Filtration, explained
Everything we know about filters, and everything we don’t.
Aurmina binds contaminants and drops them out of the water. Something still has to separate what you drink from what settled — and that something is a filter, or a steady hand and a second container.
So filtration is part of using this product.
Our recommendation, up front
Below you will find our recommendation for one particular filter: a 0.2 micron ceramic element made by Korea Ceramics. It is the element every contaminant figure on this site was produced with, it has a history with Shimanishi going back decades, and we sell it.
What we have not done is compare filter types, media and combinations against one another, so we cannot tell you which system is best. What comes first is an overview: how filters actually work, what each kind can and cannot do, and what the certification numbers mean. Read it and you can judge our reasoning for yourself. Our own gaps, and our read on the other devices sold to change your water, follow the recommendation. If you would rather not read all that and simply want the filter, skip straight to it.
Drink more for heat, exercise, or altitude. If a clinician has advised you to limit your fluids, follow their guidance. This is general hydration guidance, not medical advice.
These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.
Why there is a filter at all
Binding is not removing.
The mineral chemistry does one job: it takes contaminants that were dissolved and invisible and turns them into something with mass. Metals, organics and fine particles lose the surface charge holding them apart, clump together into visible floc, and gravity takes them to the bottom.
At that point the contaminants are no longer dissolved in your water. They are sitting underneath it. But they are still in the container, and the last step is separating the two.
You can do that two ways. Decant — let it settle fully and pour off the clear water, leaving the sediment behind. Or pass it through a filter, which is faster, more thorough, and what most people will actually do.
This is worth being precise about, because it is the opposite of how reverse osmosis works. RO uses pressure to force water through a membrane fine enough to exclude nearly everything, including the minerals. Here the separation is mechanical, and it happens after the chemistry has already done the work. The purification happens in the water; the filter then catches what should be removed, allowing the mineral chemistry to remain.
First principles
There are only four things a gravity filter can do.
Every gravity filtration product uses one of these mechanisms, or a stack of several. Once you can name the mechanism, the marketing gets much easier to read.
Mechanism one
Size exclusion
A physical sieve. Pores of a given diameter let water through and hold back anything larger. Simple, predictable, and completely indifferent to chemistry — it cannot remove anything dissolved, because dissolved substances are molecules, not particles.
Ceramic elements, hollow-fiber and pleated filters
Mechanism two
Adsorption
Molecules stick to an enormous internal surface area. A single gram of activated carbon can carry hundreds of square meters of it. This is the only common mechanism that captures dissolved organic compounds — and it has a finite capacity, so it stops working when the surface fills.
Activated carbon, granular and block
Mechanism three
Ion exchange
A charged medium swaps ions it holds for ions in the water. Useful for specific charged contaminants, and selective — a medium tuned for one ion may ignore another entirely. It also, by definition, puts something into your water as it takes something out.
Zeolites, softening and specialty resins
Mechanism four
Redox
Certain metal alloys change a contaminant chemically instead of trapping it: free chlorine is reduced to harmless chloride, and dissolved heavy metals plate out onto the medium’s surface. Because it converts rather than captures, the medium does not fill up the way carbon does.
KDF copper-zinc granules, catalytic carbon
What each medium actually does
No single filter does everything.
This is the reason gravity systems stack media in series, and the reason a claim about “a filter” means almost nothing until you know which kind.
| Medium | Mechanism | Handles | Does not handle |
|---|---|---|---|
| Ceramic | Size exclusion | Sediment and turbidity; bacteria and protozoan cysts at a 0.2–0.5 µm rating; particulates including microplastics, which are 1 µm and larger by definition | Anything dissolved — chlorine, disinfection byproducts, PFAS, nitrate, fluoride, dissolved metals. Viruses, which are smaller than the pores |
| Activated carbon | Adsorption | Chlorine, taste and odor, VOCs, many pesticides and herbicides, disinfection byproducts, some pharmaceuticals; long-chain PFAS such as PFOA and PFOS | Short-chain PFAS, which adsorb poorly. Nitrate, fluoride, most dissolved salts. Bacteria — carbon beds can in fact host them |
| Zeolite | Ion exchange | Ammonia and ammonium; certain cationic heavy metals; some hardness | Neutral organic molecules, most pesticides, PFAS. Highly variable by zeolite type and by what else is competing in the water |
| KDF(copper-zinc) | Redox | Free chlorine, converted to chloride rather than adsorbed; dissolved heavy metals such as lead and mercury, plated onto the medium; some bacterial control within the bed | Chloramine, which it handles poorly. Most organics, pesticides and PFAS. Nitrate, fluoride, dissolved salts |
A note on ratings. A micron rating describes pore size, not performance against chemistry. A 0.2 µm element will stop a bacterium and pass a dissolved pesticide molecule without noticing it.
