Home Water Filtration: How Each Method Works and How to Choose

The best water filter is the one matched to your water. Reverse osmosis is the most thorough home technology for PFAS, lead, and most dissolved contaminants. Carbon handles chlorine and many chemicals and leaves minerals in place. Ultraviolet light is the right tool when the worry is germs, not chemicals.

Of all the questions visitors ask at the nursery, what filter should I buy may be the most common. Friends and family ask too, usually after reading something alarming. The honest answer is always the same. It depends on what is in your water, and on what you want that water to do for you over the next thirty years.

That last part matters. Lifespan is how long we live. Healthspan is how long we live well: independent, clear-headed, able to work in the garden and keep up with the grandchildren. A 2024 analysis of 183 countries, using data through 2019, found the United States had the widest gap between the two. Americans spent an average of 12.4 years living with disease or disability (1). We drink water every day for decades. Choosing how to treat it is a healthspan decision, not a gadget purchase.

This piece builds on the articles on PFAS and microplastics, and on chlorine, chloramine, and disinfection byproducts. Here I explain how each method works, what it can and cannot remove, how to read a certification, and which systems are worth the money if the listing matches the claim.

1. Start with your water, not the filter

Buying a filter before you know what is in the water is like prescribing before you have made a diagnosis. Two households can need completely different systems.

• On city water, read the annual Consumer Confidence Report. It tells you which disinfectant is used, whether PFAS or other contaminants were detected, and how close the results came to the legal limits.
• Lead is a separate question. Utilities treat water at the plant. Lead usually enters from service lines and household plumbing on the way to the tap. EPA and CDC agree there is no known safe level of lead in a child’s blood. EPA estimates drinking water can account for 20 percent or more of a person’s total lead exposure (2). If the house has older plumbing, test at the tap, first draw.
• On a private well, nobody tests for you. Penn State Extension recommends testing every private supply for total coliform every year, and a broader chemical panel at least once: nitrate, arsenic, metals, and, given what we now know, PFAS (3). Coliform is more common in springs and shallow dug wells than in deeper drilled wells. A drilled well is not immune.

Once you know the problem, match the technology to it.

2. How the main methods work

Every home filter works by one of four mechanisms: straining particles, holding chemicals on a surface, trading one ion for another, or forcing water through a membrane so tight that dissolved substances stay behind. Ultraviolet light removes nothing. It inactivates germs.

Sediment filters. A pleated or spun cartridge strains sand, rust, and silt. Five microns is a common rating. They protect the equipment downstream. They do nothing for dissolved chemicals. On well water, sediment is usually the first stage.

Activated carbon. Charcoal treated to create an enormous internal surface. Chlorine, many disinfection byproducts, and many organic chemicals stick to that surface. Granular carbon is loose. Water moves through quickly, so contact time is short. A carbon block is compressed. Water has to work through fine pores, so contact time is longer, and a certified block can also trap lead, cysts, and larger microplastics. Carbon is inexpensive, needs no electricity, wastes no water, and leaves minerals in place.

Its weakness for PFAS is inconsistency. In a 2020 study of filters in North Carolina homes, Herkert and colleagues found that under-sink reverse osmosis and two-stage filters removed the PFAS they measured nearly completely. Ordinary activated-carbon filters were variable. Long-chain PFAS came out at roughly 60 to 70 percent removal. Short-chain compounds were closer to 40 percent. GenX, a replacement chemical, was removed at more than 74 percent in the dual-stage and reverse-osmosis units they tested, not at the same rate as the long-chain compounds (4). EPA lists granular activated carbon, anion exchange, and high-pressure membranes such as reverse osmosis among the treatments that can work when the system is designed and maintained for the job (5).

Catalytic carbon. If the utility uses chloramine instead of chlorine, ordinary carbon is a poor match. Catalytic carbon is surface-modified to break chloramine down. Look for a listing under NSF/ANSI 42 for chloramine reduction, not a chlorine claim reused as if it were the same chemical (6).

