Three ways of saying the same thing
Conductivity measures how easily water carries current, which depends on how many dissolved ions it holds. From that one measurement come three commonly quoted numbers:
Conductivity and resistivity are strict reciprocals — pure convention decides which one an industry uses. Ultrapure water people speak in MΩ·cm because the numbers are convenient there; everyone else uses µS/cm. TDS is the odd one out because it is an inference, not a measurement.
Why the 25 °C reference matters
Ions move more freely in warm water, so conductivity rises about 2% per degree Celsius. A sample reading 500 µS/cm at 30 °C is really 455 µS/cm at the 25 °C reference — a 9% difference from temperature alone. Without compensation, the same cooling water would appear to change composition between a cool morning and a hot afternoon.
Every industrial conductivity analyser therefore has a temperature sensor and applies this correction automatically. When two instruments disagree on the same stream, mismatched compensation settings are the first thing to check — a coefficient of 2.0 versus 2.1 shows up as a persistent offset that no amount of recalibration will fix.
Worked example
A cooling tower blowdown analyser reads 500 µS/cm at 30 °C:
- C₂₅ = 500 ÷ (1 + 0.02 × 5) = 500 ÷ 1.10 = 454.5 µS/cm
- TDS at 0.5 factor = 227 ppm
- Resistivity = 10⁶ ÷ 454.5 = 2,200 Ω·cm
Typical water quality bands
| Water | Conductivity @25 °C | Resistivity |
|---|---|---|
| Ultrapure (theoretical limit) | 0.055 µS/cm | 18.18 MΩ·cm |
| Semiconductor grade | < 0.1 µS/cm | > 10 MΩ·cm |
| Boiler feedwater | 0.1 – 5 µS/cm | 0.2 – 10 MΩ·cm |
| Distilled / RO permeate | 1 – 50 µS/cm | 20 – 1000 kΩ·cm |
| Drinking water | 200 – 800 µS/cm | 1.2 – 5 kΩ·cm |
| Cooling tower / brackish | 1,000 – 5,000 µS/cm | 0.2 – 1 kΩ·cm |
| Seawater | ≈ 50,000 µS/cm | ≈ 20 Ω·cm |
Field notes
- TDS is an estimate, not an analysis. The conversion factor assumes a particular ionic makeup. Two waters with identical conductivity can have genuinely different dissolved solids, so use gravimetric TDS when the number matters legally.
- Ultrapure water breaks the 2%/°C rule. Below about 1 µS/cm the temperature behaviour is dominated by water self-ionisation and needs a non-linear compensation curve — most pure-water analysers have a dedicated mode for it.
- Cell constant matters as much as calibration. A cell constant of 0.1 suits pure water and 10 suits high-conductivity brines; the wrong cell for the range gives poor resolution regardless of how carefully it is calibrated.
- Fouling reads low. A coated electrode under-reports, so a slowly falling conductivity trend on an unchanged process usually means the cell needs cleaning rather than the water getting purer.
- Analyser output still travels as 4-20 mA — check the scaling with the 4-20 mA converter, and see the pH electrode calculator for the other workhorse of water analysis.
Frequently asked questions
How do I convert conductivity to TDS?
Multiply the temperature-compensated conductivity in µS/cm by a conversion factor, typically 0.5 for a sodium chloride reference or 0.65 to 0.7 for natural waters. A reading of 500 µS/cm at 25 °C gives about 250 ppm TDS on the 0.5 factor.
Why does conductivity need temperature compensation?
Ion mobility rises with temperature, so the same water conducts better when warm — roughly 2 percent per degree Celsius. Without compensation the same sample would read differently morning and afternoon, so instruments normalise everything to 25 °C.
What is the resistivity of ultrapure water?
18.18 MΩ·cm at 25 °C, which corresponds to a conductivity of 0.055 µS/cm. That is the theoretical limit set by water self-ionising, and it is the benchmark for semiconductor and pharmaceutical grade water.
What conductivity is normal for drinking water?
Typically 200 to 800 µS/cm, roughly 100 to 500 ppm TDS. Below about 50 µS/cm suggests treated or distilled water; above 1500 µS/cm is brackish and above 50,000 µS/cm approaches seawater.
Provided for reference and education using linear temperature compensation. See our disclaimer.