The right arsenic water filter for a well depends on which form of arsenic is in the water, and most filter pages never ask. Arsenic comes as arsenite, As(III), and arsenate, As(V). Reverse osmosis, anion exchange and most adsorptive media remove arsenate well and arsenite poorly. If your well carries arsenite, you need an oxidation step first, or a technology that handles both forms. The New Hampshire Department of Environmental Services (NHDES) puts it bluntly: some technologies reduce both forms, but "their efficiency for removal of Arsenic-III is drastically lower." It recommends that "if water tests above 5 ppb for arsenic, a follow-up test for Arsenic-III be conducted before selecting a treatment system."
This guide covers the fix: which technology fits which form and level, what else in your water decides it, where the waste goes, and how to prove it works. For health effects and limits, see the arsenic explainer. If arsenic is one of several problems, the treatment-by-test-result guide shows how to order the units in a treatment train.
Before you buy anything: confirm the result with a second sample, then order a speciation test (As(III) and As(V)) plus pH, iron, manganese, sulfate and silica. The New Jersey Geological Survey says those levels "need to be known when designing your arsenic treatment system." Meanwhile, use bottled or other tested water for drinking and cooking, and ask your doctor or health department if anyone at home is pregnant or very young.
Step 1: find out which arsenic you have
Most lab reports show total arsenic, both forms added together. NHDES says As(III) testing "is available through the state Public Health Laboratory at $35 as well as at private laboratories." Ask the lab how to collect the sample first; in part of the field study below, speciation samples were separated on site with a disposable cartridge. Our arsenic lab finder lists state-certified labs to call.
Why it matters in practice: a study of 250 homes with arsenic treatment, listed by the U.S. Geological Survey (Yang and others, 2020), found that "the higher the untreated well water [As] and the As(III)/As ratio, the higher the rate of treatment failure." In Maine, untreated water at homes where treatment failed had a median As(III) of 14.3 µg/L, against 1.6 µg/L where it worked. Arsenite made up a median 22% of the arsenic at failed homes and 3% at working ones.
Don't assume your state's usual form applies to your well. Maine CDC calls As(V) "the kind of arsenic encountered most of the time in Maine groundwater." Penn State Extension says that in Pennsylvania, "the most common form of arsenic is likely to be arsenite."
If you can't get speciation, pick for both forms. New Jersey's 2007 information circular, written when speciation tests were "difficult and not widely available from commercial labs," concluded that "it is important to choose a treatment system that removes both arsenic species."
The tests to order before you choose
| Test | Why the installer needs it | Source |
|---|---|---|
| Total arsenic, repeated | Confirms the problem and sizes the system | NHDES; NJ Geological Survey; Oregon Health Authority |
| Arsenic speciation, As(III) and As(V) | Decides whether you need pre-oxidation | NHDES; Maryland Dept. of the Environment |
| pH | Adsorptive media lose capacity at higher pH | NHDES; NJ Geological Survey; EPA |
| Iron and manganese | Can foul RO and adsorptive media; enough iron makes oxidation-filtration an option | Maine CDC; NJ Geological Survey; NHDES |
| Sulfate (and total dissolved solids) | Competes with arsenic in anion exchange | EPA; Maine CDC; Maryland |
| Silica and phosphate | Compete with arsenic on iron-based and alumina media | NJ Geological Survey; EPA; Maine CDC |
| Hardness | Hard water can foul an RO membrane or an adsorption system | Maryland; NJ Geological Survey |
Penn State Extension (page updated May 29, 2025) says "The cost of an arsenic test typically ranges from $15 to $50." For speciation, New Hampshire's state lab charges $35 (NHDES, March 2021); ask your lab. If you already have a report, the guide to reading a well test report explains units and detection limits.
