Conductivity meter calibration: standards, cells and temperature

Conductivity meter calibration must identify the components and conditions assessed: electronics, cell, temperature and reported output. An electrical check alone does not establish cell behaviour in solution. A reference’s certified value is usable only under conditions supporting its validity. Define this scope before interpreting a result or changing the cell constant.

1. Describe the chain to be authorised

Identify the indicator, channel, cell and serial number, cable, temperature sensor, software and settings. Record the intended range, units, compensated or uncompensated mode and any reference temperature setting. Replacing the cell while retaining the indicator may change the chain covered by the certificate.

Conductivity κ differs from the conductance G of the measurement configuration. In the elementary model κ = K × G = K/R, K is the cell constant and R the relevant resistance. With K in cm⁻¹ and G in µS, κ is in µS/cm. This model does not remove electrode, connection, frequency or actual geometry effects; demonstrate its adequacy over the required range.

For example, K = 0.1000 cm⁻¹ and G = 1000 µS give 100.0 µS/cm. This is an illustrative numerical relationship, not a recommended calibration point. Electrical simulation can assess indication and configured conversion; it does not automatically reproduce immersion, surfaces, contamination or solution response.

2. Check the reference before comparison

A reference solution needs identification and batch, assigned property, value at the stated temperature, uncertainty, coverage factor where applicable, validity and storage/use instructions. The nominal salt concentration or prominent label number does not replace the assigned conductivity and its uncertainty.

Select references compatible with the operating range, required performance and cell construction. One high point does not demonstrate a much lower range. Suitability also includes matrix, solvent, compatibility and transfer conditions: do not independently dilute a reference and assign a certified conductivity to the new solution by proportional calculation.

Respect requirements for the container, opening, exposure and aliquot use. Evaporation, carryover, container interaction and CO₂ exchange may alter the liquid. Do not return used portions to the bottle. At low conductivity, a small ionic contribution may matter; “not expired” does not establish that a handled aliquot remains valid.

The historical NIST SRM 3191 certificate distinguishes the value at 25 °C from conditions after opening. The consulted copy expired on 29 November 2013: it is a documentary example, not a reference suitable for use today. Follow the actual certificate for the current material rather than applying those instructions to every solution.

3. Component, test and coverage matrix

Use this original matrix to agree service scope. Its purpose is to expose what remains outside a comparison, rather than multiply tests without a metrological question.

4. Cell constant and compensation

Retain the as-found K, original data and every newly stored K. A function labelled “calibration” may change the software parameter: distinguish that adjustment from characterisation before and after intervention. Do not force the display to a nominal value to conceal a questionable reference.

Temperature must represent the liquid interacting with the cell. A correct sensor in a different location may not describe electrode conditions. Stability, uniformity and equilibrium belong in the assessment; no universal waiting time is prescribed.

Temperature compensation calculates a value at another temperature without physically changing the sample. One possible linear model is κ_ref = κ_T/[1 + α(T − T_ref)], where α has units °C⁻¹. Use it only within the range and composition that justify it. A common coefficient is not universal, especially near very pure water conductivity.

Compare values at the same temperature or on the same documented calculation basis. Record whether compensation is disabled, the model and the reference temperature. USP ⟨644⟩ and ⟨645⟩ have different scopes: water requirements and analytical limits do not automatically become calibration criteria for every conductivity meter.

5. Uncertainty, criteria and records

Depending on the method, the budget considers reference value, stability after opening, temperature, repeatability, resolution, K, geometry, bubbles and contamination. If K is derived from the same standard used for verification, consider the dependence of the evidence; do not describe that reuse as a fully independent check.

Define e = indication − reference and c = −e for comparable quantities. Distinguish error, standard uncertainty u, expanded uncertainty U = ku and tolerance. Do not add contributions already included in the result uncertainty twice, or use a water tolerance as the instrument’s U.

The report must allow reconstruction of configuration, references, temperatures, points, modes, data before and after changes, results and limitations. Where conformity is requested, agree the requirement and decision rule beforehand. After cleaning, replacement or adjustment, define relevant checks for return to service and intermediate checks according to risk.

6. Simulated case: the aliquot has changed

A solution assigned 5.00 µS/cm at 25 °C was transferred into a previously used container and left exposed. The reading is 6.20 µS/cm. The +1.20 µS/cm difference from the certificate is not yet an established instrument error: it has not been confirmed that the assigned value still describes this aliquot.

Retain the reading, times, container and settings; suspend use of the aliquot as a reference. A planned comparison using valid material and controlled handling, without first adjusting the system, gives 5.02 µS/cm against 5.00 µS/cm, with U = 0.06 µS/cm and k = 2 for the error. The simulated internal limit is |e| ≤ 0.10 µS/cm and the rule is |e| + U ≤ limit: 0.02 + 0.06 = 0.08 ≤ 0.10.

That point satisfies the selected rule; it does not authorise the full range. The evidence directs the investigation towards reference handling without identifying a contaminant or excluding other causes by itself. Assess the impact of any previous settings based on the questionable aliquot. Do not delete the initial result or “repair” the comparison by changing K until 5.00 appears.

For analytical use, see conductivity measurement; for metrological review, see measurement uncertainty and the temperature chain. Find other pathways in the Calibration and Laboratory Metrology hub.

7. Sources and limitations

Checked on 1 October 2026. The matrix and case are original; numbers and criterion are not compendial requirements.

1. Shreiner and Pratt, NIST SP 260-142, 2004 edition, §§1.1, 2.2–2.3, 5.2, 6.3 and 7: full text consulted for principles, references and stability; historical metrology document.
2. NIST SRM 3191, batch 101203, certificate dated 16 November 2012: archived example expired in 2013, not material proposed for current use.
3. USP ⟨644⟩ and USP ⟨645⟩: official introductory previews only, bibliographic records 2020 and 2017; current full text not consulted. Check the applicable edition before a compendial decision.