Most laboratory water purification systems carry a resistivity or conductivity meter to monitor water purity, and most explanations of what the number means stop at "18.2 is pure." That figure is real and worth knowing, but it is the least useful part of the topic for anyone actually running a system: the questions that come up in practice are what the number means at a temperature other than 25°C, what a cell constant is correcting for, and what the compendium actually requires rather than what people assume it requires. This article covers all three.
Key takeaways
- Resistivity and conductivity are reciprocals of the same measurement. High purity means high resistivity and low conductivity. Neither is a different technique, and a meter reporting one can always be read as the other.
- 18.2 MΩ·cm at 25°C is the theoretical resistivity of water containing no ionic impurities at all. It is a ceiling, not a target range, and it is only meaningful for water measured inline, since drawn water starts absorbing atmospheric CO₂ within seconds.
- The cell constant corrects a real probe back to the theoretical 1 cm cube standard. Getting it wrong or leaving it unprogrammed produces a plausible-looking but incorrect reading.
- Every reading must be temperature-corrected to be comparable. Resistivity falls and conductivity rises as temperature increases, and the standard reference point is 25°C.
- Resistivity and conductivity say nothing about organic contamination. A separate TOC or chromatographic method is required for that, regardless of how good the conductivity reading looks.
- TDS can be estimated from conductivity, but only approximately. The conversion factor depends on which ions are actually present, and a figure calculated from conductivity is an estimate, not a TDS measurement.
- USP <645> is a published three-stage test with an actual numeric table, not a vague reference to "current USP." Stage 1 alone will resolve most routine measurements.
Resistivity and conductivity are the same measurement, read two ways
Resistivity and conductivity both describe how strongly water resists or conducts an electric current, and one is simply the reciprocal of the other. That ability is directly related to how much ionic material, commonly reported as total dissolved solids or TDS, is dissolved in the water. Water carrying a high TDS conducts readily and therefore shows low resistivity and high conductivity. Water with very little dissolved ionic material does the opposite: high resistivity, low conductivity. A meter that reports one can always be read as the other, and which one a given instrument displays is a matter of convention within a given industry rather than a difference in what is being measured.
The standard reference measurement for water purity is specific resistance corrected to 25°C, commonly written R-25. It is defined against a specific geometry: the resistance of a current passing between two 1 cm square plates spaced 1 cm apart, at 25°C. That geometry, one cubic centimetre of water between defined plates, is the theoretical standard every real probe is corrected back to.
Why the cell constant matters
Real conductivity cells are not built as flat 1 cm plates, because a cell has to fit into a pipe or a beaker rather than sit as an idealised laboratory geometry. Most are cylindrical, sized to the application, and the change in shape and electrode spacing changes the relationship between the measured signal and the true resistivity. The cell constant is the correction factor that maps a specific probe’s geometry back to the 1 cm cube standard, and it has to be programmed into the meter for the displayed reading to be accurate.
A worked example makes this concrete. If a cell has a constant of 1.0 cm⁻¹, the measured conductance and the true conductivity are numerically identical, because the cell behaves like the theoretical standard. A cell with a constant of 0.1 cm⁻¹ is built for lower-conductivity, higher-purity applications, and its raw signal has to be multiplied by that factor to give the correct reading. Using the wrong cell constant, or leaving a meter set to a default value that does not match the cell actually installed, produces a reading that looks entirely plausible and is simply wrong. This is worth checking specifically after any cell replacement, not assumed to carry over.
Temperature correction, and when to turn it off
Resistivity and conductivity both change with temperature, and in opposite directions: resistivity falls as temperature rises, conductivity rises as temperature rises. For readings to be comparable across time, or across different systems, they have to be corrected to a common reference temperature, and the standard for most applications is 25°C. A meter and cell combination has to measure both the electrical signal and the temperature simultaneously and apply the correction internally.
Most meters allow temperature compensation to be switched off, and there is a specific reason to do so. Current USP methodology for compendial water testing specifies non-temperature-compensated measurement for Stage 1 of the conductivity test, precisely because the compensation algorithm itself introduces an assumption about ion behaviour that the compendial method does not want in the result. If your quality system requires USP <645> compliance, confirm whether your instrument is configured to report the raw, uncompensated reading alongside temperature, rather than only the compensated figure.
What 18.2 actually means, and why it does not last
The resistivity of theoretically pure water, containing no dissolved ionic material at all, is 18.2 MΩ·cm at 25°C, equivalent to about 0.055 µS/cm. This is a ceiling rather than a target range: nothing exceeds it, because at that point there is nothing left to remove.
Water at that purity has to be measured inline, in a closed system, because open exposure degrades it almost immediately. As water is drawn from a system reading 18.2 MΩ·cm, atmospheric carbon dioxide begins dissolving into it straight away. The dissolved CO₂ reacts with water to form carbonic acid, which dissociates into ions that conduct current. The resistivity of a freshly drawn sample can fall from 18.2 to below 8 or 10 MΩ·cm in under a minute. This is not a fault in the water or the system; it is simply what open air does to water at the ionic purity where a handful of dissolved CO₂ molecules is enough to move the reading substantially. It is also why a bench reading taken from a drawn sample will always read lower than the inline figure the generation system reports, and why comparing the two is not evidence of a problem.