The decision table: technology by arsenic form
Each row names its source. "POU" is point of use (one tap, usually the kitchen sink). "POE" is point of entry (whole house).
| Technology | Arsenite, As(III) | Arsenate, As(V) | Usually installed | Waste it makes |
|---|---|---|---|---|
| Reverse osmosis (RO) | Partial. NHDES: removal "is only 60%," so add a pre-oxidation cartridge. EPA: "its efficiency is not consistent" | NHDES: 95%. Maine CDC: "usually about 95%" | POU, under the sink | Reject water: NHDES says "3-4 gallons of water for every one gallon treated," directed "to the septic system or a drywell" |
| Iron-based adsorptive media (granular ferric) | Yes. Maine CDC: "Effective for both trivalent and pentavalent arsenic removal." NHDES: As(III) cuts media life; a pre-oxidation cartridge can extend it "three to five times longer" | Yes | POE tanks or POU cartridges | Spent media. Maine CDC: disposable "as non-hazardous waste." NHDES: "regular trash as long as the vendor has a TCLP test showing that it is nonhazardous" |
| Activated alumina | Maryland: it "does not effectively reduce Arsenic III," so pre-oxidation is required | Yes | POE or POU | Spent media (disposable type); Penn State says it must be replaced "every one to two years" |
| Anion exchange | No. NHDES: "Arsenic-III is not removed because it has no charge" | Yes, if sulfate is low | POE | Salt brine. Maryland: it "can contain high concentrations of arsenic and should be disposed properly" |
| Oxidation-filtration (iron removal with arsenic) | Yes. NHDES: the media "oxidize iron and arsenic simultaneously, also addressing any Arsenic-III if present" | Yes | POE | Backwash water with iron solids |
New Jersey's circular goes further: "Systems that remove only As5 from water are not recommended," naming anion exchange and RO. "They should only be used if As5 has been determined to be the only species in the water."
Pre-oxidation: turning As(III) into As(V)
NHDES: "If Arsenic-III is present, a simple pre-oxidation step can be added to improve overall treatment efficiency." It sorts the oxidants this way:
- Usable: "chlorine, ozone, permanganate and solid manganese dioxide (pyrolusite) media."
- Not effective: "air, UV light, chlorine dioxide."
EPA's 2003 Arsenic Treatment Technology Evaluation Handbook for Small Systems notes that "Some iron based sorbents may catalyze the As(III) to As(V) oxidation and therefore would not require a pre-oxidation step." Ask the installer which kind you're getting.
If chlorine is the oxidant, Maryland's Department of the Environment says a carbon filter to remove the leftover chlorine "is generally desired," and EPA notes that some RO membranes "are sensitive to chlorine."
Point of use: one tap for drinking and cooking
For arsenic alone, a kitchen-tap system is often enough. NHDES: "For most wells, POU arsenic treatment is recommended as the most cost-effective and simplest solution since arsenic is not volatile and there is no known absorption through the skin." The condition is behavior: "all water for cooking, drinking and ice-making should be obtained from this tap."
Reverse osmosis at the sink
RO pushes water through a membrane and keeps a small tank of treated water at its own faucet. Points from the agencies:
- Water use. EPA's handbook gives RO a "Total Water Loss" of 15 to 75%, against 1 to 2% for adsorptive media and anion exchange.
- Hardness. Maryland recommends softening first "when the raw water hardness is greater than 10 grains (170 mg/L)." Maine CDC says iron and manganese "can foul RO systems."
- Parts. Maryland: the membrane "will need to be replaced after 2 to 3 years depending on water use."
- Certification. Vermont asks for NSF/ANSI 58 for RO and NSF/ANSI 53 for adsorptive filters. Read the claim, not just the number. When we searched NSF's public listings on November 3, 2026, every arsenic claim under both standards read "Arsenic (Pentavalent)," with a ceiling of "<=300 ppb" or "<=50 ppb" (all Standard 53 claims were "<=50 ppb"), and we found none for trivalent arsenic. The listings say the product is for water "at or below the referenced concentration and valency of arsenic."
Adsorption cartridges
New Jersey's circular says under-sink ferric cartridges remove both forms, make "only two quarts per minute," and that cartridges "are typically changed once per year." It adds a warning: "it is not uncommon for homeowners to exceed the useful life of the cartridges."
Where one tap stops being enough
NHDES says whole-house treatment "is only necessary if some members of the household (such as young children) are likely to drink water from taps other than the tap where treatment is installed." It recommends whole-house filtration "for arsenic levels of 100 ppb (0.10 mg/L) and higher if some household members would drink on a regular basis from an untreated tap," and warns that above 100 ppb, "ingesting even small amounts of untreated water on a regular basis (such as in the bathtub or from a bathroom tap) can significantly increase the risk of health effects." Two field studies point the same way. A Maine CDC and U.S. CDC study of 167 Maine households that used bottled water or a kitchen-tap system (Smith and others, 2016) found bathing "is not a significant arsenic exposure source," but reducing exposure "is more complicated" above 40 µg/L. There, treated or bottled water matters "not only for drinking, but also for all beverage and food preparation."
