On Outdoors StackExchange on a post about the safety of distilled water, someone claimed that pure water is very corrosive.

Water distilled for some laboratory uses (like the Z-pinch reactor) is distilled to 0.00000009% purity, and is caustic enough to burn through most things.

Water, purified enough, cannot be distributed through stainless steel pipes. Instead, they use glass pipes to transport it within the plants where it's made.

For sure, "caustic" is not the correct word, since water is neutral.

But, is it that corrosive? Or is it even less than water with salt?

  • 13
    $\begingroup$ That is entirely false. Water is certainly not caustic; that term is mostly reserved for strongly basic substances that will readily attack organic matter (e.g., human flesh). I've been drinking water (including, gasp, distilled water) for my entire life without my internal organs dissolving. As for its corrosiveness, while water will certainly react with various poor/common metals, it will not react easily with stainless steel, which has a passivation layer of chromium oxide that makes it highly resistant to corrosion. Furthermore, water of higher salinity is more corrosive, not less. $\endgroup$
    – Greg E.
    Dec 22, 2013 at 2:21
  • 8
    $\begingroup$ It's been claimed (in a comment on the linked site) that the ultrapure water is too corrosive for stainless steel pipes and has to be transported in glass tubing instead. It seems to me that the more obvious explanation is that at ultrapure levels, they don't want the water picking up contaminants from metal piping. $\endgroup$
    – chipbuster
    Dec 22, 2013 at 4:23
  • 1
    $\begingroup$ Here is another variant of the "highly corrosive" rumour: symmetrymagazine.org/article/july-2015/underground-plans : In this article ultrapure water dissolves a hammer. $\endgroup$
    – jochen
    Jul 22, 2015 at 19:24
  • 1
    $\begingroup$ Water is extremely corrosive to potassium... $\endgroup$
    – vapid
    Jul 5, 2016 at 15:44
  • 1
    $\begingroup$ Once ultra-pure water starts reacting to yield a solute, it is not ultra-pure anymore, so it can't be that bad. If you try to dissolve a highly insoluble chloride, and your water contains traces of chloride, the reaction will not go forward. However, if you have a continuous flow of ultra-pure water (or just water not containing any chloride), traces of the insoluble chloride will dissolve. Not that that would be dangerous. $\endgroup$
    – Karsten
    Mar 8, 2019 at 4:58

7 Answers 7


That quote had false statements all over it.

Type I ultra pure water (Milli-Q water, or others) are fairly common among all labs. They are not corrosive to stainless steel.

As @chipbuster said, purified water is never distributed through stainless steel partly because concern of contamination. Metal reacts not only with oxygen, but also many pollutants in the air. Overtime on metal surface will build up a layer that has all kind of compounds in them. When water flows through, ions get picked up.

However, the most major reason would be that glass is better in all respect. The reason glasswares are commonly used in the lab have little to do with pure water. Glasswares are relatively stable across wide range of pH and very hard to be oxidized or reduced. Plus, glasswares are cheap, extremely easy to manufacture, and transparent, so you can see through. You will really only need metal if extreme pressure or temperature is required.

As a matter of fact, it is also unwise to use glassware to distribute ultrapure water--you will get ion contaminations from glass as well. Purified water are actually almost always stored in plastic containers and distributed through plastic tubes.


Water is corrosive at very high temperatures and pressures (in the sense that it will react with a lot of common materials), but generally speaking, pure water is less corrosive than water with impurities. Some power plants use purified water in their cooling systems for this reason.


If we define corrosivity as impurification of the water, ultrapure water eats nearly everything, including Pyrex. Plumbing is usually polyvinylidene fluoride (PVDF, Kynar), polyperfluoroalkoxy (PFA), and ethylene chlorotrifluoroethylene (ECTFE, Halar).

Stuff still grows in ultrapure water, slowly. Chip fabs' water systems must be hard-UV irradiated, ozonized, or otherwise disinfected plus ultrafiltered against bio-particulate generation.

The product tastes awful - bitter and flat.

  • 3
    $\begingroup$ Looking at corrosion from the point of view of the solvent is misleading here. The amounts of dissolved substances are very very small and insignificant with respect to the piping or the thing being corroded. So your answer is just wrong. $\endgroup$
    – matt_black
    Mar 27, 2014 at 22:15
  • $\begingroup$ The water is ruined - and the devices it washes - because "very small" is NOT "insignificant." The SR-71's original titanium alloy was stress corroded by chloride. Read Ben Rich's book "Skunk Works" on what a marking pen did to parts, and summertime chlorination of tap water. $\endgroup$
    – Uncle Al
    Mar 28, 2014 at 1:15

The quote in the original post has nothing to do with reality. Ultrapure water will not dissolve metals like acids. Ultrapure water can be identified by its resistance of 18.2 Megaohms and it is free from organics impurities.

