The reagent nobody lists: water grades, and how purity is made, measured and lost

Ask someone to list the reagents on a bench and they will name the bottles with labels. Almost nobody names the one that outweighs all the others put together. Water rinses the glassware, dilutes the standards, fills the baths, feeds the instruments and makes up most of nearly every solution in the room. By volume it is the main ingredient of laboratory work, and it is the one most often treated as if it had no properties at all.

It has plenty. Water is an unusually good solvent, which is the reason it is useful and also the reason it is never simply water. It carries whatever it last touched: salts from the ground, organic matter from a reservoir, chlorine from the treatment works, gas from the air above it, traces of the pipe, the tank and the bottle. Purifying water means removing those passengers. Keeping it pure means stopping it from picking up new ones, which turns out to be the harder half of the job.

This guide walks through the subject in the order a careful reader would meet it: what is in ordinary water, how the published grades are defined and what the numbers mean, how each purification step works and what it leaves behind, how purity is measured and where the measurement is blind, why pure water starts to degrade the moment it is made, how to store it, how to match a grade to a task, and how to read the paperwork on a bottle someone else filled.

1. What is actually in tap water

Drinking water is treated to be safe to drink, which is a different target from being chemically empty. Everything in it falls into one of five groups, and the distinction matters because no single technology removes all five.

The amounts vary enormously from one town to the next and from one season to the next. A water system that performed well in a soft water region can be overwhelmed in a hard water one. That is why the first stage of any purification train is designed around the local feed rather than around the result that is wanted at the end.

2. The published grades, and what the numbers mean

"Pure water" is not a specification. Two standards are quoted most often for reagent water. The first, ASTM D1193, describes four types. The second, ISO 3696, describes three grades. They do not map onto each other exactly, and anyone quoting a grade should say which standard they mean.

Values as commonly cited from ASTM D1193. The standard also sets limits for sodium and chloride and, for Type IV, a pH range. Always work from the current edition of a standard, not from a summary table, including this one.

Two things in that table surprise people. The first is that Type III carries a tighter conductivity limit than Type II. The types were written around how the water is prepared, with Type II historically associated with distillation, so they are not a simple ladder on every property. The second is how small the top number is. A resistivity of about 18.2 megohm centimeters at 25 °C is close to the theoretical limit for water, the point where the only ions left are the few that water makes by splitting itself. A purification system cannot beat it, and a display reading higher is a display with a temperature or calibration problem.

ISO 3696 takes a similar approach with three grades. Grade 1 is the most demanding and is intended for the most sensitive analytical work, Grade 2 for general analytical use, and Grade 3 for most ordinary wet chemistry and for rinsing. Its conductivity limits run from 0.1 microsiemens per cm for Grade 1 through 1.0 for Grade 2 to 5.0 for Grade 3, all at 25 °C.

A grade describes water at the moment and place it was measured. It is not a property the water carries with it into a bottle, down a tube or across a room.

3. How it is made: six stages, each with one job

No single technology takes tap water to Type I. A purification system is a train of stages, each removing one kind of passenger and protecting the stage after it.

Pretreatment. A depth filter takes out particles, and activated carbon takes out chlorine and a share of the organics. Chlorine is put into drinking water on purpose and it destroys the thin membranes used in the next stage, so this step is there to protect the equipment as much as the water.

Reverse osmosis. Water is pushed under pressure through a membrane that lets water through far more easily than what is dissolved in it. A good membrane rejects roughly 95 to 99 percent of ions and almost all particles, organisms and large organic molecules. It is the workhorse of the train. What it cannot do is finish the job: a few percent of a hard feed water is still a great deal, and dissolved gases pass straight through.

Storage. Reverse osmosis is slow, so its output is collected in a tank. This is the weakest point in most systems. Still water at room temperature in contact with air is where growth begins, so good tanks are opaque, have a filtered vent, a smooth conical base that drains completely, and often a recirculation loop that keeps the water moving past a lamp.

Ion exchange. Beads of resin swap the remaining positive ions for hydrogen and the remaining negative ions for hydroxide, which combine to form water. A mixed bed of both resins takes resistivity from the low megohm range to the theoretical limit. The resin has a fixed capacity. When it is spent, the weakly held ions come off first, and silica is among the earliest to escape. Electrodeionization does the same work with an electric field that regenerates the resin continuously.