An observational New Jersey study of 49 people (Spayd and others, 2015) found lower urinary arsenic after nine months of whole-house treatment than of point-of-use treatment.
Point of entry: whole-house options
Adsorptive media tanks
New Jersey's preferred design uses two tanks in series: a "worker" tank and a "safety" tank, each with "at least one cubic foot of adsorption media," a 5-micron sediment pre-filter, a raw-water sampling tap and a sampling tap between the tanks. The circular says to set the backwash valves to run "at least once per month, each tank on a separate day." NHDES says media is replaced "every one to three years," and that "media longevity is especially affected by the presence of As-III and pH greater than 7.5."
Anion exchange
This works like a softener, but the resin trades chloride for negatively charged ions, and EPA's handbook ranks sulfate ahead of arsenate in what the resin prefers. Run past exhaustion, incoming sulfate "may displace previously sorbed As(V) and nitrate," so treated water can carry more arsenic than the well. Maine CDC: "all of the arsenic on the resin can be released at once." NHDES warns of arsenic "dumping" if salt is not replenished.
Lead warning. NHDES: anion resin "must be followed by a calcite neutralizer to prevent leaching of lead and copper from home plumbing fixtures." "Do not allow installation of anion resin or a mixed cation/anion bed unless a neutralizer is installed, as lead will be leached from the plumbing at potentially toxic levels." Brine goes to the septic system: it "should not be discharged to a drywell."
Oxidation-filtration
NHDES: "if natural iron levels are 0.1 mg/L or higher, iron-arsenic oxidation followed by filtration is the most cost-effective approach to reducing both contaminants on a whole-house basis." It puts the best ratio at "20 parts iron to 1 part arsenic (example, 0.2 mg/L iron to 10 ppb arsenic)," with pH 7 to 7.5. EPA's handbook gives the same 20:1 minimum and says these conditions "customarily result in an arsenic removal efficiency of 80-95%." Penn State says arsenic removal by ordinary iron filters "will vary considerably," so test. The iron guide covers the iron side.
Competing ions: what else in your water decides it
Most numbers below come from EPA's 2003 handbook, written for small public water systems. Treat them as questions for an installer, not pass/fail lines.
| In your water | Technology it affects | Level the source gives | Source |
|---|---|---|---|
| Sulfate, total dissolved solids | Anion exchange | Not "economically viable" over 50 mg/L sulfate or 500 mg/L TDS. Maine CDC: "EPA recommends" under 25 mg/L. Maryland: efficiency reduced above 50 ppm | EPA handbook; Maine CDC; Maryland |
| Phosphate | Iron-based media | "Each 0.5 mg/L increase in phosphate above 0.2 mg/L will reduce adsorption capacity by roughly 30%"; optimal under 1 mg/L | EPA handbook |
| Silica | Activated alumina; iron oxide media; oxidation-filtration | Problem level for alumina 30 mg/L. Silica reduces iron oxide filters' effectiveness (Maine) and weakens arsenic removal on iron particles (EPA) | EPA handbook; Maine CDC |
| pH | Adsorptive media; activated alumina | Iron-based media optimal at pH 6 to 8.5 (EPA). Media life "especially affected" by pH above 7.5 (NHDES). Alumina optimal at 5.5 to 6 (EPA) | EPA handbook; NHDES |
| Iron, manganese, hardness | Adsorption systems | Iron above 0.5 ppm, manganese above 0.05 ppm, hardness above 300 ppm "may foul an arsenic adsorption system" | NJ Geological Survey |
The field study below adds that adsorption to iron hydroxides "is very sensitive to pH, with higher pH (>8) favoring As(V) desorption." The pH and TDS guide covers that test.