Ion exchange. Tiny beads swap ions. A softener uses cation resin to trade calcium and magnesium for sodium or potassium. Anion resin works the other way. Because many PFAS carry a negative charge, anion exchange is one of the treatments EPA lists for them (5). Some pitchers combine ion exchange with carbon and can strip minerals along with contaminants.

Reverse osmosis. Pressure pushes water through a membrane. Water molecules pass. Most dissolved salts, metals, and larger molecules do not, and they leave in a wastewater stream. Home units nearly always pair the membrane with sediment and carbon stages. In the Herkert study, every under-sink reverse-osmosis unit and every two-stage filter tested cut the measured PFAS sharply, including GenX at more than 74 percent (4). That is the strongest home evidence we have. It is still one region and a modest number of units.

The trade-offs are real.

• Wastewater. Older tank systems can send about three gallons down the drain for every gallon filtered (7). Newer tankless units are more efficient. Read the rated ratio, and remember a rating is not a measurement in your kitchen.
• Minerals. Reverse osmosis removes calcium and magnesium along with the contaminants. I come back to that below.
• Chloramine. The membrane alone does not reliably stop it. A carbon stage ahead of the membrane matters on chloraminated city water (6).
• Idle time. In EPA point-of-use testing, one reverse-osmosis unit’s PFAS reduction slipped when it was run only now and then (8). If the system has sat, run the tap briefly before you fill a glass.

Distillation. A distiller boils water and condenses the steam. Metals and minerals stay behind. Chemicals that evaporate easily can travel with the steam (9). Distillers are slow and use a lot of electricity. They suit small volumes.

Ultrafiltration and microfiltration. Looser than reverse osmosis. They block particles, bacteria, cysts, and larger microplastics, and they let dissolved minerals and most dissolved chemicals through. They waste little water. In a 2023 laboratory study of pour-through devices, the unit with a physical membrane gave the most consistent removal of plastic fragments and fibers larger than 10 micrometers (10).

Method	Removes well	Does not remove	Wastes water?	Power?
Sediment	Sand, rust, silt	Dissolved chemicals, germs	No	No
Granular carbon	Chlorine, taste, many organics	Minerals, nitrate, most germs, short-chain PFAS	No	No
Certified carbon block	The above, plus lead, cysts, some PFAS and microplastics when certified	Nitrate, fluoride, dissolved salts	No	No
Catalytic carbon	Chloramine, when certified	Minerals, nitrate	No	No
Ion exchange	Hardness on cation resin; some PFAS on anion resin	Depends on the resin	Softeners use rinse water	Sometimes
Reverse osmosis	PFAS, lead, arsenic, fluoride, nitrate, most dissolved solids, larger microplastics	Chloramine without a carbon stage ahead	Yes	Some models
Distillation	Metals, minerals, nitrate	Some volatile chemicals	Little	Yes, a lot
Ultrafiltration	Particles, bacteria, cysts, larger microplastics	Most dissolved chemicals	Little	No
UV, Class A	Inactivates bacteria, viruses, cysts	Any chemical or particle	No	Yes

3. Microplastics

Most good home filters catch the larger microplastic particles. Very few can catch nanoplastics, and only a tight membrane can in principle.

The Toronto study challenged three pour-through devices with fragments and fibers larger than 10 micrometers. Removal ran from 71 percent to 100 percent. The membrane unit did best (10). That is reassuring for the larger particles. It says nothing about the nanoplastics Columbia found in bottled water, most of them well below one micrometer.

The NSF/ANSI 401 microplastics claim is tested with particles 0.5 to 1 micrometer across (11). Nanoplastics are smaller. A reverse-osmosis membrane is far tighter than carbon, which is why I treat it as the best home option for plastics. Direct evidence on home systems and true nanoplastics is still thin.

• Look for a certified microplastics claim, not the words on the box.

• Store filtered water in glass or stainless steel. There is little sense in filtering plastic out and pouring it into a plastic jug.

