community scale where iron is already present. Not a household technology.
What does not work
Each of these is commonly assumed — or actively marketed — to help with arsenic. None of them do.
Boiling
Arsenic does not evaporate. Boiling reduces water volume and concentrates the arsenic that remains.
UV
UV inactivates organisms. It has no effect whatsoever on a dissolved metalloid.
UF and sediment filters
Pore sizes are orders of magnitude larger than a dissolved arsenic species.
Activated carbon
Standard carbon has very low affinity for arsenic. Carbon blocks specifically certified to NSF/ANSI 53 for arsenic are a different product and are rare in India.
Ion exchange softener
Softening resin targets calcium and magnesium. It does not remove arsenic, and sulphate competition makes anion resins unreliable for it too.
As(III) versus As(V): the distinction the marketing skips
Arsenic exists in water in two forms. Arsenate — As(V) — is negatively charged at normal drinking-water pH. Arsenite — As(III) — is electrically neutral.
That single difference decides whether your purifier works. RO membranes reject charged species efficiently and neutral molecules of similar size much less efficiently. So the same membrane that removes over 95% of As(V) may remove only half of As(III).
The problem is that the two forms are not evenly distributed. Oxygen-rich surface water tends towards As(V). Reducing groundwater — the deep, low-oxygen aquifers where arsenic contamination is a problem in the first place — tends towards As(III).
In other words, the form that RO handles poorly is the form you are most likely to have. A specification sheet quoting an arsenic reduction percentage without stating which species it was tested on has told you nothing usable.
Important
Verify the output, not the brochure
Arsenic is the clearest case on this site for testing treated water rather than trusting a rating. The contaminant is invisible, the failure mode is silent, and the membrane degrades gradually rather than stopping.
The practical routine in an affected area: test raw and treated water at installation, then test treated water annually, and again after any membrane change. That is roughly ₹1,200–3,600 a year. Set against the cost of the system, it is the cheapest part of owning it, and it is the only part that tells you whether the rest is working.
Frequently asked questions
Does RO remove arsenic from water?
Can you taste or smell arsenic in water?
Does boiling water remove arsenic?
What does an arsenic water test cost in India?
Which states in India have arsenic in groundwater?
Sources
Every limit and standard quoted on this page comes from one of the documents below. Where a figure is not in one of these, we have not published it.
Bureau of Indian Standards · published 2012-05-01 · checked 2026-08-01
BIS states two figures for many parameters: an 'acceptable limit' and a 'permissible limit in the absence of an alternate source'. The second is a fallback for communities with no other supply, not a target.
Guidelines for Drinking-water Quality, 4th edition incorporating the 1st addendum
World Health Organization · published 2017-04-01 · checked 2026-08-01
Ground Water Quality in Shallow Aquifers — annual and state reports
Central Ground Water Board, Ministry of Jal Shakti, Government of India · checked 2026-08-01
District-level groundwater quality reporting. Coverage and sampling year vary by state — always check the report date for your district rather than assuming a national figure applies.
NSF/ANSI 58: Reverse Osmosis Drinking Water Treatment Systems
NSF International / American National Standards Institute · checked 2026-08-01
The certification standard that defines verified reduction claims for RO systems. A unit 'using NSF-certified components' is not the same as a unit certified to NSF/ANSI 58.
This page states what published standards say and what the physics of each technology allows. It does not state what is in your supply — no page can. Where the two questions meet is a test result, and every entry in this library says which test to ask for.