Waste and regeneration: where the arsenic goes
Removed arsenic leaves in one of four streams:
- Brine from anion exchange. In an EPA study cited in the handbook, arsenic in spent brine "ranged from 1.83 mg/L to 38.5 mg/L, with an average value of 16.5 mg/L." For public systems, EPA says liquid waste above 5.0 mg/L arsenic "would therefore be classified as a hazardous waste." New Jersey calls the disposal of extracted arsenic "to the environment near the home" one reason anion exchange is "problematic."
- Spent media. Iron-based media holds onto arsenic. Maryland's 2008 summary, citing a New Jersey evaluation, says ferric media "did not return captured arsenic to the environment through regeneration (backwash)." Maine CDC and NHDES both treat spent media as ordinary waste if it is non-hazardous; NHDES wants a vendor TCLP test to show it.
- RO reject water. NHDES: "directed to the septic system or a drywell."
- Backwash water. New Jersey says to pipe the backwash line "to a suitable disposal location according to local plumbing codes."
One more place arsenic can hide: New Jersey's circular says minerals built up in a water heater "may release arsenic back into the treated water," and that homeowners with high arsenic "should consider testing their hot water after installing a whole-house treatment system."
Why real systems fail
In the Maine and New Jersey study, treatment cut median arsenic "from 71.7 to 0.8 μg/L" in Maine and "from 8.6 to 0.2 μg/L" in New Jersey. Still, 19% of Maine systems left treated water above EPA's 10 µg/L limit (a limit for public water systems; private wells are not covered), and 16% of New Jersey systems above that state's 5 µg/L standard. In Maine, failure rose with the arsenic going in:
Show the numbers
| Item | Value | Note |
|---|---|---|
| Untreated well 10 to 50 µg/L | 9 % of systems with treated water above 10 µg/L | 5 of 54 systems |
| Untreated well 50 to 100 µg/L | 16 % of systems with treated water above 10 µg/L | 7 of 43 systems |
| Untreated well above 100 µg/L | 31 % of systems with treated water above 10 µg/L | 17 of 55 systems |
| All Maine systems in the study | 19 % of systems with treated water above 10 µg/L | 29 of 156 systems |
156 Maine households with arsenic treatment, each sampled once. Failure = treated kitchen-tap water above EPA's 10 µg/L MCL (a limit for public water systems; private wells are not covered). The 4 homes with untreated water under 10 µg/L (no failures) are not shown as a bar, so the three bins total 152. Source: Yang Q. et al., Reduction in drinking water arsenic exposure and health risk through arsenic treatment among private well households in Maine and New Jersey, USA, Science of the Total Environment (2020), Table 3, accessed 2026-11-03.
Two other findings are worth acting on. "POE systems failed less than POU systems," though for systems alone in a home the difference was not statistically significant. And upkeep mattered: among 93 Maine homes that reported who looked after their system, those maintained by homeowners failed 20% of the time (13 of 65), against 7% (2 of 28) for those maintained by a vendor.
Prove it works: testing before and after
Every agency we checked says to test treated water. They differ on when.
| Source | After installation | Ongoing |
|---|---|---|
| NHDES (fact sheet accurate as of March 2021) | "quarterly arsenic testing of the treated water for the first year of treatment" | "semi-annually after that"; check the pre-oxidant or salt feed tank weekly |
| New Jersey Geological Survey (2007), two-tank whole-house system | Sample between the tanks "a few days after installation"; it "should find less than 3 ppb" | Between the tanks "every six months"; at 5 ppb, schedule media replacement in tank 1, then retest a few days after |
| Oregon Health Authority | "Continue regular testing with a treatment system" | "Every year at the faucet that you drink out of" and "Every 3 years at the wellhead" |
| Vermont Department of Health | "Retest your drinking water for arsenic after any treatment system is installed" | No separate schedule for treated water; general advice for every well is arsenic every five years |
| Washington State Department of Health | "After installation and routine maintenance, your water should be tested" | Not stated separately |
| Wisconsin DNR (2017); Wisconsin DHS (page revised Feb. 24, 2026) | DNR: "verify with sampling to make sure the system continues to operate" | DHS general well advice: "Test once a year if arsenic was present in previous tests or you live in Outagamie, Winnebago, or Brown county"; otherwise at least every five years |
New Jersey says to sample "while stressing the system": run the water for 10 minutes at 3 to 5 gallons per minute (a bathtub and a sink running cold at the same time usually does it) before you fill the bottle. Keep a raw-water sample in the routine too, as Oregon's wellhead test does. For other contaminants, see how often to test well water.