4. Ultraviolet light, when germs are the concern

If the water comes from a well, spring, cistern, or surface source, bacteria can be a bigger immediate risk than any chemical. Ultraviolet treatment is the chemical-free way to handle that, after the well itself is sound.

A positive coliform test can come from a storm or from pump work. Penn State recommends shock chlorinating, fitting a sanitary well cap, and retesting. If bacteria come back, continuous disinfection is warranted (3).

Water flows past a lamp inside a quartz sleeve. The light disrupts the ability of bacteria, viruses, and cysts to reproduce. It adds no chemical and does not change the taste.

• Buy Class A, certified to NSF/ANSI 55. Class A is for disinfecting contaminated water, at a dose of 40 mJ/cm2. Class B only reduces harmless bacteria in water that is already disinfected (9, 12). For a well, Class A is the one you want.

• The water has to be clear. Sediment, iron, or cloudiness shades the germs. A sediment prefilter goes ahead of the unit. Installers generally want iron below about 0.3 mg/L and hardness below about 7 grains per gallon, or the sleeve coats (3, 13).

• The lamp fades while it still glows. Replace it about once a year. Clean the sleeve. A sensor that warns when output drops is worth having (3).

• Penn State does not recommend ultraviolet treatment where total coliform exceeds 1,000 colonies per 100 mL, or fecal coliform exceeds 100 (3). That water needs a source fix, not a lamp.

Ultraviolet light removes no chemical, metal, PFAS, or microplastic, and it leaves no disinfectant in the pipes downstream (3). On a well, the usual order is sediment, then iron or hardness treatment if the water needs it, then ultraviolet light last before the water enters the house. For chemicals, add reverse osmosis or certified carbon at the kitchen tap.

5. How to read the certifications

There are no federal performance rules for home filters. NSF/ANSI standards are voluntary. The numbers are the order they were written, not a ranking (9).

Standard	What a listing under it can cover
NSF/ANSI 42	Aesthetic effects: chlorine, taste, odor. Not a health claim.
NSF/ANSI 53	Health contaminants named on the sheet: lead, cysts, specific chemicals. PFOA and PFOS reduction is certified here for many carbon units.
NSF/ANSI 55	UV. Class A is for disinfecting microbiologically unsafe water, at 40 mJ/cm2. Class B only reduces non-pathogenic bacteria in already treated water.
NSF/ANSI 58	Reverse osmosis. Read which contaminants the model is listed for. Total dissolved solids is not the same as PFAS.
NSF/ANSI 62	Distillation.
NSF/ANSI 177	Shower filters, free chlorine only. Not chloramine.
NSF/ANSI 244	Supplemental microbiological protection for already treated city water. Not a well disinfectant.
NSF/ANSI 401	Named emerging contaminants, including a microplastics claim tested at 0.5 to 1 micrometer. Not nanoplastics.

Three rules.

• Match the contaminant, not just the standard. A unit can be listed under Standard 58 for total dissolved solids and not for PFOA. Read the performance data sheet for the exact model.

• NSF, WQA, and IAPMO all count. All three can certify against the same NSF/ANSI standards. A WQA Gold Seal or an IAPMO listing is not a lesser mark.

• Tested to is not certified to. A one-time lab report the company paid for is not a listing with audits and retesting. Search the certifier’s database by model number before you buy.

6. A filter is only as good as its upkeep

An old filter can be worse than none. In the Herkert study, some whole-house carbon systems increased PFAS in the finished water. Four of eight point-of-entry systems did that (4). The paper did not prove the cause in every case. Saturation, channeling, and a spent bed that releases what it caught are the practical reading. Carbon saturates. Membranes foul. A UV lamp fades while it still glows.

• Replace cartridges on schedule, by date or gallons, whichever comes first. Write the change date on the housing.

• Replace reverse-osmosis membranes at the maker’s interval.

• Change ultraviolet lamps yearly and clean the quartz sleeve (3).

