USP 645 Conductivity and Water Spec Lookup

It is temperature dependent, which is what makes it awkward. The limit is read from the Stage 1 table against the measured, non-temperature-compensated conductivity at the measured temperature. At 25 degrees Celsius the Stage 1 limit is 1.3 microsiemens per centimeter; at 20 degrees it is 1.1, and at 80 degrees it is 2.7.

Because the Stage 1 table is built against raw conductivity at the actual temperature. Most instruments default to compensating readings back to 25 degrees, and a compensated value compared against the Stage 1 table is simply the wrong comparison. Turning compensation off for the Stage 1 test is the step most often missed.

Nothing is rejected yet. Stage 1 is a screening test, and failing it moves you to Stage 2, where a sample is equilibrated to 25 degrees with atmospheric carbon dioxide and re-measured against a fixed limit. If Stage 2 also fails you proceed to Stage 3, which adds pH to the assessment.

Stage 2 uses a fixed limit of 2.1 microsiemens per centimeter on a sample equilibrated to 25 degrees in contact with atmosphere, with stirring until the reading stabilizes. Because it is fixed rather than temperature dependent, it is a cleaner comparison than Stage 1 but takes considerably longer to perform.

Stage 3 brings pH into the assessment, comparing the measured conductivity against a pH-dependent limit after adding potassium chloride to the equilibrated sample. Reaching Stage 3 usually means the water is carrying an ionic contribution that the earlier stages could not resolve.

No. USP does not set a pH specification for Water for Injection or Purified Water. pH control is subsumed into the conductivity requirement of chapter 645. This surprises people regularly, and it is why a pH excursion alone is not a monograph failure.

Not more than 500 parts per billion, which is 0.5 milligrams per liter, under USP chapter 643. It applies to both Water for Injection and Purified Water — the two grades share the same chemical limits and differ on endotoxin and on how they are produced.

Not more than 0.25 endotoxin units per milliliter, under chapter 85. Purified Water has no endotoxin specification. This is the attribute that genuinely separates the two grades, alongside the production method.

They are recommended action levels rather than monograph specifications, and they live in the informational chapter 1231: 10 colony forming units per 100 milliliters for Water for Injection, and 100 per milliliter for Purified Water. Exceeding an action level triggers investigation rather than automatic rejection.

Three things: conductivity by the three-stage method of chapter 645, total organic carbon by chapter 643 at not more than 500 parts per billion, and bacterial endotoxins by chapter 85 at not more than 0.25 units per milliliter. Microbial action levels and pH are not in the monograph.

Inline, continuously, at a point representative of the water being released — commonly on the distribution loop return. Grab samples drift because dissolved carbon dioxide changes the reading, which is one reason the Stage 1 test is specified on a non-compensated inline reading.

Because it forms carbonic acid, which is ionic and raises conductivity. That is exactly why Stage 2 requires deliberate equilibration with atmosphere: it standardizes the carbon dioxide contribution rather than letting it vary between samples.

The limit changes with temperature, not the requirement. A loop stored and circulated hot is compared against the Stage 1 value for that temperature, which is higher than the 25 degree figure. Using the 25 degree limit on an 80 degree loop will fail water that is entirely compliant.

The instrument must be suitable for the range and calibrated, with the cell constant known and verified. The practical point is that a conductivity cell reading near the limit needs to be trustworthy at that end of its range, and calibration records form part of the qualification evidence.

The European Pharmacopoeia takes a different approach to water conductivity and the two are not interchangeable. If you are releasing to both, the governing specification is whichever is stricter for your case, and that has to be established rather than assumed.

Diagnose before dosing. Common causes are carbon dioxide ingress, resin or membrane exhaustion, a loop that has lost its sanitization regime, or an instrument reading in compensated mode against the Stage 1 table. The last of those costs nothing to check and is more common than people expect.

Both meet the same chemical limits for conductivity and total organic carbon. WFI adds a bacterial endotoxin limit of 0.25 EU per mL and, historically, a restricted set of production methods. The practical difference is the microbial control regime around the system rather than a difference in the chemistry specification itself.

Modern pharmacopoeial positions permit methods other than distillation where the process is validated and consistently produces water of equivalent quality, which opened the door to membrane-based production. It remains a validation-heavy route, and the controls around microbial and endotoxin performance carry the burden that distillation used to discharge by physics.

Because a single number cannot distinguish harmless dissolved carbon dioxide from genuine ionic contamination. Stage 1 is a non-temperature-compensated in-line measurement against a table; water passing there needs no further testing. Stages 2 and 3 progressively strip out the carbon dioxide contribution so that only real contamination remains.

Not necessarily that the water is unacceptable. It means the sample failed the simplest and most conservative test, and the analysis must continue to stage 2 and if needed stage 3. Many stage 1 failures are atmospheric carbon dioxide dissolving into the sample rather than a system problem.

Because organic carbon is a proxy for everything the purification train should have removed and for the nutrient load that supports microbial growth. The limit is deliberately generic: it does not identify the compound, it establishes that the total burden is low enough that the system is behaving as designed.

Action levels, and the distinction matters. 10 CFU per 100 mL for WFI and 100 CFU per mL for purified water are levels at which investigation and response are triggered, not pass or fail specifications in the pharmacopoeial sense. Treating them as pass/fail leads to the wrong response when a trend is developing.

Because holding the loop at 80 degrees C or above suppresses microbial growth continuously rather than relying on periodic sanitisation. The cost is thermal cycling on the system, higher energy use, and the pump suction problem that hot water creates, since vapour pressure at that temperature consumes most of the available NPSH.

Iron oxide formation on stainless surfaces, typically in hot WFI service. It is a surface phenomenon rather than general corrosion, but it discolours product-contact surfaces, can shed particulate, and indicates that the passive layer is not performing. Derouging and repassivation restore the surface.

On a risk-based schedule established during validation, with the frequency higher during initial qualification and settling into a routine programme afterwards. Sampling points are chosen to represent the worst case in the loop, including points of use, rather than only the easy locations.

It is decisive. Water leaving the generation train can meet every limit and still fail at a point of use if the loop has dead legs, insufficient velocity, poor drainability or cold spots. The specification describes the water; the loop determines whether it is still that water when it arrives.