What agencies have said treatment costs
Agency cost figures are scarce and dated, and none is adjusted for inflation here. Get current quotes, including maintenance, as Vermont's health department advises.
| System | NHDES fact sheet (accurate as of March 2021) | New Jersey circular (2003 cost survey) |
|---|---|---|
| POU adsorption cartridge | Equipment "begin at $150"; cartridges "$80 to $90" each | $365 installed; $0.32 per day upkeep |
| POU reverse osmosis | No price given | $700 installed; $0.33 per day |
| POE adsorptive media | $1,500 to $3,000 installed; replacement "about $1,000/CF" (per cubic foot of media) | $2,740 installed; $0.67 to $1.00 per day |
| POE anion exchange | About $1,500 to $3,000 with a calcite neutralizer; salt "about $100" | $2,000 installed; $0.27 per day |
| POE oxidation-filtration | Whole-house single step: "$1,500 to $3,000" | Not listed |
Maine CDC's arsenic page (undated) says a kitchen RO system usually costs "between $800 and $1,500," varying with water chemistry. For test prices, see our well water testing cost guide.
State programs and rules that apply
- Wisconsin. The DNR's Well Compensation Grant Program can help "replace, reconstruct or treat" a contaminated private well. Arsenic qualifies at "50 ug/l (parts per billion)" or higher. Family income "may not exceed $65,000," and the program "will pay 75% of eligible costs up to $16,000," a maximum grant of $12,000. Work done before DNR approval "will not be reimbursed." The DNR's 2017 brochure says to check the state's approved treatment-device list before buying. Labs in Wisconsin.
- New Jersey. The state set 5 ppb as its arsenic drinking water standard, effective January 2006, and its geological survey recommends whole-house granular ferric adsorption as the first choice.
- Maine. Maine CDC's arsenic flyer (March 2020) offers help by phone: "We can help you decide how to solve the problem." The Maine Monitor reported on August 17, 2026 that MaineHousing's well water abatement program for lower-income homeowners (arsenic, and since 2022 other contaminants) had spent over $385,000 of its latest $400,000, and that lawmakers had not introduced new funding. Ask MaineHousing whether it is open. Labs in Maine.
- Vermont. The health department says funding "is available to help you pay for water treatment" and lists loan and grant programs. Arsenic testing in Vermont.
- Maryland. Maryland's 2008 summary says a newly built well needs a Certificate of Potability, and if arsenic "is found to be above 0.010 mg/L then the water must be treated."
State arsenic lab pages: Oregon, Minnesota, Washington, or find labs certified for arsenic in your state.
Worked example: "Arsenic 0.032 mg/L, As(III) 0.024 mg/L"
Your confirmed report, from an untreated tap, reads: total arsenic 0.032 mg/L, As(III) 0.024 mg/L, iron 0.8 mg/L, pH 7.2, sulfate 30 mg/L.
- Units and form. 0.032 mg/L is 32 µg/L. As(III) is 24 µg/L, three-quarters of the total.
- Pre-oxidation. As(III) is present, so NHDES's advice is a pre-oxidation step, and New Jersey and Maryland rule out As(V)-only systems (plain RO, anion exchange). From the decision table, the candidates are oxidation before RO, iron-based media, or oxidation-filtration.
- The iron route. Iron is above NHDES's 0.1 mg/L, and the ratio is 800 µg/L iron to 32 µg/L arsenic, 25 to 1, above the 20:1 that NHDES and EPA give. pH 7.2 sits inside NHDES's 7 to 7.5. Whole-house oxidation-filtration fits this water on paper.
- The one-tap route. 32 µg/L is under the Maine study's 40 µg/L and NHDES's 100 ppb whole-house line. RO with a pre-oxidation cartridge, or a ferric cartridge, at the sink also fits if everyone uses that tap for drinking, cooking and ice.
- Prove it. Test treated water quarterly for a year, then twice a year (NHDES), with a raw sample alongside.