• Think twice about whole-house carbon on city water. Stripping the residual disinfectant where water enters the house can leave the pipes more open to bacterial growth (4). Filtering at the kitchen tap avoids that.

7. Matching a system to the situation

Your situation	What I would look at first
City water, chlorine, no special finding	A carbon block certified to NSF/ANSI 42 and 53, at the pitcher, faucet, or under the sink
City water with chloramine	Catalytic carbon certified for chloramine, or RO with a real carbon stage ahead of the membrane
PFAS on the report or the well test	Under-sink or countertop RO listed for PFOA and PFOS, or carbon certified for that reduction
Older plumbing, possible lead	A filter certified for lead under 53 or 58, plus a first-draw tap test
Private well, bacteria	Fix the wellhead first. Then sediment, then a Class A UV unit for the house
Private well, chemical concerns	Test first. RO at the kitchen tap for drinking and cooking
Renting, no plumbing changes	Countertop RO, or a pitcher certified for the contaminant you care about
Tight budget	A certified pitcher, filters changed on time, beats a neglected expensive system

8. Systems worth a look

Models and listings change often, and new products keep coming out. Before you buy, check that the exact unit carries the certification described here.

Some of the links below are Amazon affiliate links. If you buy through them, Mindful Living Sanctuary may earn a small commission, at no extra cost to you. I would make the same short list without the links. Certifications change. Confirm the current listing for the exact model before you buy.

Countertop reverse osmosis, no plumbing. AquaTru Classic and Classic Smart countertop models (AT2000 series) are listed by IAPMO for PFOA and PFOS reduction under NSF/ANSI 53; older sheets used NSF Protocol P473 (14). Confirm the current sheet for the exact unit, including whether 42, 53, 58, and 401 are on that model. A countertop unit is the practical choice for a renter. Check price on Amazon

Under-sink reverse osmosis. The iSpring RCC7AK is a tank-style system with a remineralization stage. Retailers describe an NSF/ANSI 58 listing for total dissolved solids and several metals. Confirm that sheet yourself, and do not assume the listing includes PFOA unless PFOA is named (15). The Waterdrop G3P800 is tankless and more water-efficient than older tank units. IAPMO has listed related Waterdrop models under 42, 53, and 58. On at least some of those listings, Standard 58 covers total dissolved solids, and the longer contaminant list is the company’s own data. Read the sheet. iSpring on Amazon Waterdrop G3P800 on Amazon

Under-sink carbon, if you want to keep minerals. Aquasana’s Claryum under-sink line, including the current AQ-6300M, is listed for chlorine, chloramine, lead, and PFOA/PFOS reduction. That is the maker’s summary. Confirm the NSF or WQA listing for that model before you treat it as certified (16). Check price on Amazon

Pitcher. The Brita Metro 6-cup pitcher with the Elite cartridge (the blue one, model OB06) is certified by WQA to NSF/ANSI 42, 53, and 401, including about 99.5 percent lead reduction and a microplastics claim at 0.5 to 1 micrometer. PFOA and PFOS reduction, about 98 percent on Brita’s data sheet, is an IAPMO listing, and it applies only in named Brita pitchers; Metro is one of them (17). The original white Brita filter does not carry the lead, PFAS, or microplastics claims. Elite filters are rated about 120 gallons, or six months. Check price on Amazon

Whole-house ultraviolet light for a well. The Viqua Pro series includes NSF/ANSI 55 Class A units, with an intensity sensor and a flow rating in gallons per minute. Viqua also sells Class B models. Confirm the class on the exact unit. Plan on a plumber or a water-treatment installer (18). Check price on Amazon

If chlorine in the shower bothers you, look for NSF/ANSI 177. That standard covers free chlorine only. It says nothing about chloramine (9).

9. A word about Berkey and other gravity filters

Many homesteaders own a Berkey. I understand the appeal. No plumbing, no power, useful in an outage. The Black Berkey elements are not certified to NSF/ANSI standards. The company relies on laboratory tests it commissions (19).