A note on the numbers you may have already seen
The 245-plus contaminant reductions published on our landing page were measured with a 0.2 micron ceramic gravity filter and 24 hours of contact time. No carbon stage, no ion exchange, no redox medium — the mineral chemistry and a ceramic element, nothing else.
Three implications follow from that. First, the results are not achieved by the filter alone: a ceramic element removes nothing dissolved, so whatever left the water was taken out as a result of the flocculating chemistry. Second, the reductions are a floor rather than a ceiling — time-series testing shows reductions still climbing at 48 hours. Third, adding a carbon filter in series would bring a second mechanism to bear on the dissolved organics the ceramic cannot touch — and the housings we list further down include a carbon stage as well as a ceramic dome, all of them compatible with the Korea Ceramics element we sell.
We have not tested what extra purity that combination produces, so we are not going to put a number on it — though on the mechanisms above it is likely higher than the ceramic gravity filter alone. Which is probably why a good many of our customers already run a carbon stage below the ceramic.
Drink more for heat, exercise, or altitude. If a clinician has advised you to limit your fluids, follow their guidance. This is general hydration guidance, not medical advice.
These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.
A note on ratings. A micron rating describes pore size, not performance against chemistry. A 0.2 µm element will stop a bacterium and pass a dissolved pesticide molecule without noticing it.
How to read a certification
What the NSF/ANSI numbers mean.
The standards are written jointly by NSF International (NSF, formerly the National Sanitation Foundation) and the American National Standards Institute (ANSI). They are the closest thing this industry has to an accurate common language. Each one covers a defined set of contaminants, and a product certified to one says nothing about the others.
NSF/ANSI 42
Aesthetic effects — chlorine taste and odor, chloramine, particulates. A filter can carry this certification while removing nothing that affects health.
NSF/ANSI 53
Health effects. Lead, mercury, arsenic, Giardia and Cryptosporidium, herbicides, pesticides, benzene, PCBs, MTBE, trihalomethanes. Which of these a given product is certified for varies — check the specific list, not the number.
NSF/ANSI 58
Reverse osmosis systems. Nitrate, nitrite, fluoride, perchlorate, dissolved metals — the contaminants other mechanisms tend to miss.
NSF/ANSI 401
Emerging contaminants — prescription drugs, flame retardants, detergents, newer pesticides. The standard that covers what regulation has not caught up to.
NSF/ANSI 244 & P231
Microbiological reduction — bacteria, viruses and cysts. This is the standard that matters for a ceramic element.
NSF P473
PFOA and PFOS specifically. If a product claims PFAS reduction without this or equivalent third-party data, the claim is unverified.
The one we stand behind
A 0.2 micron ceramic element, and why this one.
Everything above is general. This is specific, and it is the only filter recommendation on this site.
Ceramic domes look alike. They are not alike. Effective pore size is the product of a long chain of manufacturing decisions: the ceramic body itself, particle-size distribution, moulding pressure, wall thickness, how the element is dried, the temperature it is fired at, and how tightly all of that is held from one batch to the next. Get any of it wrong and the number printed on the box stops describing the object in your hands.
The failure modes are physical and unglamorous. A microscopic crack, an irregular pore, or a poor seal where the element meets its mounting stem lets water travel around the barrier rather than through it. Push density the other way and flow slows to something nobody will live with. Two elements can carry the same 0.2 micron rating and behave nothing alike, which is why we name a particular element rather than a category.
Every contaminant number we publish was produced with a 0.2 micron ceramic gravity element made by Korea Ceramics. Not a class of filter. That element. When we say 147 analytes were driven below detection, this is the thing the water passed through on its way to the laboratory.
