PISONICS
PS7020 Series
Ultrasonic Concentration Meter
Sound velocity · inline spool · no moving parts
Chlorine Tail-Gas Caustic Scrubber — Online Residual-Alkali Monitoring
—— Residual alkali and available chlorine are one number said two ways ——
【Key measurement: residual NaOH wt% / available chlorine wt%】
The Duty
Chlorine, or chlorine-bearing tail gas, enters the tower and is absorbed by circulating caustic. There is only one reaction:
Cl₂ + 2 NaOH → NaOCl + NaCl + H₂O
Fresh caustic is typically made up at about 20 wt%. The loop circulates until residual alkali is down to around 5 wt%, then the batch is switched out to the hypochlorite tank. At that moment the liquor carries three solutes: unreacted NaOH, the product NaOCl, and by-product NaCl.
Plants state the switch-out criterion either way. Some hold a residual alkali figure; others go by available chlorine — the Chinese chlor-alkali emission-standard drafting note, describing the twin caustic towers that handle waste and emergency chlorine, puts it as: circulate the sump liquor back over the tower, and when the available chlorine as NaClO reaches 10 wt%, pump it to the hypochlorite storage.
These are not two criteria. Run the balanced equation and available chlorine of 10.0 wt% lands at exactly 6.7 wt% residual alkali. Same point on the same reaction path — counted from the product side instead of the reactant side. Everything below follows from that.
One Scalar Is Enough
The three solutes are not free to move independently. Every 2 mol of NaOH consumed yields exactly 1 mol of NaOCl and 1 mol of NaCl — stoichiometry locks them onto a single reaction path. The composition of the loop is therefore a one-parameter family: fix how far the reaction has gone and all three concentrations are fixed with it. One scalar locates you.
Starting from 100 g of 20 wt% caustic, the path runs like this:
| Extent | Residual NaOH | NaOCl | NaCl | Available Cl | Density | Sound velocity (est.) |
|---|---|---|---|---|---|---|
| Fresh | 20.00 % | 0.00 % | 0.00 % | 0.00 % | 1.219 g/cm³ | 2097 m/s |
| 1/4 | 15.45 % | 3.59 % | 2.82 % | 3.42 % | 1.213 g/cm³ | 2020 m/s |
| 1/2 | 11.21 % | 6.95 % | 5.46 % | 6.62 % | 1.207 g/cm³ | 1949 m/s |
| 3/4 | 7.23 % | 10.09 % | 7.92 % | 9.61 % | 1.201 g/cm³ | 1882 m/s |
| Switch-out | 5.00 % | 11.85 % | 9.31 % | 11.29 % | 1.198 g/cm³ | 1844 m/s |
The composition columns are exact, straight from the equation. Density and sound velocity are first-order estimates built from handbook per-wt% increments; they are here to compare the two routes by order of magnitude and are not calibration values.
Reading Our Own Exclusion List
The PS7020 product page excludes “ternary and higher multi-component systems (sonic velocity alone cannot uniquely determine concentration”). The operative words are uniquely determine. What is excluded is a system whose components vary independently. A ternary system pinned to one dimension by stoichiometry still maps sound velocity to concentration one-to-one.
| Both are “NaOH + NaOCl + NaCl” | Degrees of freedom | Independent quantities needed | Selection |
|---|---|---|---|
| Hypochlorite production tower (making a product) chlorine and caustic dosed independently; the product is accepted on available chlorine and free alkali | Two | Two | PS7100, spectroscopic |
| Tail-gas absorption tower (holding a safety margin) the chlorine load is whatever the process vents; the loop simply runs down one reaction path | One | One | PS7020, sound velocity |
So this page does not undercut our available-chlorine + free-alkali page. That duty needs two numbers you can accept a product against. This one needs a single number telling you how much absorbing capacity is left.
