A pH probe can pass calibration and still be biased in a sample that behaves differently from the standard buffers. The often-overlooked reason is the liquid-junction potential at the contact between the sample and the reference-electrode solution. Different ions cross this interface at different rates, and the resulting voltage becomes part of what the meter reads.
This guide explains the hidden measurement term with a checkable 25 °C Nernst-slope model: a fictional +12 mV sample-specific shift can create roughly 0.20 pH unit of display bias. It then separates slope and temperature compensation from matrix effects and shows how to test whether a reported pH value is defensible.
eduKateSG · The Gold Standard of Chemistry
Choose your reading route
Understand the instrument · Chapters 1–3. Reference, voltage and error calculation.
Check the method · Chapters 4–6. Buffers, temperature and sample matrices.
Test and transfer · Chapters 7–8. FAQs, linked chemistry and sources.
Chapter index
Understand the instrument · Chapters 1–3
Check the method · Chapters 4–6
Test and transfer · Chapters 7–8
A modern pH meter usually measures a potential difference involving a pH-sensitive glass membrane and a reference-electrode system. Many combination probes place both within one body. The electronics may display a simple pH number, but the instrument’s signal belongs to a larger electrochemical circuit that includes contact between different solutions.
IUPAC defines a reference electrode as an electrode intended to maintain a virtually invariant potential under the prevailing measurement conditions, providing a stable comparison for the indicator electrode. That phrasing is important: the reference is a controlled measurement component, not a metaphysical source of zero voltage.
An Ag/AgCl reference commonly uses a filling solution containing chloride. A porous junction or other interface allows ionic contact between that solution and the unknown test solution. At the junction, different ions can diffuse with different mobilities. Their transport can establish an additional electric potential difference.
IUPAC calls this interface a liquid junction and the resulting potential a liquid-junction potential. The operational pH cell has such a junction between the sample or standard and the reference filling or bridge solution. Its behaviour can differ in standard buffers and in a difficult real sample.
That is why a meter can pass ordinary calibration and still encounter bias in a strongly different matrix. The measurement is an electrochemical potential difference, and the reference junction forms part of what is measured.
The useful scientific question is not ‘Does my instrument show a number?’ but ‘Which potentials contribute to that number, and which remain sufficiently unchanged when I replace the calibration buffer with the actual sample?’
A simplified pH-cell response can be represented as E_measured = C − S×pH + E_j, where C gathers a chosen constant offset, S is a positive response magnitude in millivolts per pH unit and E_j represents a junction-dependent term under the stated sign convention.
At 25 °C, an ideal Nernstian response to a monovalent hydrogen-ion activity change has a slope magnitude near 59.16 mV per pH unit. The magnitude follows 2.303RT/F at 298.15 K, using gas constant R, absolute temperature T and Faraday constant F. Actual electrode response can depart from the ideal due to condition, aging, calibration and sample effects.
For teaching, choose the arbitrary offset so that E=0 mV at pH 7 when the junction contribution has the same value as in the calibration standards. Then write E_ideal=−59.16(pH−7) mV. This is one polarity convention chosen for clarity, not a universal meter wiring diagram.
At pH 6 the ideal voltage is +59.16 mV; at pH 4 it is +177.48 mV. A difference of three pH units corresponds to 177.48 mV in this ideal model. That ratio is a calibration relationship, not the absolute electrochemical potential of a hydrogen ion in solution.
A two-buffer calibration can estimate a slope and intercept from standard solutions. However, it does not guarantee that a sample with a very different ionic composition has exactly the same junction term as both buffers.
Keep the distinction between calibrated pH and chemical activity. The hydrogen-ion activity guide explains why pH itself is defined through a standardised activity convention, rather than direct counting of free ions.
| Toy true pH | Ideal signal relative to pH 7 | Meaning |
|---|---|---|
| 4.00 | +177.48 mV | Three units below pH 7 |
| 6.00 | +59.16 mV | One unit below pH 7 |
| 7.00 | 0.00 mV | Chosen response reference |
| 8.00 | −59.16 mV | One unit above pH 7 |
Now suppose two suitable buffers have been used to set the toy ideal calibration response. In the unknown test sample the desired pH value is 6.00 and the pH-sensitive component would give +59.16 mV under that response. But the junction contributes an additional +12.0 mV relative to the calibration conditions.
The measured potential is then 59.16+12.00=71.16 mV. If the instrument assumes the calibration relationship E=−59.16(pH−7) with no changed junction term, its reported pH becomes 7−71.16/59.16≈5.797.
The sample’s actual pH in this invented example is 6.000. The displayed value is lower by approximately 0.203 pH unit, solely because the example introduced an uncompensated 12 mV sample-specific junction shift. The sign of a real error depends on the instrument’s conventions and physical circumstances.