Shimanishi Kaken, Matt Bakos and Dr. Kory each have their own history with Korea Ceramics. Bakos has imported Themarox into North America for twenty years and bought their elements directly for much of it. Dr. Kory’s history is the shortest and the plainest: their element has run in his own house for the past year and a half, across a good number of them, and he trusts it enough to put his name on this page.
Korea Ceramics was later acquired by Outriger, and the element now ships almost exclusively within Korea. We found one importer bringing it into the United States, bought their entire current inventory, and we sell it here. You should factor that in. We are not telling you it outperforms other ceramic elements — we have not tested other ceramic elements. We are telling you it is the one our results were produced with, the one with the older connection to this chemistry, and the one we run at home.
What this recommendation is not. It is not a finding about anyone else’s filter. We have not compared ceramic elements against one another, and we have not tested filter combinations. This element is recommended on provenance, history and daily use — three things we can account for — and not on a comparison, because we have not run one. When we run one, it will be published here whichever way it goes.
What we don’t know yet
The questions we haven’t answered, stated openly.
These are the gaps in our own data. We would rather you read them here than discover them later.
PFAS
We have not tested Aurmina against PFAS and we make no claim about it. Third-party testing of this mineral chemistry did not show meaningful performance, and we would rather say so. PFAS removal is an adsorption problem, which means it belongs to activated carbon — and EPA reports carbon works well on long-chain compounds such as PFOA and PFOS, and less well on short-chain ones. A ceramic element alone, however fine, will not address PFAS at all.
Microplastics
Not yet tested. We have tried, and the analytical logistics have so far defeated us. Microplastics are found not only in bottled water but throughout the wider water environment: rivers, reservoirs, groundwater, treated municipal water and household tap water. Studies generally report higher particle counts in bottled water than in tap water, though results vary enormously because laboratories use different collection methods, detection limits and definitions. In bottled water some particles may originate in the source water, while others may be introduced by bottling equipment, filtration membranes, the bottle itself, or repeated opening and closing of the cap.
Using the common convention, microplastics range from roughly 1 µm to 5 mm; particles smaller than 1 µm are generally called nanoplastics. A 0.2 µm ceramic element is therefore, in principle, fine enough to retain conventional microplastics by size. That mechanical expectation is not the same as a measured or certified removal claim, and it does not fully address nanoplastics — some of which are small enough to approach or pass a 0.2 µm barrier.
A 2024 study in PNAS, using a new stimulated Raman scattering imaging method, examined three brands of bottled water and estimated an average of roughly 240,000 plastic particles per litre, about ninety percent of them nanoplastics. That finding also illustrates why older studies, which could not reliably detect the smallest particles, may have substantially underestimated total counts. It should not be read as a universal concentration for every bottled or tap supply.
The health significance remains unsettled. Smaller particles are of particular scientific interest because their size may make biological uptake more plausible, but current evidence does not establish the human-health risk posed by the concentrations ordinarily encountered in drinking water. Until we have tested the complete system using an appropriate, validated method, we are not going to claim a microplastic or nanoplastic reduction we have not measured.
Which filter is best
We use gravity systems ourselves and we like them: no electricity, no plumbing, no water sent to the drain, and a replacement schedule measured in years rather than months. We do stand behind one ceramic element, for the reasons set out above — but liking and using something is not the same as having tested it against the alternatives, and we have not run that comparison. Nobody should read our recommendation as a finding about other people’s filters. When we run it, it will be published with the data behind it.
What filtration adds to the cost
Our cost comparisons have counted the bottle and not the hardware, and that is not a like-for-like comparison against a reverse osmosis system whose installation we do count. A gravity housing runs roughly two hundred and fifty to three hundred dollars and lasts years; a ceramic element is on the order of ten dollars and lasts one to two. Carbon-bearing elements are the shorter-lived component — a stage rated for three hundred gallons is replaced two to three times a year by a household of two. We are working the full arithmetic and will publish it on the economics page rather than leave it implied.
Other approaches
The rest of the water aisle.
Beyond filtration there is a whole aisle of machines and additives that promise to alter and improve drinking water. We sell mineral chemistry, which you should factor in. We have also researched water chemistry and water physics deeply enough to have a view. Some of these have no support at all. One or two have real science behind them, and real limits. Our take:
Magnetized watermixed evidence
There is more literature here than skeptics allow. Magnetic treatment of irrigation and livestock water has been studied for decades, and reviews report gains in yield, biomass and nutrient uptake across a range of crops — with the largest effects on brackish, saline or otherwise poor-quality water, which is the interesting part. The caveats are equally real: the mechanism is not established, results vary between studies, and these reviews are narrative rather than systematic, so null results are underrepresented. What we have not seen is evidence that a magnetic field durably changes drinking water in a way any standard method can measure. Promising in a field. Unsupported on a countertop.