EPA issued stop-sale orders in December 2022 and May 2023 covering Black Berkey elements and several Berkey systems. The stated basis was microbiological claims that brought the products under federal pesticide law, as unregistered pesticide devices, including the silver used in the elements. Both dockets were still listed as active in September 2026, with no later filings (20). Those orders are about registration and marketing claims. They are not a performance test. Without a certification, I cannot verify the reduction claims. If you want a gravity filter, choose one listed for the contaminants you care about.

10. Our situation: a drilled well and reverse osmosis

We have a drilled well here in Oswego County, and we drink a lot of water. For drinking and cooking we run it through reverse osmosis at the kitchen tap. That is the broadest protection against what we cannot see or taste, including PFAS, metals, and the larger microplastics. It also means we are not waiting to find out what might show up in the groundwater later.

The question I get most is minerals, and it is a fair one. A World Health Organization review, drawing on work through 2005, concluded there was enough evidence to support an inverse relationship between magnesium in drinking water, and possibly calcium, and deaths from ischemic heart disease (21). Most of that evidence is population-level. It cannot prove cause and effect. For most people eating a varied diet, food is the main source of these minerals.

We eat a whole-food, plant-based diet built on leafy greens, beans, whole grains, nuts, and seeds, which are among the richer sources of magnesium. I am comfortable drinking reverse-osmosis water. If your diet is less mineral-rich, or you prefer the taste, a remineralization cartridge that adds a little calcium and magnesium back is an inexpensive stage on most systems.

A well also puts the testing on us. No filter substitutes for knowing what is in the water.

11. Honest caveats

• The home PFAS study is one region. Herkert and colleagues tested filters in North Carolina homes. The pattern matches EPA’s treatment research. Your water and your model can differ (4).

• The microplastics certification stops at half a micrometer. Nanoplastics are smaller, and home data on them are thin (11).

• The mineral question is not settled. The WHO conclusion rests mostly on population studies (21).

• Prices and listings change. Check the current sheet for the exact model.

• A filter does not fix the source. A lead service line, a poorly sealed wellhead, or a contaminated aquifer is best addressed at the source when you can.

• This is not medical advice. I am a veterinarian and a board-certified internist. I am not your physician.

12. Where we land

Clean water is one of the simpler investments in healthspan I know. It does not take heroics. It takes the right information and a little routine maintenance.

1. Know your water first. Read the utility report, or test the well.

2. Match the technology to the problem. Carbon for chlorine and taste. Catalytic carbon for chloramine. Reverse osmosis for PFAS and the broadest cut. Class A ultraviolet light for germs.

3. Buy certified, and check the claim for your contaminant in the NSF, WQA, or IAPMO listing.

4. Keep up with maintenance. A spent filter can make the water worse.

5. Do not let perfect be the enemy of good. A certified pitcher, filters changed on time, is a real improvement over nothing.

13. Further reading

If you open two, open the Herkert filter study and the NSF standards page.

• Herkert et al., Environmental Science & Technology Letters, 2020

• NSF, standards for water treatment systems

• EPA, basic information about lead in drinking water

• Penn State Extension, coliform bacteria

14. On our website

• PFAS and Microplastics in Tap Water: What is in the water, and what the health evidence actually supports.

• Chlorine, Chloramine, and Disinfection Byproducts: What treatment plants add on purpose, and what those additives leave behind.

