Revision note (October 2026): when first published on 9 September 2026, this article quoted published material placing the turning point of the sulphuric acid velocity curve below about 90 %. Our own titration-referenced calibration data does not support that — the insensitive region is 35–45 %. The article has been revised to the PS7020 Sulfuric Acid Application Note Rev. B; what follows is the revised text.
The instinctive reaction to inline measurement of concentrated sulphuric acid is that the accuracy demand is too tight. In fact velocity resolution is not the bottleneck; what decides the delivered accuracy is temperature compensation.
Resolution is not the constraint
Between 85 % and 99 %, the sound velocity of sulphuric acid is single-valued and steep, and instruments of this class resolve velocity to the order of 0.01 m/s. Calibrated against titrated sulphuric acid standards, the PS7020 achieves the following measured accuracy by band:
| Concentration band | Measured accuracy (wt%) | Remarks |
|---|---|---|
| 85–99 % | ±0.05–0.1 | Best-performing region: the velocity curve is single-valued and steep |
| 80–85 % | ±0.1 | Single-valued; slope still strong |
| 75–80 % | ±0.3 | Slope begins to flatten |
| 45–75 % | ±1 | Usable for trend and control, not for custody-grade concentration |
| 35–45 % | Not measurable | Turning point of the velocity curve: no sensitivity to concentration, double-valued across it; a calibrated span must not cross this band |
| 15–35 % | ±0.4 | Single-valued on the dilute branch |
| 5–15 % | ±1 |
The concentrated-region calibration covers 20–95 °C.
The real constraint: temperature
The sound velocity of concentrated sulphuric acid is strongly temperature-dependent, with a negative coefficient of a few m/s per kelvin. Across a few wt% in the concentrated region, the velocity changes by an amount of the same order as across a 70 K temperature swing — so whatever residual the temperature compensation leaves is what the final accuracy will be.
The right response to a concentrated-acid enquiry is therefore not to quote an accuracy figure first, but to ask for the real concentration × temperature operating window at each point. A point usually runs in a narrow window — for example 97.5–98.5 % at 75–85 °C in an absorption circuit — and a local fit within that window vastly outperforms a global one. The temperature element belongs in the acoustic path, thermally coupled to the wetted body, not in a thermowell elsewhere in the line.
One band to stay out of: 35–45 %
The sulphuric acid velocity curve turns between 35 % and 45 %: there it has no sensitivity to concentration and it is double-valued across the band. So a calibrated span must not cross 35–45 %. 92–98 % sits in the best-performing band; if start-up, shutdown or flushing takes the process through 35–45 %, it needs a second configuration or a second physical quantity to disambiguate.
Some published material puts the turning point below 90 %, and this article quoted it when first published. Our measured result stands; that statement is withdrawn.
Materials: the lowest concentration at the highest temperature, not nominal 98 %
Corrosion of metals by sulphuric acid is at its minimum around 98–99 %, and worsens sharply as concentration falls and temperature rises. Materials must therefore be specified for the lowest concentration this point can see at the highest temperature it can see, not for the nominal 98 %.
- Hastelloy C-2000 / C-276: baseline recommendation
- Tantalum (clad): upgrade option, excellent across the whole range to roughly 150 °C; not with oleum, free SO₃ or fluoride
- Alloy 20: acceptable; check acid velocity
- 316L: concentrated acid near ambient only; not recommended above roughly 40–50 °C
- Titanium, Nickel 200, Monel 400: not recommended in oxidizing concentrated acid; Zirconium 702: not above approximately 70 %
How to define acceptance
A concentration specification needs an executable acceptance method or it will be argued about forever. Ours: within an agreed operating window, take an agreed number of paired samples and compute the standard deviation of meter reading minus titration, written into the calibration certificate.
One further note: displayed decimal places are not accuracy. A product stated at ±0.05 wt% that displays three decimals shows a last digit already below its own uncertainty — it merely looks finer.
The full selection notes are on the application page Inline concentration of 92–98 % sulfuric acid.
