A representative enquiry: sulphuric acid 92–98 wt %, 20–90 °C, DN50, no bubbles or solids, ±0.03 wt % required.
The instinctive reaction is that the accuracy demand is too tight. Run the numbers and it turns out not to be the issue.
Resolution is not remotely the constraint
Between 80 % and 100 %, the sound-velocity sensitivity of sulphuric acid is roughly 10–20 m/s per wt %.
Resolving 0.03 wt % therefore requires 0.3–0.6 m/s of velocity resolution. Instruments of this class resolve velocity to the order of 0.01 m/s — more than an order of magnitude of headroom. Velocity resolution is not the bottleneck.
The real bottleneck: a 70 K temperature span
20 °C to 90 °C is 70 K. The velocity change that span produces is of the same order as the change across the entire 92–98 % concentration range.
Which means: whatever residual your temperature compensation leaves is what your final accuracy will be. Concentration resolution cannot rescue it.
So the correct response to this kind of enquiry is not to quote an accuracy figure, but to ask for the real concentration × temperature operating window at each measurement point. A local fit within a narrow window vastly outperforms a global fit. A point that lives permanently at 92–94 % and 55–65 °C and a point that genuinely roams the full envelope support completely different commitments.
One more trap: the calibrated range must not cross the inflection
The velocity–concentration curve for sulphuric acid has an inflection point, located below about 90 %. The 92–98 % band sits on the monotonic branch above it, which is why the band is usable — but the calibrated range must never extend down across the inflection, or one velocity maps to two concentrations and the reading becomes irreducibly ambiguous.
This is why sound velocity is the generally recommended method for sulphuric acid in the 60–100 % band, while more dilute ranges need separate consideration.
Materials: the controlling case is 92 % at 90 °C, not nominal 98 %
This is the most commonly mis-specified point in the whole selection.
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 92 % + 90 °C worst-case combination, not for the nominal 98 %.
Baseline is Hastelloy C-2000 or C-276. Tantalum cladding is an upgrade option, but cannot be used with oleum or free SO₃. Zirconium 702 only reaches about 70 %; titanium, Ni200 and Monel are ruled out at these conditions.
How to define acceptance
A concentration specification needs an executable acceptance method or it will be argued about forever. Ours: the standard deviation of "meter reading − titration" across paired samples, written into the calibration certificate.
One further note: displayed decimal places are not accuracy. Some products state ±0.05 wt % while displaying three decimals — the last digit is already below the instrument's own uncertainty and merely looks finer.