Far-infrared, “energized” and ionizing devices
Two different things get sold under this heading, and they deserve different answers.
The passive add-ons — far-infrared and tourmaline balls, mineral cartridges sold as “energizers,” and the scalar, quantum and frequency-imprinted plates, discs and stickers — are frequently bundled onto otherwise reasonable gravity systems. The filter media in those systems may be perfectly good. The attachment is sold on a story rather than a result, and we have found nothing measurable by a standard method behind any of it.
Electrolysis ionizers are a real device, and we are not going to pretend otherwise. Passing a current through water genuinely splits the stream, raises the pH on one side and drives the oxidation-reduction potential negative. Those are measurable changes. The question is what they are worth — and the most authoritative answer comes from researchers who are friendly to this chemistry. A 2022 review of electrolyzed-reduced water concluded that dissolved molecular hydrogen is “the exclusive agent responsible” for the effects, that the alkaline property “will not meaningfully influence the pH of the body,” and that the negative ORP reading is fully explained by dissolved H₂ and pH, so that “no additional enigmatic explanations such as ‘free electrons’ or ‘stored energy’ are required.” So an ionizer is best understood as a hydrogen-water machine, and its other two claims are answered in the cards below.
“Structured” or “coherent” water
Coherence is not a defined property in water chemistry, and there is no accepted method for measuring it. We used this language ourselves in early materials and removed it, because we would rather name a mechanism than a mood. We are not going to lump every product here together, though: Analemma, the best known of them, publishes its own research including a microbiome study, and we have seen no independent replication in either direction. What we can say is that it treats water a vessel at a time, which is not a scale anyone runs a household on. Whether it adds anything on top of mineral chemistry, we have no idea.
Vortexed water
There is a long tradition here, going back to Viktor Schauberger, and the people in it have spent decades watching water move and are genuinely attached to what they see. We are not going to wave that away. What we can tell you is what we found when we looked: nothing published in the independent literature on vortexing drinking water. It is also worth knowing that several products sold as vortex devices contain magnets as well, which makes it hard to attribute any reported result to the vortex itself. This is an approach to water physics rather than water chemistry, which is a different question from the one we work on. We are intrigued, and we would like to see the data.
Alkaline water
The central claim — that drinking high-pH water alkalizes the body — does not survive contact with the stomach, which restores its own acidity by secreting more acid, or with the blood, which is held between pH 7.35 and 7.45 whatever you drink. A 2026 clinical review concluded that claims of systemic alkalinization, enhanced immunity and disease prevention are not supported by robust evidence, and that any modest effects are “likely driven more by mineral composition than pH alone.” Which is the part worth keeping: what is dissolved in water matters. Its pH number, by itself, does not.
Hydrogen waterreal mechanism, real limits
This one is not in the same category, and we are not going to pretend otherwise. Dissolved molecular hydrogen has a plausible mechanism as a selective antioxidant, and there is a genuine literature behind it: a 2023 review counted 81 clinical trials. But that literature is early and inconsistent — a 178-patient Parkinson’s trial found no benefit after smaller studies had. And the physical limits are real. Hydrogen’s solubility in water is about 1.57 mg/L at ordinary pressure and temperature, so a therapeutic dose means drinking litres a day. It is also the smallest molecule there is: it begins leaving the moment a container is opened, and diffuses through most plastics, so the concentration in an open flask falls from the moment it is opened. Interesting chemistry. Not a durable property of your drinking water, and not something that scales.
Unqualified micron claims
“Filters to 0.2 microns” describes a pore, not a result. It tells you what the filter physically blocks and nothing at all about dissolved chemistry, which is where most of the contaminants people worry about actually live.
The test we try to apply to ourselves is simple. Can independent laboratories measure the intervention’s effects using validated methods? If not, it does not yet belong on a label — including ours.
We will keep this page current.
When the testing is done, the results land here — including the ones that do not flatter us.