Ultraviolet light. Two wavelengths do two jobs. Light at 254 nanometers damages the genetic material of microorganisms and stops them multiplying. Light at 185 nanometers has enough energy to break organic molecules into charged fragments that a following resin bed can capture, which is how total organic carbon is brought down to single figures.

Final filtration. A membrane with pores around 0.2 micrometers at the outlet holds back any organisms and particles shed by the system itself. Where the by-products of bacteria matter, an ultrafilter with far smaller pores is used instead or as well.

Distillation, the oldest method, still has a place. Boiling and condensing removes ions, particles and organisms in one step and needs no consumables beyond energy. Its limits are that volatile compounds travel over with the steam, the output is slow, and a still that is not cleaned concentrates scale in its boiler.

4. Measuring purity: what resistivity sees

Pure water conducts electricity very poorly. Dissolved ions carry current, so the more ions, the higher the conductivity. Resistivity is simply the same measurement turned upside down, and it is preferred at the high purity end because the numbers are easier to read: 18.2 megohm centimeters is friendlier than 0.055 microsiemens per cm.

Three things about the measurement are worth knowing before trusting a display.

• It depends heavily on temperature. The conductivity of pure water changes by several percent for every degree. Meters correct the reading to 25 °C, and the correction is a calculation, not a measurement. A report should say that the value is temperature compensated.
• It is measured in the line, not in a beaker. Water at 18.2 megohm centimeters poured into an open container falls to around 1 within minutes as it takes up carbon dioxide from the air. A handheld probe in a cup can never confirm Type I. Only a cell inside the closed system can.
• It is blind to anything without a charge. Most organic molecules, particles, microorganisms and dissolved gases other than carbon dioxide carry no current. Water can read a perfect 18.2 and be heavily contaminated with all four.

That third point is the one that catches people. Resistivity is an excellent measurement of ions and no measurement at all of anything else. A water specification that quotes only resistivity has described one of the five groups in the first table.

5. The measurements that cover the rest

Total organic carbon. A total organic carbon analyzer oxidizes the organic matter in a sample and measures the carbon dioxide that results. The answer is a single number, usually in micrograms of carbon per liter, often written as parts per billion. It says how much organic carbon is present, not what the compounds are. For Type I water the limit is 50, and modern systems routinely deliver below 5. It matters most for chromatography and for any work where an unidentified peak costs a day.

Microbial count. A measured volume is passed through a membrane, the membrane is placed on a growth medium, and the colonies that appear over several days are counted. The result is given as colony forming units per volume. It is slow, and it only counts organisms that will grow under the chosen conditions, so it is a trend indicator more than an absolute figure.

Bacterial by-products. When certain bacteria die and break apart they release fragments of their outer wall, known as endotoxins. These are not alive, pass through a 0.2 micrometer filter, and survive boiling. They are measured with a dedicated assay and reported in endotoxin units per volume. Removing them takes ultrafiltration or a charged filter.

Particles and silica. Particle counters report the number above a stated size per volume. Silica is measured separately because it is weakly ionized, contributes little to conductivity, and is one of the first things a tired resin bed lets through.

6. Purity does not keep

The purer water is, the more aggressively it takes things up. A solvent with nothing dissolved in it has, in a sense, room for everything. Three processes begin as soon as the water leaves the final filter.

• It absorbs carbon dioxide. Air contains carbon dioxide, which dissolves to form a weak acid. Within minutes the pH of exposed pure water drifts to roughly 5.6 and resistivity falls to about 1 megohm centimeter. Nothing has gone wrong. This is simply what water in equilibrium with air looks like, and it is why a pH reading on high purity water is close to meaningless.
• It leaches its container. Glass gives up sodium, silica and boron. Many plastics give up plasticizers, mold release agents and unreacted monomers. The amounts are tiny, and so are the limits being protected.
• Things grow in it. Organisms adapted to nutrient poor water can live on the trace organics in the water and the surfaces around it. They attach to walls and build films that shed cells and by-products for as long as they are left alone.

The practical consequence is simple and widely ignored. Type I water is made, not kept. It should be drawn at the moment of use. A carboy filled on Monday from a Type I outlet does not hold Type I water on Tuesday, whatever the label on the carboy says.

7. Storing the water that can be stored

Lower grades tolerate storage if a few rules are followed. Each rule answers one of the three processes above.