Now the punch line is that ultrapure water is "contaminated" very quickly. This is all at what we analytical chemists call "ultratrace level" contaminants. Merck did a study (not published, I saw that in a conference) that it just takes 30 minutes in a lab for the resistance to decrease 10-12 Megaohms. The most common contaminant is surprisingly NH4(+) ion in a typical laboratory atmosphere along with chloride from HCl traces. CO2 lowers its pH to around ~ 6. Most likely UPW will have a very unpleasant or bland taste.

Yes supercritical water is extremely corrosive. It dissolves silica, attacks most metals, and so on.


Remember your equation for free energy

There is a germ of truth to this claim. The thermodynamics should be familiar, and is already covered in another Chemistry.SE post, "Reaction quotient and Gibbs free energy at the start of a reaction." The gist is that dissolving something in theoretically (hypothetically) pure water has a fabulous entropy advantage. If it's in a container with a mole of surface ions exposed, you have Avogadro's number of ions trying to get into it from the walls of its container, and then that one ion has to go back for the reverse reaction to occur. Equilibrium will favor the dissolution of substances even when the standard free energy is highly positive (endergonic).

The catch, of course, is that infinitely pure water doesn't stay perfectly pure for long! The glass pipes are to prevent the pipes from ruining the water, not the water ruining the pipes because it only takes a minuscule amount of solute to reduce the free energy difference substantially. Additionally, the reaction has to occur at a measurable rate by kinetics, so any compound not normally at risk of dissolving in water at all (such as well crosslinked polymers) won't dissolve no matter what the delta G is.

In the biological context this started with, we can be sure that the water is absorbing some substances from the body. For example, a glucose transporter uses free energy from ATP hydrolysis to establish a very strong concentration gradient across the cell membrane - nonetheless, some net molecules of glucose are sure to move out of the cell when absolutely none are available to enter. But this is true whenever you drink water completely lacking in glucose, regardless of whether it contains electrolytes or not.


This "question" was just trashed out several days ago, I am flabbergasted that it even got one upvote let alone 19.

Pure Water has no dissolved ions except equal amounts of H3O+ and OH-from self ionization, no dissolved gases, no organics, no VOCs, and no bacteria or viruses. Can this be attained? No! Can it be approached? Yes! Talk to Millipore, Barnstead, Culligan, and other companies in the super-pure water business for details and specs.

What can [pure]Water do: It can dissolve anything that is soluble in water, Ionic materials, gases, polar organics, and even nonpolar organics, to their limits of solubility. How is this controlled? by limiting contact with anything and constant recycling thru the purification system. Water will dissolve ions from any metal, ceramic, borosilicate glass [I know I worked for Bausch and Lomb Soflens; isotonic saline in borosilicate vials ruined lenses by leaching ions from vials], organics from any plastic, gases from any atmosphere, even ions from teeth. The only substance that I think might resist Water is fused silica. This is what water does, the universal solvent. To prevent this the solution is to limit contact or the common ion effect [the latter means the water is no longer pure].

Corrosiveness of water depends on an anode and anode reaction, a cathode and cathode reaction, an electrical connection between the anode and cathode, and ion transport between the anode and cathode. Remove any one of these and corrosion stops whether it is galvanic or electrolytic. Removing ions from water reduces the ion transport to the limit of the 18.2 Mohms so Water will never be completely noncorrosive or nonconductive, but the purer it is the less corrosive. Also, water reduces the possible anode and cathode reactions.

The idea that Water causes greater single electrode potentials seems specious. The charge separation on the metal will be inhibited by just that and to say that it is lowered by ions in solution is true but circular to the discussion because if they come from the metal the concentrations will be lower. Charge separation in a conductor is unlikely to be positive in entropy change.

Water is not biologically dangerous unless it contains contaminants, poison, etc. Drinking too much water will cause problems and relying on drinking to supply necessary minerals is doubtful. Reverse osmosis water and distilled water, to my taste at least, have a crisp, clean taste. In general, the cleaner water is the better to drink. The only caveat here in RO or DI water is that fluoride ion is removed or at least lowered and that is usually added and not natural. Fluoride in water has made a huge reduction in dental caries. [Again personal experience I [BF] have cavities in almost all my teeth, My children [AF]have only 2 or 3 between them.

Bacteria and virus contamination is nonexistent in good reverse osmosis water and is removed by ultrafiltration and UV sterilization in ultrapure Water systems. Pure Water will not support bacteria unless enough is added to feed those not killed and nutrients are added.

Water will leach monomers, oligomers, plasticizers, and initiators from plastics. Since it is a small molecule Water will diffuse thru some plastics. Pure water is probably better at these feats than less pure water. It is better at osmosis.


