All three industrial pH sensing technologies convert hydrogen-ion activity into a measurable potential, but the conversion mechanism differs fundamentally: a hydrated glass membrane for the glass electrode, an MOS-transistor gate coated with an amphoteric metal oxide for the ISFET, and a pure antimony metal surface that develops a potential via surface oxide for the antimony electrode [S1][S3][S6].
The practical consequences are direct. Glass remains the accuracy benchmark at roughly ±0.1 pH, antimony trails at about ±0.2 pH, and ISFET reaches approximately 3× better accuracy than glass in side-by-side tests while offering 5× the mechanical ruggedness of a glass membrane [S1][S8].
Glass electrode: ion-exchange at a hydrated silica membrane
A glass pH electrode uses a thin-walled glass bulb (typically Li₂O-Na₂O-SiO₂ formulation) as the hydrogen-selective element; hydrogen ions on the outer gel layer exchange for alkali ions in the glass, producing a phase-boundary potential that follows the Nernst equation at about 59.16 mV/pH at 25 °C [S4][S5]. The reference half-cell is normally Ag/AgCl immersed in a KCl filling solution, and the measured voltage between the two is converted to pH by the meter [S4].
Glass is selective for hydrogen ions, but not perfectly so, and at high pH and high sodium concentrations the membrane starts responding to Na⁺ as well, producing the classic "sodium error" that limits useful range to roughly pH 0-12 in low-alkali samples and to pH 0-13 in low-Na matrices [S5][S7]. Operating temperature generally tops out near 100 °C for general-purpose glass and around 110 °C for the low-impedance "blue glass" formulation used in hot caustic and hot-process applications [S7].
For specifiers weighing corrosion and material limits, the same constraint that drives glass-quartz and sight-glass selection on pressure-vessel viewports also drives the choice between standard and alkali-resistant glass here: it is the chemistry of the wetted surface, not the brand label, that sets the service envelope.
ISFET: gate voltage controlled by surface protonation
ISFET sensors replace the glass membrane with an MOS transistor whose metal gate is replaced by a hydrogen-sensitive insulator such as SiO₂, Si₃N₄, Al₂O₃, or the industry-preferred Ta₂O₅ [S2][S3]. Hydronium or hydroxide ions bind to this amphoteric surface, generating a surface charge that modulates the source-drain current; the transmitter electronics hold that current constant and read out the required gate voltage U<sub>GS</sub>, which is proportional to pH [S2][S3].
Process-ready ISFET sensors appeared in the mid-1990s, and a key breakthrough was the introduction of a Ta₂O₅ gate in 2002 that enabled steam sterilisation and fast response down to low temperatures; the launch product was the Endress+Hauser CPS471 hygienic sensor [S2]. A subsequent dual-layer Ta₂O₅ gate was developed specifically to survive hot NaOH clean-in-place (CIP) cycles, which were destroying the original single-layer gates and forcing extra maintenance [S2].
Behaviorally, an ISFET still obeys the Nernst slope, but the zero point is shifted: at pH 7, the mV output is not zero but commonly sits around +100 mV or higher, which means the transmitter must be configured for the ISFET's specific zero/slope pair rather than a glass electrode's defaults [S3]. The chip also integrates an amplifier that gives a low-impedance output, eliminating the high-input-impedance meter requirement of glass and the associated cable-noise problems [S3].
The chip is housed in a polymer (epoxy or PEEK) body, which is the reason ISFET is the default in food, beverage, and life-science hygienic lines where any glass fragment is a regulatory and product-loss event [S2][S3].
Antimony electrode: metal-oxide redox at an antimony surface
The antimony pH electrode is a solid metal probe whose surface carries a thin Sb₂O₃ layer; the equilibrium Sb₂O₃ + 6 H⁺ + 6 e⁻ ⇌ 2 Sb + 3 H₂O generates a potential that depends on pH, with a typical slope around 55-60 mV/pH at 25 °C depending on the source [S1][S6]. The half-cell is mechanically robust, contains no fragile membrane, and tolerates hydrofluoric acid and other species that attack glass [S6][S9].
That mechanical strength comes with two well-documented penalties. Published test data shows the antimony system reaches only about ±0.2 pH accuracy under ideal conditions, roughly 2× worse than a well-conditioned glass electrode at the same temperature, and the effective measurement range is generally narrower than glass, typically reported as roughly pH 1-11 depending on the buffered reference and the antimony source [S1][S9]. The antimony electrode is therefore specified almost exclusively where glass cannot survive, such as HF-bearing pickling baths, slurries with embedded glass fragments, or in-line measurements inside strong reducing media [S6][S9].