An equal-sized −12.0 mV junction shift would move the toy displayed value in the opposite direction to approximately 6.203. This is a deliberately transparent sensitivity calculation. It does not claim that all real reference junctions create 12 mV shifts.
One should not conclude that every 0.20 pH discrepancy is caused by a junction or that adding a universal offset fixes every unknown solution. The educational result is narrower: an electrochemical error of a few millivolts can be consequential because the ideal pH response is only about 59 mV per unit near room temperature.
Check the units. Dividing 12.0 mV by 59.16 mV per pH yields about 0.203 pH unit. The arithmetic is meaningful only after explaining that the voltage shift is an uncompensated differential relative to the calibration. The absolute cell voltage contains other terms and cannot simply be divided by 59.16.
| Quantity | Toy value | Role |
|---|---|---|
| Correct sample pH | 6.000 | Hypothetical target value |
| Ideal electrode signal | +59.16 mV | Chosen 25 °C response |
| Extra junction contribution | +12.00 mV | Relative to calibration |
| Actual toy voltage | +71.16 mV | Sum of assumed terms |
| Displayed pH if ignored | ≈5.797 | Biased inferred value |
| Display bias | ≈−0.203 pH | Under this defined polarity |
A pH meter calibrated in accepted standard buffers often performs very well within the conditions for which the method was validated. A reference electrode’s salt bridge or junction is designed to make its potential small and reasonably reproducible. Common filling-electrolyte choices seek similar cation and anion mobilities to reduce diffusion-induced voltages.
The IUPAC Gold Book describes high-concentration reference filling solutions with cations and anions of nearly equal mobility as a way to keep liquid-junction potential small and approximately constant as one replaces standard solutions with test solutions. That reduces the problem; it does not prove the potential is identical for every sample matrix.
Consider a routine buffer at moderate ionic strength and a very low-conductivity water sample. The ionic transport conditions at the reference junction can differ appreciably. The electrode can also respond slowly or exhibit unstable readings because the sample’s low buffer capacity, carbon dioxide exchange and liquid-junction characteristics become important.
A highly saline or chemically unusual sample raises different challenges. Chloride activity, ionic strength and sample-specific reactions may influence the reference system or measurement method. A probe advertised as suitable for ordinary aqueous samples may not deliver the same uncertainty in concentrated process liquors.
Now consider a soil suspension. IUPAC has a separate term for anomalous liquid-junction potential arising in contact with charged suspended particles. It notes that such potentials can exceed ordinary junction effects under relevant conditions. That is a concrete reason to use field-specific soil pH procedures rather than transferring a clear-water method without validation.
In any of these cases, the standard-buffer calibration establishes that the meter responds to standards. To establish accuracy in the unusual matrix, the analyst may need a suitable reference method, matrix comparisons, replicate measurements and an uncertainty statement that includes residual junction effects.
The ideal Nernstian voltage slope depends on absolute temperature. Near 25 °C its magnitude is about 59.16 mV per pH unit. At 35 °C, using 308.15 K in the same formula, the magnitude is approximately 61.15 mV per pH unit, about 3.4% larger.
A temperature-compensated meter can adjust the expected electrode slope when temperature changes. It cannot automatically restore identical chemical pH values across temperatures, because acid–base equilibria and reference-buffer pH values can also change. Nor can it infer an unknown sample-specific junction potential from temperature alone.
Suppose the true pH is 6.0 and the meter’s ideal response is referenced to pH 7. At 35 °C the magnitude of a one-pH-unit signal would be about 61.15 mV rather than 59.16 mV. If an idealised instrument incorrectly used the 25 °C slope, the reported pH would be 7−61.15/59.16≈5.966. This is a small but checkable example of slope mismatch under a chosen sign convention.
The earlier 12 mV junction scenario is a different error source. Correcting the temperature-based slope would not remove that hypothetical voltage term. Similarly, a reference electrode can drift over time or become contaminated. Temperature electronics cannot identify every cause of drift.
Record the sample temperature, buffer certification conditions, meter’s compensation setting, electrode type and calibration date. In high-stakes analytical contexts, calibrate and verify according to the specific method rather than relying on a generic ‘ATC enabled’ display.
Finally, avoid confusing temperature dependence of the Nernst equation with temperature dependence of the water’s true pH. One belongs to the instrument response, the other to chemical equilibrium. Both can matter in the same observation.
| Assumed temperature | Ideal slope magnitude | One-pH-unit toy signal |
|---|---|---|
| 25 °C | ≈59.16 mV/pH | 59.16 mV |
| 35 °C | ≈61.15 mV/pH | 61.15 mV |
Suppose two calibrated probes disagree by 0.15 pH unit when placed into a low-ionic-strength laboratory sample. The observed discrepancy might involve reference-junction potentials, incomplete equilibration, contaminated membranes, calibration differences, temperature, dissolved carbon dioxide exchange or changes caused by stirring and exposure to air. A single discrepancy cannot identify one cause.