As stated by others, the original post if full of falsehoods. There are also falsehoods in the replies. I have a lot of experience with ultrapure water, my company makes very large ultrapure and bench top labwater systems.

1) UPW (ultrapure water) is not caustic. Conversely, in it's purest form it because acidic when exposed to atmosphere because it readily absorbs CO2, which converts to carbonic acid. However, by default, UPW can only have a pH of 7 (again when not exposed to air). So proper pH measurement can only be done inline or with proper buffers. We actually wrote a paper on this topic because researchers were always complaining that their water was acidic.

2) UPW will corrode iron, brass, poor quality steel. However, it will not corrode stainless steel as long as the welds are proper, even then hard to make SS conducive to corrosion by UPW. Example, resistivity meters that we use are made from SS (need to be to conduct electricity across 2 metal probes to get a reading).

3) UPW will not corrode Pyrex as stated above. After all, HPLC and LCMS grade water (not going to define so look it up) is placed in glass containers. Distilled water is even made in glass stills for laboratory use. Having said this, it will leach some silica and boron, so they actually should use special glass. But it won't corrode it. I had a bottle of LCMS grade water sitting in the lab for years. Also, exotic plastics don't need to be used, such as PFA, PVDF, Halar, etc. The possible exception is in a fab of a semiconductor plant. We use polypropylene, PEX (cross-linked polyethylene), and even PVC (but not ideal because in some cases the glue can leach TOC's - total organic/oxidizeable carbon that are undesireable, for example in a waste water treatment plant.).

4) the comment that "pure water is less corrosive than water with impurities", is not completely true. Put a brass or iron bolt in distilled water and see what happens compared to tap water. Also, power plants don't use UPW because tap water is more corrosive. All steam powered plants (coal, gas, nuclear, etc.) use UPW, because residuals will form on the turbine blades and throw them off balance and destroy them. Also, impurities from the water will leave residue on the evaporators in the cooling towers and render them useless. One can't compare the corrosiveness of tap water and UPW. I've seen very corrosive tap water (classic example is the city water issue in Flint, MI. The water is corroding the pipes and leach lead into the water because it contains chlorates). UPW can be corrosive to other metals because it forms carbonic acid when exposed to air. etc......

By definition, ultrapure water is defined as ~17.0-18Meg-ohm or type I water. Meaning does not contain dissolved ions to conduct electricity. In theory, if you throw your TV into a tub full of UPW it won't short out. If swimming in a lake of UPW and lightening hits the water you won't be electrocuted, because it contains no ions to conduct electricity. The theoretical limit for UPW is ~18.26 Meg-ohm.

DO NOT DRINK UPW, it will corrode metal fillings, leach calcium from your bones, eat the enamel in your teeth, even can cause your cells to explode (due to osmosis). Fortunately, before you can do to much damage you likely will feel sick to the stomach. The irony is that the purer the water is the worst it is for you to drink. Though, you can denature UPW by mixing it with Tang, powdered drink mix, tea, coffee, etc. It will make very strong tea and coffee. Even though distilled water is not considered UPW (by definition, it is a low grade Type II), it isn't good for you to drink because it has no minerals. Also tastes insipid. While on the subject, home RO or 'purified water' you buy in the store, is considered as ASTM Type IV, but in our industry we call it Type III.

Yes, UPW is one of the strongest solvents known to mankind. But that doesn't mean it will eat through everything. It just means it's always trying to reach equilibrium by leaching out ions, VOC's and CO2.

  • 1
    $\begingroup$ Through what mechanism could water that you drink "leach calcium from your bones"? Further, do you have sources for the claims that it will dissolve fillings, eat the enamel of teeth, or cause cell lysis in a human body? People drink RO and various qualities of distilled water all the time with no ill effects - only the slight loss of mineral content that tap/well water contains. A mouthful of 100% pure water at 0ppm TDS only has to mix with about one milligram of saliva in order to become 1 ppm level (RO - comparable) water that huge numbers of people have in their kitchens. $\endgroup$ Mar 1, 2016 at 1:50
  • $\begingroup$ This was a great answer until the relatively dubious claims that highly pure levels of water are somehow dangerous to the human body. I'm shocked that somebody seemingly well versed in the subject of water purity would consider that some easily consumable amount of water would have any significant affect on an organism containing perhaps 100 times that amount of water. Also, suggesting severe damage to an organism by fact that the water is missing trace amounts of minerals requires proof that the organism is solely dependent on water as their source to begin with. These things you cite are in $\endgroup$
    – user29956
    May 11, 2016 at 7:23
  • $\begingroup$ @JimmyWilliam This was really a comment, not an answer. Please use the answer field only for actual answers. With a bit more rep, you will be able to post comments. For the moment I've converted your post to a comment. $\endgroup$ May 11, 2016 at 9:10

Not the answer you're looking for? Browse other questions tagged or ask your own question.