Why Density Goes Blind Here
On the same path the two measurement routes are nowhere near equally sensitive. Alkali leaving the solution lowers the density; the NaOCl and NaCl forming in its place push it back up, and the two nearly cancel. Sound velocity responds far more strongly to the alkali than to the two products, so nothing cancels.
| Change over the whole run | Per 1 point of residual alkali | As residual-alkali uncertainty | |
|---|---|---|---|
| Density | only 0.021 g/cm³ | 0.0014 g/cm³ | at the 1 ‰ class, ±0.001 g/cm³ → ±0.73 points |
| Sound velocity | about 253 m/s | about 16.9 m/s | at ±1 m/s stability → ±0.06 points |
Roughly an order of magnitude apart. A density instrument — Coriolis, tuning fork — cannot see this loop change. Not for want of accuracy: the measured quantity itself barely moves. Choosing the PS7020 here is therefore not settling for the cheap option; sound velocity is the right method for this duty. It is also why this page does not recommend PS7200 or PS7400.
Two Measuring Points, Two Meters, Different Metallurgy
| Point | Medium | System | Wetted material | Calibrated span |
|---|---|---|---|---|
| ① Make-up side caustic feed line, before mixing | NaOH + water | Binary | 316L stainless | narrow, 15–20 % (this is the point the regulation names) |
| ② Circulating liquor pump discharge, upstream of the make-up tee | NaOH + NaOCl + NaCl | One-dimensional path | Ti-Pd / duplex / C-276 see below | residual alkali 20 % → 5 % or equivalently available chlorine 0 → 11 % |
Do not order the circulating meter as standard build. Standard PS7020 wetted parts are 316L. At switch-out the liquor is about 12 % sodium hypochlorite over about 9 % sodium chloride — an oxidising, chloride-bearing medium in which 316L is simply out. On that there is no disagreement. The make-up side is plain caustic and 316L is fine there.
That leaves the three chloride-stress-corrosion-resistant families AQ 3062—2025 names in A.2.11: titanium-palladium alloy, duplex stainless, Hastelloy (or CPVC / PVDF verified against chloride permeation). Enquiries specifying C-276 for this duty are common, and C-276 is one of them. We do not decide for you — the rest of the stream decides: residual free chlorine, pH (hypochlorite gets markedly more aggressive on the acid side), chloride level, temperature, and whether the design has crevices. One trap to avoid: A.2.10 of the same standard prohibits titanium and titanium alloys in contact with dry chlorine, so discussing titanium on the wet-chlorine and caustic side does not make it usable everywhere on the unit. Send us the conditions, the material is confirmed against the duty and written into the order drawing — which is what the product page means by “special media, material confirmed against the duty”.
If the budget stretches to one meter only, fit ②. Make-up strength is usually a purchased specification — 20 % caustic off a tanker, verifiable on receipt. The margin left in the circulating liquor has no second source. For the make-up side on its own see also our caustic storage tank page.
Not Optional: What GB 11984—2024 Requires
GB 11984—2024, the technical code for chlorine safety in chemical enterprises, was issued 2024-11-28 and came into force 2025-08-01, replacing GB 11984—2008. It is a mandatory national standard and applies to chemical enterprises that produce or use chlorine (the standard notes that chlorine includes liquid chlorine). Online monitoring on this system is not an engineering option. The clauses read as follows.