What resistivity and conductivity do not tell you
Resistivity and conductivity measure ionic purity only. Dissolved organic material does not carry a meaningful ionic charge and does not register on a conductivity reading, so water that reads at 18.2 MΩ·cm can still carry organic contamination that a conductivity meter has no way to detect. Screening for that requires a separate method, either total organic carbon, TOC, analysis or a chromatographic technique aimed at the specific contaminant class of concern. A conductivity specification and a TOC specification are answering two different questions, and meeting one says nothing about the other.
Estimating TDS from conductivity
Natural and municipally treated water carries a wide range of TDS, from under 50 ppm to several hundred, depending on the mineral content of the source. Potable water typically carries calcium, magnesium and sodium as the dominant cations, paired with carbonate, sulfate and chloride, all dissolved from rock and mineral contact as water moves through the ground and into the supply system. Because these dissolved salts are what allow the water to conduct current, conductivity or resistivity can be used to estimate TDS reasonably quickly and far more cheaply than a full gravimetric TDS determination.
The estimate depends on a conversion factor, and the factor depends on which ions are present. A common working approximation is TDS in ppm equal to conductivity in µS/cm multiplied by a factor around 0.5 for a sodium-chloride-dominated water, ranging up toward 0.7 or higher for water with a different ionic mix. Two examples show the arithmetic: a sample with a specific resistance of 4,000 Ω·cm has a conductivity of 1 ÷ 4,000 Ω·cm, or 250 µS/cm, and at a factor of 0.5 that gives an estimated TDS of about 125 ppm. A sample at 600 Ω·cm has a conductivity of about 1,667 µS/cm, giving an estimated TDS of roughly 835 ppm at the same factor. Both figures are estimates rather than measurements, and the further the actual water chemistry departs from the sodium-chloride assumption, the less reliable the estimate becomes.
USP <645>: what the compendium actually requires
This is the section most articles on this topic skip, and it is the part worth bookmarking. USP General Chapter <645>, Water Conductivity, sets out a three-stage test for compendial water. Stage 1 is intended for online, non-temperature-compensated measurement, compared directly against a published table of temperature-conductivity limits. If the reading exceeds the Stage 1 limit for the measured temperature, testing proceeds to Stage 2.
Temperature (°C) | Stage 1 conductivity requirement (µS/cm) | Temperature (°C) / requirement |
0 | 0.6 | 50 / 1.9 |
5 | 0.8 | 55 / 2.1 |
10 | 0.9 | 60 / 2.2 |
15 | 1.0 | 65 / 2.4 |
20 | 1.1 | 70 / 2.5 |
25 | 1.3 | 75 / 2.7 |
30 | 1.4 | 80 / 2.7 |
35 | 1.5 | 85 / 2.7 |
40 | 1.7 | 90 / 2.7 |
45 | 1.8 | 95 / 2.9 |
This table is reproduced from USP <645>, Table 1, Stage 1 — Temperature and Conductivity Requirements, for non-temperature-compensated measurements only. Two independent captures of the compendial text agree exactly on every value shown, across two different USP editions. The table continues beyond 95°C; confirm the top end against your current USP-NF subscription rather than a secondary source, since that range is where the sources I could access became inconsistent with each other.
Stage 2, if Stage 1 is exceeded, works differently: a sample is stirred to equilibrate with atmospheric CO₂ while conductivity is monitored, and once the rate of change falls below 0.1 µS/cm per 5 minutes, the reading is taken at 25°C ± 1°C. If that reading does not exceed 2.1 µS/cm, the water passes. Stage 2 measurements may be temperature-compensated or not, which is a specific departure from the strict non-compensation requirement of Stage 1.
Stage 3, reached only if Stage 2 also fails, saturates the Stage 2 sample with potassium chloride and measures pH, then compares the conductivity against a published limit that varies with the measured pH. In practice, purified water and water for injection systems running at or near ambient temperature rarely reach Stage 3, and the full pH-dependent limit table is a compendial reference to consult directly if your result gets that far, rather than something worth carrying in a general working guide.
Instrument requirements under USP <645>
- Temperature measurement accuracy within ±2°C. This is what the Stage 1 lookup depends on being correct.
- Cell constant known and accurate to within ±2%, determined against reference instrumentation or solutions traceable to a national standards authority.
- Display resolution of at least 0.1 µS/cm on the lowest range the instrument reports.
- Regular calibration against the manufacturer’s schedule, with the cell constant re-verified rather than assumed to remain stable indefinitely.
Reading a conductivity meter without being misled
- Check whether the display is compensated or raw. The two will read differently above and below 25°C, and comparing a compensated figure against an uncompensated compendial limit is a common source of an apparent out-of-specification result that is not actually one.
- Confirm the cell constant programmed into the meter matches the cell physically installed, particularly after any cell replacement or meter swap.
- Do not compare a bench sample against an inline generation reading and expect agreement. CO₂ uptake in the seconds after drawing a sample will always push a bench reading lower than the closed-system figure.
- Remember conductivity is silent on organics. A good conductivity reading is necessary and is not sufficient on its own for water quality, and a TOC excursion can sit underneath a passing conductivity result. Our guide to pharmaceutical grade water and USP standards covers the wider specification picture, including where TOC and microbial limits fit alongside conductivity.
Resources and further reading
This article was produced under Lab Manager’s AI Editorial Guidelines