15. References

1. Garmany A, Terzic A. Global healthspan-lifespan gaps among 183 World Health Organization member states. JAMA Netw Open. 2024;7(12):e2450241. Data through 2019. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2827753

2. U.S. Environmental Protection Agency. Basic information about lead in drinking water. https://www.epa.gov/ground-water-and-drinking-water/basic-information-about-lead-drinking-water

3. Swistock BR, Clemens S, Sharpe WE. Coliform bacteria. Penn State Extension. Updated September 8, 2025. https://extension.psu.edu/coliform-bacteria

4. Herkert NJ, Merrill J, Peters C, et al. Assessing the effectiveness of point-of-use residential drinking water filters for perfluoroalkyl substances. Environ Sci Technol Lett. 2020;7(3):178-184. https://pubs.acs.org/doi/10.1021/acs.estlett.0c00004

5. U.S. Environmental Protection Agency. Reducing PFAS in drinking water with treatment technologies. https://www.epa.gov/sciencematters/reducing-pfas-drinking-water-treatment-technologies

6. Water Quality Association. Chloramine fact sheet. https://wqa.org/resources/chloramine/

7. Madison Water Utility. Household filter study: PFAS. Notes older reverse-osmosis reject ratios. https://www.cityofmadison.com/water/water-quality/perfluorinated-compounds/household-filter-study-pfas

8. U.S. Environmental Protection Agency. Effectiveness of point-of-use and point-of-entry systems to remove PFAS. https://www.epa.gov/sciencematters/epa-researchers-investigate-effectiveness-point-usepoint-entry-systems-remove-and

9. NSF. Standards for water treatment systems. https://www.nsf.org/consumer-resources/articles/standards-water-treatment-systems

10. Cherian AG, Liu Z, McKie MJ, Almuhtaram H, Andrews RC. Microplastic removal from drinking water using point-of-use devices. Polymers. 2023;15(6):1331. https://www.mdpi.com/2073-4360/15/6/1331

11. NSF/ANSI 401 microplastics reduction claim, particles 0.5 to 1 micrometer. Example of a certified claim: Coway press release, 2024. https://www.prnewswire.com/news-releases/coway-water-purifiers-certified-for-microplastic-removal-302102266.html

12. NSF/ANSI 55 Class A dose of 40 mJ/cm2. Summary: https://espwaterproducts.com/pages/nsf-certified-uv-water-purification-systems

13. Mid-Atlantic Water. Ultraviolet pretreatment notes, iron and hardness. https://midatlanticwater.net/collections/ultraviolet-uv-water-disinfection-systems

14. IAPMO R&T product listing, AquaTru drinking water treatment units, PFOA and PFOS, file W-11604, current as of July 5, 2023. Confirm the live listing for the model you buy.

15. Retailer summaries of iSpring RCC7AK and Waterdrop G3P800 certifications. Confirm the certifier’s sheet for the exact model. https://www.drinkingwaterfilter.com/types/best-reverse-osmosis-system/

16. Aquasana. Manufacturer summary of Claryum under-sink certifications. Confirm on the NSF or WQA listing. https://www.aquasana.com/info/new-york-times-publication-wirecutter-names-aquasana-best-under-sink-water-filter-pd.html

17. Brita. Performance data sheet, Elite filter, model OB06. PFOA/PFOS claim limited to named pitchers. https://www.brita.ca/wp-content/uploads/2024/07/Brita_Performance_Data_Sheet_Elite_PFOA_PFOS_021323_EN_FR.pdf

18. Mid-Atlantic Water. Viqua PRO20 product page. Confirm Class A on the unit. https://midatlanticwater.net/products/viqua-pro20-uv-system

19. WaterFilter.guide. Berkey review, certification status. Secondary source. https://waterfilter.guide/water-filter-brands/berkey

20. U.S. Environmental Protection Agency, Region 8. Stop Sale, Use, or Removal Order, Berkey International LLC, FIFRA-08-2023-0038, May 8, 2023. A December 2022 order, FIFRA-08-2023-0011, covers related elements. Both dockets still listed active as of September 2026. https://yosemite.epa.gov/OA/RHC/EPAAdmin.nsf/Filings/70BF71F4C6BC372C852589AA0000E556

21. World Health Organization. Nutrients in Drinking Water. Geneva: WHO; 2005. Includes Kozisek F, Health risks from drinking demineralised water. https://iris.who.int/server/api/core/bitstreams/ee809f3c-d909-4138-9242-002bce630853/content