The mechanism: two simultaneous attacks
A standard pH electrode contains two electrochemical subsystems: the glass membrane (pH-sensitive) and the reference electrode (a fixed-potential anchor). In clean aqueous solutions, H₂S affects neither. In real O&G process streams, H₂S primarily attacks the reference electrode — dissolved sulphide enters through the liquid junction and poisons the Ag/AgCl element. The glass membrane is generally resistant to H₂S at process-relevant concentrations.
1. Reference poisoning — the Ag/AgCl pathway
The most common pH reference design uses a silver wire coated with silver chloride (Ag/AgCl), immersed in a concentrated KCl filling solution. The half-reaction that maintains the stable reference potential is:
When dissolved H₂S enters the filling solution through the liquid junction, it reacts with the silver wire directly:
Silver sulphide (Ag₂S) has a solubility product of approximately 6 × 10⁻⁵⁰ — it is essentially insoluble. Once formed, it precipitates as a black coating on the silver wire, altering the effective electrode area and, critically, shifting the reference potential. The half-reaction is no longer Ag/AgCl but a mixed potential incorporating the Ag/Ag₂S system (E° = −0.691 V vs. SHE), a shift of up to 0.9 V in the theoretical limit; in practice, mixed potentials typically produce shifts of 50–400 mV depending on the extent of poisoning.
2. Liquid junction blockage — the KCl pathway
Even before H₂S reaches the silver wire, it causes a secondary problem at the liquid junction. The junction connects the KCl filling solution to the process stream. In H₂S-rich environments:
- Dissolved sulphide species (HS⁻ and H₂S) react with trace heavy metals (Cu²⁺, Pb²⁺) dissolved in the filling solution, precipitating metal sulphides that physically block the porous ceramic junction.
- Colloidal sulphur formed by partial oxidation deposits in the junction pores.
- The junction potential — already a source of calibration uncertainty — shifts unpredictably as flow through the junction drops.
A partially blocked junction increases the junction resistance, slows electrode response, and introduces a residual liquid junction potential (RLJP) that cannot be corrected by simple two-point calibration.
How large is the error?
The magnitude of the error depends on H₂S concentration, temperature, process ionic strength, and how far the poisoning has progressed. From field data across sour water stripper and SAGD produced water applications:
| H₂S (mg/L) | Exposure time (hours) | Typical pH error | Visual indicator |
|---|---|---|---|
| < 1 | Any | ± 0.05 | None |
| 1 – 10 | < 24 | ± 0.1 – 0.3 | Light discolouration of junction |
| 1 – 10 | 24 – 168 | ± 0.3 – 0.7 | Black tinge on Ag wire (if visible) |
| > 10 | < 8 | ± 0.5 – 1.5 | Often none until junction is inspected |
| > 50 | > 2 | Total loss of reference stability | Readings drift continuously |
These ranges assume a standard Ag/AgCl double-junction reference with ceramic porous pin junction. Actual errors in your process will vary.
Why the glass membrane is usually not the problem
Engineers often suspect the glass membrane first — it is the visible sensing element. In H₂S environments, the glass membrane is typically not the primary failure mode. Most commercial pH glass formulations are chemically resistant to H₂S at concentrations below several thousand mg/L. The Ag/AgCl reference fails orders of magnitude faster.
This is why replacing the glass bulb (or buying an entirely new combination electrode and using the same KCl filling solution) does not fix the problem: the new glass membrane is calibrated against a still-poisoned reference.
Four engineering solutions, in order of preference
1. Double-junction reference with non-silver inner chamber
The most effective approach eliminates silver from the reference path. Designs using alternative reference elements that avoid silver — such as Tl/TlCl (Thalamid® by Metrohm), Ross-type double junction, or iodide/iodine references — maintain stable potential in H₂S without sulphide poisoning. The inner filling solution uses a non-reacting electrolyte (e.g., 3 M KNO₃), with the outer KCl chamber acting as a sacrificial barrier.
This is the preferred option for continuous in-line monitoring in sour water streams above 10 mg/L H₂S. Electrode cost is 3–5× standard, but maintenance frequency drops by the same factor.
2. High-flow sleeve junction
A PTFE sleeve junction with high KCl outflow (typically 0.5–2 mL/day) physically prevents process ingress. The continuous outward KCl flow keeps H₂S from reaching the silver wire. This requires pressurised filling solutions (or a pressurised reference reservoir) and regular filling solution top-up.
Practical for applications where the electrode is accessible for weekly maintenance. Not suitable for remote or unattended installations.
3. Solid polymer electrolyte (SPE) pH sensors
ISFET and solid-state polymer electrolyte sensors eliminate liquid junctions and Ag/AgCl references entirely. They have no junction to block and no silver to poison. Their limitation in H₂S environments is different: the SPE membrane can be attacked by dissolved sulphide at pH < 4 where H₂S is the dominant species (pKa₁ = 6.98 at 25°C).
For pH 5–12 sour water applications, solid-state sensors are an excellent choice. Below pH 5 in high H₂S, evaluate membrane compatibility with the specific polymer type.
4. Sample conditioning and dilution
In some applications — particularly lab or at-line measurement of process samples — conditioning the sample before electrode contact is more economical than specialised electrodes. Stripping dissolved H₂S by acidification (drop sample pH to < 2 briefly, then neutralise to measurement pH) or by nitrogen sparging removes sulphide before electrode contact.
This adds complexity and introduces its own sources of uncertainty, but allows standard electrodes to be used reliably.
What to specify in your next RFQ
When sourcing pH electrodes for H₂S service, require the following from vendors:
- Reference type: non-Ag/AgCl inner reference, or documented H₂S compatibility test data (not just a marketing claim)
- Junction type: sleeve or ceramic; specify maximum H₂S concentration tested
- Body material: HDPE (PEAD), PTFE, or PEI (Ultem) — not polysulfone, which is attacked by some amine solvents present in sour water
- Process pressure rating: match to your vessel or pipeline pressure, not just the electrode's rated maximum
- Calibration interval: ask for MTBF data specific to H₂S concentration range, not generic intervals
Documentation and QA implications
If your process operates under ISO 17025, ASTM D1293, or an internal QMS that requires pH measurement traceability, a poisoned reference is not just a calibration problem — it is a data integrity problem. Any pH records collected with a compromised electrode are suspect.
Recommended practice for H₂S environments: calibrate before and after each measurement session (or each shift for continuous analysers), keep the calibration records, and establish acceptance criteria for slope and offset. A slope below 54 mV/pH at 25°C in a new electrode is a hard stop — recalibrate with a known-good electrode before accepting data.
Summary
- H₂S attacks the Ag/AgCl reference irreversibly via Ag₂S formation. The poisoning shifts the reference potential by up to 0.9 V equivalent.
- Junction blockage by metal sulphides and colloidal sulphur creates unpredictable residual liquid junction potentials.
- The glass membrane is rarely the primary failure mode in H₂S service.
- Errors of 0.5–1.5 pH units are common in sour water above 10 mg/L H₂S with standard electrodes.
- Solutions: non-Ag inner reference, high-flow sleeve junction, ISFET/solid-state, or sample conditioning — in order of preference for continuous monitoring.
- Standard recalibration does not fix a poisoned reference. Replace it.