• Keep it closed. A sealed container with little headspace limits carbon dioxide uptake and airborne contamination. A vent that must be open should carry a filter.
• Keep it dark and cool. Light supports algae, and warmth speeds every kind of growth.
• Keep it short. Date every container when it is filled and agree a maximum holding time. Days, not months.
• Draw from the bottom, never dip into the top. A tap at the base means nothing enters the container to take water out. A pipette or a beaker lowered through the neck contaminates everything left behind.
• Empty, clean and dry on a schedule. Topping up a half used container carries the old film into the new water.

8. Matching the grade to the task

Using the highest grade for everything feels safe and is wasteful. Type I water is expensive to produce, wears out cartridges, and is so aggressive that it is a poor choice for filling a water bath or feeding some equipment. The sensible approach is to ask what would interfere with the task and choose the grade that controls that one thing.

9. Looking after a purification system

A water system is a set of consumables arranged in a line, and its output is only as good as the most neglected one. The display on the front reports resistivity at one point. It says nothing about the age of the lamp, the state of the tank or the film inside the dispensing tube.

• Change consumables on the calendar, not on the alarm. A resin cartridge releases silica and organics before resistivity visibly drops. A lamp loses output long before it goes dark.
• Flush before drawing. Water that has sat in the dispensing line overnight is not the water the cell measured. Run the first portion to drain.
• Sanitize the tank and the loop on a schedule. Films do not clear themselves, and a new cartridge downstream of a fouled tank has a short life.
• Keep the outlet clean and off the bench. A dispensing tip resting in a puddle draws that puddle back into the line.
• Write it down. Date, readings, what was changed and by whom. When a result is questioned months later, the log is the only evidence of what the water was on that day.

10. Water in a bottle: reading someone else's paperwork

Not every bench has a purification system, and packaged water is a reasonable alternative for modest volumes. The difficulty is that the buyer cannot see how it was made, so the label and the accompanying document have to carry the whole description. A complete one answers seven questions.

1. Which standard and which grade? A named standard with its edition, not an adjective such as "ultra" or "high purity".
2. Which properties were measured? Resistivity or conductivity alone describes ions only. Look for total organic carbon and, where it matters, a microbial or by-product result.
3. When and where was each measurement taken? At production, in the line, before filling. A reading taken at filling is a statement about that moment.
4. Is it a result for this lot, or a typical value? "Typical" means somebody measured some other lot once.
5. What is the lot number, and does it match the bottle? A document that cannot be tied to the container in the hand describes a different container.
6. What is the container made of, and how was it closed? The material decides what will leach. The closure decides whether the first opening can be recognized.
7. What dates are given? A fill date, a date by which it should be opened, and guidance on how long it remains suitable once opened.

Once a bottle is open, everything in section 6 applies to it. Record the opening date on the label, pour from the bottle instead of reaching into it, never return unused water, and treat the stated grade as a description of the unopened container.

11. Eight common mistakes

1. Quoting resistivity as if it were purity. It is the ionic part of purity and nothing more.
2. Checking Type I water with a handheld probe in a beaker. The reading will be low, and it will be the air that is being measured.
3. Storing Type I water. It has stopped being Type I by the time the cap is on.
4. Measuring the pH of high purity water and adjusting something because of the answer.
5. Using a wash bottle for weeks. A squeeze bottle is a small, warm, frequently opened storage tank.
6. Running a system until the alarm sounds. By then the silica and organics have been in the output for some time.
7. Connecting an outlet with whatever tubing is to hand. Soft tubing can add more organics than the whole train removed.
8. Using the best water for everything. It costs more, and in a bath or a boiler it does harm.

12. A ten-point water check

1. Is the required grade written down for each task, with the standard it refers to?
2. Is the property that matters for that task actually being measured, or only resistivity?
3. Is the resistivity reading taken in the line and temperature compensated?
4. Is Type I water drawn at the time of use and never held?
5. Is the first portion from the outlet flushed to drain?
6. Does every storage container carry a fill date and an agreed maximum holding time?
7. Are containers closed, kept out of the light and drawn from a tap, not dipped into?
8. Are cartridges, lamps and filters changed on a schedule, with the dates recorded?
9. Is the tank and distribution loop sanitized at a set interval?
10. For packaged water, does the document name the lot on the bottle and state when each value was measured?

None of this is difficult, and very little of it needs equipment beyond what a purification system already has. It needs the habit of treating water as a reagent with a grade, an age and a history, which is what it has been all along.