| Clause | As written | What it means for this page |
|---|---|---|
| 6.1.3 d) emergency chlorine absorption, liquid chlorine storage | Sodium hydroxide in the circulating absorption liquor shall be 15–20 %, and its temperature leaving the tower not more than 45 °C. Facilities for online monitoring of the circulating liquor's sodium hydroxide concentration and temperature shall be provided; periodic laboratory analysis shall be carried out; and the liquor shall be replaced or renewed before the sodium hydroxide falls below 5 %. | This page is about that online monitoring facility. Starting at 20 % and changing out at 5 % is the standard's own framing, not ours |
| 5.2.3 electrolysis | On abnormal operation at full load or emergency shutdown, the electrolysis unit's emergency chlorine absorption system shall absorb for not less than 30 min; sodium hydroxide concentration shall be monitored online; a chlorine detector shall be fitted at the vent. | The electrolysis side is under the same online-monitoring mandate |
| 5.2.3 unit separation | The electrolysis unit's emergency chlorine absorption system and the chlorine compressor's chlorine-bearing seal-gas absorption system shall not be combined with the sodium hypochlorite production unit. | The standard itself separates the emergency absorber from hypochlorite production — which is where this page's one-dimension / two-dimension distinction comes from, not something we invented to sell the cheaper instrument |
| 6.1.3 a) c) g) | Independently installed, using a two-stage absorption process; capable of 24 h continuous operation, with the caustic circulation tanks able to be switched, held in standby and made up; circulation pumps and emergency chlorine fans provided with spares, on the most critical class of primary electrical load. | The system is designed to be available at all times; the instrumentation has to match that availability |
| 5.1.1 a) b) | Liquid chlorine systems shall have nitrogen trichloride discharge and treatment facilities; discharge frequency is set by keeping NCl₃ in the discharge below 0.5 % by mass; sodium hydroxide in the treatment facility shall be 5–20 %. | Another caustic measuring point on the same site, over a similar span |
Three of its requirements land exactly on what this page already recommends:
① online monitoring of concentration and temperature — temperature is not an optional compensation input but the second quantity the clause requires in its own right. This page says temperature compensation is not optional; the standard says it harder.
② “periodic laboratory analysis” — the standard requires laboratory testing in addition to the online instrument. That is the shift titration this page argues for. Not conservatism on our part: a requirement.
③ “before the sodium hydroxide falls below 5 %” — before, not at. The standard itself asks for margin. The nominal 8 % alarm this page recommends is what that word costs.
The boundary is worth stating precisely, without stretching it. Clause 6.1.3 governs the emergency chlorine absorption system for liquid chlorine storage; 5.2.3 governs the one on the electrolysis unit. A purely process tail-gas tower that absorbs only normal process vent gas and carries no emergency duty does not fall directly under either. In most fine-chemical plants, though, the tower does both — clause 4.10 requires the chlorine capture systems serving liquid chlorine storage buildings, cylinder stores, filling areas and vaporiser rooms to be connected to the emergency absorption system. Where it does both, 6.1.3 applies. Where it does not, the mandate does not bite but the engineering is the same, and the margin and re-anchoring below apply just as much.
Building the Scale
- Do not use a handbook binary caustic table. Handbooks give NaOH in water; this loop runs a reaction path, and at the same residual-alkali figure the two curves sit at different sound velocities.
- Build it from the plant's own titrations. Sample from the same spool as the probe, at the same moment, and log four things together: time, titrated residual alkali, sound velocity, temperature.
- Cover the whole path. Sample from fresh caustic all the way to switch-out — at least 5 concentration points across 3 temperatures, which is also what the PS7020 accuracy-class table asks for at the 0.5 ‰ class.
- If the gas carries HCl, calibrate on the real mixture. HCl also eats alkali but makes no hypochlorite, so it runs a path of different slope (see below). Chlorination tail gas in fine chemicals is commonly Cl₂ + HCl; a stable ratio can simply be calibrated in, an unstable one has to be covered by margin.
- Bring the raw sound velocity out on its own loop. The PS7020 can map sound velocity to either 4–20 mA output independently. Put that on the DCS and compare it directly against each titration, rather than reconstructing it from the concentration curve afterwards.