For process engineers comparing ruggedness to accuracy, the antimony electrode belongs in the same family of "specialty solids" as refractory glass fiber and optical glass probes: each is chosen because the glass alternative fails the service, not because it beats the glass alternative on metrology.
Selection criteria and side-by-side comparison
Three selection axes separate the three technologies in practice: (1) accuracy, (2) chemical/physical ruggedness, and (3) integration cost. On accuracy, glass still wins at roughly ±0.1 pH under lab conditions and ±0.05 pH with high-impedance reference, with ISFET close behind in hygienic service and antimony trailing [S1][S7][S8]. On ruggedness, ISFET is the clear leader for CIP and impact, antimony is the only one of the three that survives HF and embedded solids, and standard glass is the most fragile [S2][S6][S9].
On integration, glass uses the same KCl-reference architecture and the same high-impedance mV wiring every plant already has; ISFET requires a meter that accepts the ISFET zero/slope pair and the higher current output, or a Memosens-style digital transmitter that handles the conversion in the cable head [S2][S3]. Antimony needs the same reference-electrode handling as glass but generates a slightly lower slope, so two-point buffer calibration must use the antimony-specific buffers specified by the vendor [S1][S9].
For decision-making, the rule of thumb that survives contact with plant practice: specify glass for the widest pH and temperature range, antimony where HF or abrasive solids forbid any glass-oxide contact, and ISFET where CIP, in-line cleaning, or food/pharma hygienic rules forbid any glass fragment at all [S2][S6][S9].
Service envelopes, failure modes, and standards context
General-purpose glass is published for pH 0-14 and up to 100 °C; "blue glass" (low-impedance) extends to roughly 110 °C and pH 0-13; HF-resistant glass variants use a polymer body and a different membrane chemistry to push the fluoride tolerance window, but they trade slope accuracy for that resistance [S7]. ISFET with the dual-layer Ta₂O₅ gate is published for hygienic service, including repeated hot NaOH CIP cycles at 80-85 °C, with stable response down to refrigerated temperatures near 0 °C in food-process lines [S2]. Antimony is documented for roughly pH 1-11 and temperatures limited mainly by the epoxy or polymer housing, with continuous service reported up to about 70-80 °C depending on the manufacturer [S1][S9].
The dominant failure modes are also different. Glass fails by membrane cracking, by junction poisoning of the KCl reference, and by sodium error in high-pH / high-Na samples [S5][S7]. ISFET fails by gate-layer attack under aggressive CIP (the issue the dual Ta₂O₅ layer was developed to suppress), by encapsulation failure of the chip, and by reference-electrode drift [S2][S3]. Antimony fails by surface abrasion exposing fresh metal, by drift of the Sb₂O₃ layer in strongly reducing or strongly oxidising media, and by a more limited Nernstian linearity that compresses accuracy at the pH extremes [S1][S6][S9].
Industry standards applicable to the measurement chain (rather than to the sensor chemistry itself) include ASTM E70 for pH measurement of aqueous solutions, Good Practice Guide on pH measurement from the National Physical Laboratory for traceability, and the metrological language of ISO/IEC 17025 calibration laboratories; sensors themselves are commonly qualified against ATEX/IECEx for hazardous-area use and against 3-A sanitary standards for food and dairy hygienic service [S2]. CIP validation draws on EHEDG Doc. 2 for hygienic cleaning design.
Specifying the right technology per use case
For pharmaceutical water-for-injection (WFI) loops, biotech bioreactors, and dairy HTST lines, ISFET with a hygienic body and a dual-layer Ta₂O₅ gate is the correct call: a glass fragment inside a stainless line is a documented batch-loss event, and the dual-layer gate has been published to survive standard hot NaOH CIP regimes [S2].
For boiler-feed water, condensate, cooling-tower make-up, and general chemical-plant pH, the standard glass combination electrode with refillable KCl reference is still the metrological workhorse and remains the cheapest path to ±0.1 pH accuracy [S4][S7]. For semiconductor wet-etch, HF pickling, and abrasive slurry services, antimony is the only viable technology of the three, with the explicit accuracy trade-off accepted at the specification stage [S1][S6][S9].
Trackable signals to watch: published ISFET long-term drift rates under modern Ta₂O₅ dual-layer gates, antimony electrode CIP survivability data from vendors that now offer polymer-bodied Sb probes for hygienic service, and the spread of Memosens-style digital ISFET transmitters that allow hot-swap calibration off the loop. For engineers also weighing glass tube vs metal tube variable-area flowmeters, the same CIP-and-hygiene logic applies: specify the wetted material for survivability first, then accept the accuracy and price tier that the chemistry forces on you.