Start with a controlled repeat in appropriate check standards independent of the calibration set. Record measured potentials if available, stabilisation time and response slope. Inspect the physical probe and junction for blockage or deposits, and follow the manufacturer’s validated cleaning and storage procedures.
Next repeat the test using a suitable method for the actual sample matrix. For very low conductivity, aggressive solvents, strongly saline media or suspensions, the selected electrode design and procedure may need to differ from a general-purpose classroom pH meter. An appropriate double-junction or flowing-junction design can sometimes be advantageous, but selection depends on compatibility and validation.
Check whether the samples are changing while being measured. Low-buffer-capacity water may exchange CO₂ with air, changing its real pH; a temperature gradient may affect both equilibrium and electrode response. Rapid transfer between matrices can produce delayed junction equilibration. Compare readings after consistent, documented stabilisation, not arbitrary durations chosen after seeing the result.
Treat duplicate electrode agreement as a useful check, not conclusive proof. Two probes of similar design may share biases in the same unusual matrix. Independent standards and appropriately different reference methods can expose systematic effects better than simply averaging two displays.
NIST’s pH metrology programme shows how primary measurement methods and certified pH reference materials support traceability. A routine meter is part of that chain, not its origin. When matrix effects matter, the limitations of transferring the calibration from standards to the unknown should appear in the uncertainty budget.
Do not translate the hypothetical 12 mV example into a universal service rule. Some actual junction differences may be much smaller or larger. A defensible laboratory result includes an empirically justified uncertainty, selected method and the evidence that the electrode is suitable for its intended sample.
Does a reference electrode always measure zero volts?
No. It provides a sufficiently stable potential under specified conditions so changes in the working or sensing electrode can be interpreted relative to it.
What is a liquid junction?
It is contact between electrolyte solutions of differing composition. Differences in ionic transport can create a potential across that interface.
Can perfect buffer calibration eliminate every junction error?
No. A test sample may have a different junction potential from the calibration standards, leaving residual bias.
Is a 12 mV junction shift always a 0.203 pH error?
Only in the article’s ideal 25 °C, monovalent-response sensitivity calculation when the full 12 mV appears as an uncompensated differential signal. Real slopes and conditions may differ.
Why are potassium-chloride reference solutions common?
Their ion-mobility properties and chemical compatibility can support a small, stable junction potential under appropriate measurement conditions. Exact designs vary.
Can temperature compensation correct a bad junction?
It generally adjusts the electrode’s expected temperature-dependent response. It cannot know every sample-specific junction potential.
Why are low-conductivity waters difficult?
The junction and sample may equilibrate slowly, the liquid may have little buffer capacity, and exposure or contamination can alter pH. Suitable specialised procedures may be needed.
Is a soil suspension equivalent to clear buffer solution?
No. Contact with charged particles can create additional junction effects and sample-dependent behaviour. Soil measurements require appropriate methods.
Can I add a fixed 0.20 to every pH reading?
No. The hypothetical example is not an instrument correction. Any adjustment must be justified by a validated method and the actual conditions.
The core learning is that a number from an instrument represents a system of components and conventions. A stable reference electrode, a sensing electrode, calibrated standard buffers and a sample-specific interface all contribute to a pH measurement. Understanding the chain prevents an exact-looking digital display from becoming an unjustified chemical claim.
Repeat the worked example after replacing +12 mV with +6 mV. Under the same slope and polarity, the approximate pH bias halves to −6/59.16≈−0.101. Then ask why the instrument may not follow this exact arithmetic if temperature, slope or the matrix changes. The answer is that the calculation is a transparent conditional model, not a substitute for verification.
Subject and connected guides
The Gold Standard of Chemistry — parent subject guide.
Hydrogen-ion activity and pH — the thermodynamic quantity behind the calibrated scale.
Ionic strength, activities and equilibrium constants — understand why solution composition matters.
Carbonate alkalinity and titration endpoints — distinguish pH from a separate analytical water measurement.
Industrial pH meter calibration — practical operation and verification.
The Gold Standard of Education — apply transparent models to new evidence.
How X Works — wider scientific connections.
Authoritative sources and scope
References reviewed on 22 October 2026. All displayed mV values, pH readings and bias examples are original simplified teaching cases with an explicit sign convention. They are not manufacturer specifications, laboratory validation results or recommended pH corrections.
IUPAC Gold Book: Liquid Junction — definition and the connection to operational pH cells.
IUPAC Gold Book: Reference Electrode — required reference stability and common designs.
IUPAC Gold Book: Filling Solution — reference filling compositions and mobility-related junction control.
IUPAC Gold Book: Anomalous Liquid-Junction Potential — additional concerns in charged-particle suspensions.
NIST/IUPAC: Measurement of pH—Definition, Standards and Procedures — operational standards and metrological practice.
NIST: pH Metrology — primary measurements and reference-material traceability.