Three Ways It Drifts Off the Path
The scale holds while the loop stays on that reaction path. Three things take it off — and this has to be said first: all three err in the same direction.
| Mechanism | How it happens | Magnitude (estimated) | Direction |
|---|---|---|---|
| Water loss | absorption is exothermic and the gas flow carries vapour out; less water concentrates all three solutes together | 5 % water lost → residual alkali reads about 0.8 points high | Unsafe way |
| CO₂ ingress | the fan draws in air; 2 NaOH + CO₂ → Na₂CO₃ + H₂O consumes alkali without making hypochlorite | for each point of alkali eaten the meter reports only about 0.2 — roughly four-fifths under-reported | Unsafe way |
| HCl in the gas | HCl + NaOH → NaCl + H₂O again eats alkali with no hypochlorite; the path slope is only about 82 % of the pure-Cl₂ one | calibrated on pure Cl₂ but running a mixture: up to 2.6 points over the full run | Unsafe way |
All three show up as “more alkali on the display than in the tower” — it looks like there is absorbing capacity left when the loop is close to breakthrough. A scrubber is a safety device and this error is not symmetrical: reading alkali low wastes a little caustic, reading it high puts chlorine out of the stack. What follows is not conservatism, it is the operating envelope.
How to Operate It
- Alarm with margin — set switch-out at a nominal 8 % residual alkali, not 5 %. The caustic that costs is far cheaper than one breakthrough, and cheaper than the difference to a PS7100.
- Keep the shift titration — the instrument's job here is continuous trend and early warning, not absolute assay. One titration per shift re-anchors the scale and catches all three drift mechanisms.
- Trend the raw sound velocity alongside — when the reading jumps, look at sound velocity first. A jump is usually gas or water ingress, not a real concentration change.
- Temperature compensation is not optional — the product page puts it plainly: sound velocity often varies more with temperature than with concentration. On this loop 1 °C uncompensated is worth roughly 0.1 point of residual alkali. Put the temperature element close to the sound path.
- Re-zero after a changeover — a new batch, a large water addition, or a restart after a long idle period all put you back at the start of the path. Get the start wrong and the whole scale shifts with it.
Installation
- Full-bore vertical run on the circulating pump's discharge. Not the tower return line, not the pump suction — “aerated duties, where bubbles make the sonic velocity fall sharply and the reading jump” is on the PS7020's own exclusion list, and an absorber return line is precisely where entrained gas lives.
- Keep ② upstream of the make-up tee. ② has to see the circulating liquor itself; blended with fresh caustic it no longer reports the loop's real margin.
- Sample valve on the same spool as the probe. The scale is built from titrations, and the further the sample point is from the probe the looser that scale gets.
- Temperature element close to the sound path, for the reason above.
- Allow for long idle periods on an emergency-duty tower. A tower that sits still slowly absorbs CO₂; re-anchor against a titration before putting it back in service.
Against a Real Enquiry
The table below runs line by line against an enquiry as it actually arrived. Where we match, it says so; the three lines where we do not are spelled out underneath — a gap hidden at enquiry stage only becomes a more expensive argument at acceptance.
| As specified | PS7020 | |
|---|---|---|
| Medium: NaOH + sodium hypochlorite solution | The duty this whole page is about | ✓ |
| Temperature 0 – 80 °C | Standard −20 to +80 °C; high-temperature build to +120 °C | ✓ |
| Working pressure PN16 | ≤ 2.0 MPa | ✓ |
| Flanged connection, DN50 | Spool type, standard bores DN25 / DN50 | ✓ |
| Ex d IIB T6 Gb | Ex d IIC T6 Gb (optional) — IIC covers IIB; the higher group is not a problem | ✓ |
| 4–20 mA output, 24 V supply | 4–20 mA × 2, NAMUR NE43; DC24V is one of the standard supply options | ✓ |
| Wetted material Hastelloy C-276, FFKM seals | Special media, material confirmed against the duty — see the section above; seals selected with it | To order |
| Measuring span 0 – 50 % | Determined by real-liquid calibration. But a wider span means a lower accuracy class we can commit to — the 0.5 ‰ class requires a span of ≤ 10 percentage points, so 0–50 % can only be discussed at the 5 ‰ class | ⚠ |
| Ingress protection IP66 | PS7020 is IP65 | ✗ |
| Ultrasonic 24 V automatic cleaning | Not on the standard build | ✗ |
| Named an ultrasonic dual-parameter concentration meter | PS7020 is single-parameter (sound velocity) | ✗ |
IP66 against IP65 — one step apart; IP66 asks for resistance to powerful water jets. Outdoors on a tower deck, with frequent washdown, that is a reasonable thing to ask for. Either the protection is raised at order, or the difference is stated plainly and the user judges it. It does not get quietly passed over.
