PISONICS
Selection note
Emergency Chlorine Absorbers
Five methods, judged against this one duty
How to Measure Caustic Strength in an Emergency Chlorine Absorber
—— The standard wants 15–20 % monitored online, and the method most often proposed fails in exactly that window ——
【Key measurement: sodium hydroxide wt%】
State the Problem First
GB 11984—2024 clause 6.1.3 d) requires the circulating absorption liquor to hold 15–20 % sodium hydroxide, with online monitoring facilities for concentration and temperature, periodic laboratory analysis, and replacement before the caustic falls below 5 %.
So the instrument has three jobs:
- Hold the 15–20 % window and raise an alarm below the lower limit;
- stay useful all the way down to 5 %, because that is where the change-out criterion sits — the span cannot cover only the compliance window;
- measure temperature too, and specifically the temperature of the stream leaving the tower.
Note the second one. The real measuring range is 5–20 %, not 15–20 %. A method that is accurate inside the compliance window but goes deaf at 10 % falls silent exactly when you most need it to speak.
Five Ways to Do It
| Method | How it behaves on this duty | Verdict |
|---|---|---|
| Manual sampling and titration | Absolutely accurate, the referee method, and required by the clause itself. The cost is frequency and exposure: the sample point is on a chlorine-bearing system, so each sample means operating a valve in protective equipment, and the result arrives hours later — whatever happened in between is unrecorded | Keep it but it is not online |
| pH | 5 % caustic is about 1.3 mol/L and 15 % about 4.3 mol/L — a nominal difference of roughly 0.5 pH units across the whole span, while a glass electrode already suffers alkaline error above pH 12. Neither the resolution nor the accuracy is there | Not usable |
| Conductivity | The method most often proposed, and the one this page is really about — its turning point lands inside the window the standard requires. See the next section | Unreliable in this window |
| Density / Baumé | Density of caustic soda rises monotonically with concentration (about 1.010 g/cm³ at 1 % and 1.430 at 40 %, 20 °C), so the direction is sound and hydrometers have always been used in the field. An online densitometer can do it, but temperature compensation is demanding, and density barely moves when a standby tower's caustic carbonates | Workable watch the temperature compensation |
| Ultrasonic sound velocity | Calibrated on real liquid, with monotonicity established by that calibration on your own loop. No moving parts, nothing to consume, no optical window, and the raw sound velocity can be brought out on its own loop for cross-checking | Recommended PS7020 |
| ORP | The Shandong standard DB37/T 1933—2022 suggests an online ORP instrument on the absorber's circulating liquor. But ORP reads the oxidation state of the liquid — whether it has absorbed chlorine and how much available chlorine is present — not free alkali. It cannot stand in for the quantity the clause names | Valuable but a different quantity |
The Conductivity Problem
Conductivity is the first method most people think of for caustic strength: cheap, mature, and available in a contactless electrodeless form. Applied to sodium hydroxide it runs into one property that cannot be designed around — conductivity rises with concentration, then falls, and the turning region sits in the 15–20 % band. Three independent lines of evidence:
- Rosemount / ASTI, “Conductance Data for Commonly Used Chemicals” (51-6009): for NaOH, maximum conductance 410,000 µmho/cm, point of inflection 15 % by weight at 25 °C.
- A separate engineering reference table gives NaOH at 25 °C as 78,300 µS/cm at 5 %, 140,000 at 10 %, 226,000 at 20 %, and marks 30 % as saturated. A different data series, but it too stops behaving like a simple rising curve past 20 %.
- The plainest evidence comes from a conductivity vendor: a Chinese electrodeless caustic concentration meter is offered in two ranges, 0–10 % and 25–40 % — 10–25 % is simply not sold. The window the standard asks you to hold is 15–20 %.
To be honest about it: the literature does not agree on where exactly the maximum falls. Some sources put it at 15 %, others draw the curve still rising at 20 %. That disagreement is itself the argument — the people who make conductivity instruments cannot pin the turning point, and it lies in the band you have to hold. Near a turning point a single conductivity reading does not determine concentration uniquely: the same value may sit inside the window or already below the lower limit. Temperature moves the turning point as well, and the liquor temperature in this tower is not constant.
None of which makes conductivity useless. On dilute caustic — say 0–8 % — it works well and responds quickly. It simply happens that the band GB 11984 asks you to hold is the band it handles worst. If there is already a conductivity instrument on site, keep it for indication and rate-of-change, but do not lean on it as the compliance measurement.
Why Sound Velocity, Then
First, what we are not claiming: we do not assert that sound velocity has been proven strictly monotonic across the whole range. Published sound-velocity data for caustic soda is not easy to come by, and selling on a claim whose source you cannot produce is precisely what this page is trying to help you avoid.
What the argument does rest on is checkable on your own site:
- The scale is built from your own loop's liquid. Monotonicity is not our promise but the calibration's demonstration — five concentration points across three temperatures show directly whether the curve is monotonic and whether the sensitivity is there. If it is not, we will say so.
- The raw quantity can be brought out separately. The PS7020 maps sound velocity to either 4–20 mA output independently. On the DCS, every titration can be compared against it — an instrument that exposes its own raw measurement for checking is worth more than one that only hands you a conclusion.
- There is no maintained wetted interface. No optical window, no reference electrode, no membrane. On a unit that sits in standby all year and may go untouched for months, that matters more than accuracy class.
- Temperature compensation is mandatory — our product page says so plainly: sound velocity often varies more with temperature than with concentration. The standard happens to require temperature as well, so one instrument returns both quantities.
What We Recommend
| Role | What | What it covers |
|---|---|---|
| Primary online | PS7020 sound velocity (concentration + temperature) | Continuous trend, step-change alarms, holding the 15 % lower limit, and visibility all the way to the 5 % change-out point |
| Absolute reference | Manual sampling and titration (required by the standard) | The true free-alkali figure, and carbonation in particular — the path an online instrument cannot see clearly |
| Optional supplement | ORP (suggested by the provincial standard) | Whether this tower has absorbed chlorine and the liquor's oxidation state. A different quantity from caustic strength; neither replaces the other |
Why the answer is not simply “buy the expensive one and forget it”. The clause requires online monitoring facilities and periodic laboratory analysis, in the same sentence. No solution escapes the second half. So the right question is not which instrument replaces the laboratory, but which instrument stops the interval between analyses from being a blind spot.
Before We Quote
- How is caustic strength judged today — titration, hydrometer, conductivity, or a fixed change-out interval?
- How often is a sample taken, by whom, and from where — that decides where the online instrument belongs.
- Is there a vertical full-bore run on the circulation pump discharge, and at what bore?
- Liquor temperature range; is there a caustic cooler?
- Is the circulation tank open to atmosphere? That sets the carbonation rate, and therefore how often the analysis has to be done.
- Hazardous-area classification and electrical supply class.
- Any historical laboratory results (concentration + sampling time + liquor temperature) — with those we can work out whether the scale is feasible before quoting.
What this page is: a method and selection note, not a case study. There are no unit counts, saving percentages or payback periods here — we do not invent those. Where the page says what the industry does, it cites something you can check for yourself (published vendor literature, granted patents) rather than implying we have installed base there. Send us the process conditions and you get a selection opinion you can check.