The reverse case: 0.015 mg/L, all As(V), sulfate 80 mg/L. No oxidation step is needed, and sulfate is past EPA's 50 mg/L line for anion exchange. A kitchen RO unit fits, and an NSF/ANSI 58 "Arsenic (Pentavalent) <=50 ppb" claim covers 15 µg/L. In neither case is a softener the answer: New Jersey's circular says softeners "do not remove arsenic." If one is offered, read water softener vs filter.
Next step
Get a confirmed total, a speciation result, and pH, iron, manganese, sulfate and silica from a state-certified lab. Then ask each installer which row of the decision table your water falls in, and why. Start with the arsenic lab finder.
Questions people ask
What is the best arsenic water filter for well water?
Does reverse osmosis remove arsenic from well water?
Do I need a whole-house arsenic filter?
Will a water softener remove arsenic?
How do I know if my well has arsenic III or arsenic V?
How often should I test after installing an arsenic filter?
How much does arsenic treatment cost?
- New Hampshire Department of Environmental Services, Arsenic in New Hampshire Well Water, DWGB-3-2 (2021; "accurate as of March 2021") (accessed 2026-11-03)
- U.S. EPA, Arsenic Treatment Technology Evaluation Handbook for Small Systems, EPA 816-R-03-014 (July 2003) (accessed 2026-11-03)
- New Jersey Geological Survey, Information Circular: Arsenic Water Treatment for Residential Wells in New Jersey (prepared by Steven Spayd, 2007; costs from a 2003 survey) (accessed 2026-11-03)
- Maryland Department of the Environment, Arsenic Water Treatment for Individual Wells in Maryland (February 2008; copy hosted by University of Maryland Extension) (accessed 2026-11-03)
- Maine CDC Drinking Water Program, Arsenic in Drinking Water (treatment technologies, pros and cons, checklist) (accessed 2026-11-03)
- Maine CDC, Article summary: Assessing arsenic exposure in households using bottled water or point-of-use treatment systems (Smith AE et al., Science of the Total Environment 544, 2016) (accessed 2026-11-03)
- Maine CDC, Have you tested your well water for arsenic? (flyer, March 2020) (accessed 2026-11-03)
- USGS Publications Warehouse, Yang Q. et al., Reduction in drinking water arsenic exposure and health risk through arsenic treatment among private well households in Maine and New Jersey, USA (Science of the Total Environment, June 26, 2020) (accessed 2026-11-03)
- Yang Q. et al. (2020), full text on PubMed Central (Tables 3 and 5; speciation, POE vs POU, installation and maintenance) (accessed 2026-11-03)
- Spayd SE, Robson MG, Buckley BT, Whole-house arsenic water treatment provided more effective arsenic exposure reduction than point-of-use water treatment at New Jersey homes with arsenic in their well water (Science of the Total Environment 505, 2015), PubMed Central (accessed 2026-11-03)
- Penn State Extension, Removal of Arsenic from Wells in Pennsylvania (updated May 29, 2025) (accessed 2026-11-03)
- Vermont Department of Health, Arsenic in Drinking Water (accessed 2026-11-03)
- Vermont Department of Health, How to Treat Your Drinking Water (accessed 2026-11-03)
- Vermont Department of Health, Funding Assistance for a Healthier Home (accessed 2026-11-03)
- Washington State Department of Health, Arsenic and Your Private Well (accessed 2026-11-03)
- Oregon Health Authority, Domestic Well Safety Program, Routine Domestic Well Testing (accessed 2026-11-03)
- Wisconsin DNR, Arsenic in Drinking Water Wells, PUB-DG-062 (2017) (accessed 2026-11-03)
- Wisconsin Department of Health Services, Arsenic (last revised February 24, 2026); fact sheet Arsenic in Private Wells, P-45012 (09/2023) (accessed 2026-11-03)
- Wisconsin DNR, Well Compensation Grant Program (accessed 2026-11-03)
- NSF, Certified Drinking Water Treatment Units: NSF/ANSI 58 listings (Arsenic (Pentavalent) Reduction claims) (accessed 2026-11-03)
- NSF, Certified Drinking Water Treatment Units: NSF/ANSI 53 listings (Arsenic (Pentavalent) Reduction claims) (accessed 2026-11-03)
- The Maine Monitor, Maine struggling to get homeowners to test and treat private wells for arsenic (John Forbes, August 17, 2026) (accessed 2026-11-03)