“Ultrasonic automatic cleaning” — this line is usually copied from a refractometer or spectrometer template. Those instruments have an optical window, fouling changes the reading directly, and so they are fitted with ultrasonic or high-pressure cleaning. Sound velocity takes the arrival time of the signal rather than its amplitude; the product page puts it as “insensitive to amplitude-type degradation such as face fouling, couplant ageing and gain drift”. This meter does not rely on cleaning to hold its reading. If the real worry is crystalline build-up blocking the line, the conversation to have is about spool geometry and velocity, not about adding a cleaner.
“Dual-parameter” — those two words decide the whole solution. It is not a category with one agreed definition; behind it lie at least three different technical routes serving three different needs. If what you want is a single progress number — how much absorbing capacity is left — the route on this page is the cheapest that works. If the outlet liquor is sold as product and has to be held to available chlorine and free alkali as separate acceptance figures, that is a two-dimensional problem and this page's method cannot do it. That question has a page of its own: what an “ultrasonic dual-parameter concentration meter” actually means — three routes and how to choose, including what we do and do not have.
Specification
| Item | Rating |
|---|---|
| Principle | Ultrasonic sound velocity (transit time ÷ fixed path) |
| Valid sonic velocity range | 400 – 3500 m/s; outside it the fault bit is set |
| Sonic velocity resolution | 0.01 m/s |
| Primary outputs | Mass concentration, volume concentration, density, Baumé, solids content; selectable units |
| Diagnostic outputs | Sound velocity, probe temperature, received signal energy, signal quality, fault status word |
| Accuracy | 0.5 ‰ / 1 ‰ / 5 ‰ classes, i.e. density ±0.0005 / ±0.001 / ±0.005 g/cm³; the class is a function of calibration conditions |
| Process temperature | Standard −20 to +80 °C; high-temperature build −20 to +120 °C at order. Absorption is exothermic and enquiries for the circulating loop commonly state 0–80 °C; select against the real duty |
| Working pressure | ≤ 2.0 MPa |
| Wetted materials | 316L stainless; titanium on the circulating side for this duty |
| Standard bores | Spool type DN25 / DN50; other bores confirmed at order |
| Analogue outputs | 4–20 mA × 2, NAMUR NE43; each independently configurable |
| Digital comms | RS485 Modbus-RTU; HART 5 slave |
| Ingress / Ex | IP65; Ex d IIC T6 Gb (optional, specified at order) |
Full parameters per the PS7020 datasheet; see also the PS7020 product page.
Before We Quote
- Gas composition: pure Cl₂, or with HCl / CO₂ / organics? Is the ratio stable?
- Make-up caustic strength, and whether it is fixed (20 % delivered? 32 % diluted? recovered caustic?)
- Is this tower continuously loaded or on emergency standby (a standby tower sits idle, which makes the CO₂ problem worse)
- Is there circulation cooling, and how far does temperature swing?
- Is switch-out currently called on residual alkali or available chlorine, and at what threshold?
- Loop bore, material, pressure — and whether a vertical full-bore run exists on the pump discharge
- Any history of “time + titrated residual alkali + temperature” — with that we can work out whether the scale is feasible before anyone quotes
What this page is: a method and selection note, not a case study. The PS7020 has no publicly citable installed base on this duty, so there are no unit counts, saving percentages or payback periods here — we do not invent those. The composition table is computed from the reaction equation; the sound-velocity and density figures are first-order estimates used to compare orders of magnitude. The real scale can only be built from your own loop's titrations. Send us the process conditions and you get a selection opinion you can check.
