ICPT 2026 · Seoul · 21 October

Three things change sound speed in CMP slurry. Only one of them is H2O2.

Our ICPT 2026 paper shows how the Rhosonics Chemical Concentration Meter separates them: inline, up to ten times a second, within ≈ ±0.03 wt% of titration from 20 to 28 °C.

Talk to a specialist Read the paper

Feasibility study presented at ICPT 2026, from the paper Inline Real-Time Monitoring of Hydrogen Peroxide Concentration in CMP Slurries Using Ultrasonic Technology for Advanced Process Control by Raymond Maas and Stefan Seuleanu, Rhosonics.

≈ ±0.03wt%Accuracy against titration, three titrated samples, 20 to 28 °C
≈ ±0.015wt%Spread at 0.9618 wt%, with temperature held within ±0.5 °C
10readings/sUp to 10 per second on a 4-20 mA output, versus one titration value per sample
21 OCTICPT 2026Poster session in Seoul, Wednesday 13:10 to 15:10
01 · The signal

H2O2 changes the speed of sound. That makes it measurable inline.

Titration gives you one sample, measured in a lab, at one temperature. Sound speed can be read in the slurry loop, continuously. The catch: H2O2 isn’t the only thing that moves it.

 TitrationRhosonics CCM
Where it measuresA sample, taken to the labInline, in the recirculation loop
How oftenWhen someone takes a sampleUp to 10 times a second
TempératureThe lab’s, not the loop’sMeasured and compensated live
When you knowAfter sampling and analysisImmediately, in your control system

Titration still matters. It provides the reference samples that calibrate the CCM, and it checks it.

02 · Trap one: temperature · Poster panel 03

Temperature moves sound speed too, and not by the same amount at every concentration.

Sound speed vs temperature · paper fig 5 Speed of sound against temperature at four H2O2 concentrationsSound speed changes with temperature at every concentration, by a different amount at each. Redrawn from Figure 5 of the paper.m/s142014601500154001020304050 °C0.05 wt%1 wt%2 wt%3 wt%
Speed of sound in m/s at four H2O2 concentrations. Redrawn from Figure 5 of the paper.

Sound speed shifts strongly with temperature, and the size of that shift depends on how much H2O2 is in the slurry. A single temperature correction, tuned for one concentration, is wrong for all the others.

How the CCM handles it

Its model lets the temperature correction change with the H2O2 level, so the reading holds as your loop warms and cools.

03 · Trap two: an ageing slurry · Poster panel 04

As slurry ages, sound speed drifts even when H2O2 doesn’t.

Slurry ageing · paper fig 6 Figure 6 of the paper, redrawnOver 48 hours the uncompensated H2O2 reading rises from about 1.039 to about 1.067 wt%. With conductivity compensation it stays close to the titration value of 1.051 wt%.wt%1.041.051.061.070 h12243648 huncompensated: ≈ +0.028 wt% in 48 hcompensated: stays at titration target
H2O2 in wt% over 48 hours. Dashed line: titration value, 1.051 wt%. Redrawn from Figure 6 of the paper.

We circulated a copper CMP slurry at about 1 wt% H2O2 for more than two months to age it. Its base properties changed, and a sound-speed reading drifted with them. Possible causes:

  • Water lost to evaporation
  • H2O2 decomposing
  • Changes in slurry stability
How the CCM handles it

Conductivity follows the base slurry, not the H2O2. Adding it to the model separates the two. Uncompensated drift: ≈ +0.028 wt% in 48 hours. Compensated: on target.

04 · The result · Poster panel 01

With both traps handled, the reading stays within ≈ ±0.03 wt% of titration.

Reading minus titration vs loop temperature · paper figs 10 to 12 CCM reading minus titration across the loop temperature rangeFor titrated samples of 0.9618, 1.9780 and 2.8552 wt% H2O2, the CCM reading stays within plus or minus 0.03 wt% of titration from about 20 to 28 degrees C. Redrawn from Figures 10 to 12 of the paper.0.9618±0.03 wt%1.97802.85522022242628 °C
Line: running median of CCM reading minus titration, in wt%. Grey band: ±0.03 wt%. Redrawn from Figures 10 to 12 of the paper.

One copper CMP slurry, three titrated H2O2 levels, loop temperature swept from 20 to 28 °C.

  • 0.9618 wt%, 20 to 28 °C≈ ±0.03
  • 1.9780 wt%, 20 to 28 °C≈ ±0.03
  • 2.8552 wt%, 20 to 26 °C≈ ±0.02
  • Spread at steady temperature≈ ±0.015
  • Low concentrations told apart0.04 · 0.07
See the full data in Seoul

ICPT 2026 poster session. Wednesday 21 October, 13:10 to 15:10.

Add to calendarRead the paper
05 · Your slurry

The method works. Next, we calibrate it to your slurry.

CMP slurries differ in abrasives, additives and oxidiser levels, so the CCM’s model is modular. Its coefficients are fitted to your formulation from reference samples at several H2O2 levels and temperatures.

Four measurements, taken at one point in the line

UltrasoundSound speed

The H2O2 signal.

Ultrasound echoDensité

From the echo at the sensor face, with sound speed.

InductiveConductivity

Follows the base slurry, to separate ageing from H2O2.

SensorTempérature

At the same point, with a correction that changes with concentration.

One model, fitted to your slurryH2O2 wt% + density · 4-20 mA
  1. 01 · Install

    Inline, at the mixing tank

    In the recirculation loop, before the slurry reaches the polishers.

  2. 02 · Calibrate

    To your slurry

    With titrated reference samples at several H2O2 levels and temperatures.

  3. 03 · Monitor

    Continuously

    On the display and in your control system, compensated for temperature and slurry ageing.

06 · Experience

CMP slurry is new for us. Fab chemistry isn’t.

Rhosonics concentration meters have measured developers, etchants and process chemicals inline in display and chip plants since 2009. The CMP work builds on that.

1992Founded in the Netherlands, specialists in ultrasonic measurement
2009First TMAH and photoresist analysers in flat-panel display plants
200+B30 series concentration meters installed
20+Countries with Rhosonics concentration and chemical density meters

Chemistries we already measure inline

TMAH developerDisplay and semiconductor

Display makers and chemical suppliers in Asia

KOHDisplay and etching

Display makers in Korea and China

Silane solutionsElectronic chemicals

Chemical producers in Taiwan

Carbonate de potassiumSemiconductor packaging

Semiconductor packaging plants in Taiwan

Hydrofluoric acidChemical production

Chemical and nuclear fuel producers in Europe

Sulphuric and hydrochloric acidChemical production

Chemical producers in Asia and Europe

DS Linetech
Our partner in Korea

DS Linetech works with us to solve difficult chemical concentration measurements for some of the biggest flat-panel display and semiconductor companies in South Korea.

Track record

  1. 2009Analysers in flat-panel display plants
  2. 2023Partnership with S3 Alliance, German maker of automated wet-processing tools
  3. 2024First CMP slurry paper, ICPT Wiesbaden Read it
  4. 2025CMP slurry monitoring presented at SEMICON China
  5. 2026Calibrated H2O2 results at ICPT Seoul
07 · Next step

Let’s talk about your slurry, at ICPT or after.

At the show? Find Stefan at our poster on Wednesday 21 October, or leave your details and we’ll arrange a meeting in Seoul. Can’t make it? Tell us what you’re running, CMP slurry or any other fab chemical, and we’ll set up a call.

Find us at ICPT 2026

Poster session, SeoulPresented by Stefan Seuleanu
Wed 21 Oct, 13:10 to 15:10

The authors

Stefan SeuleanuHead of R&D and Product, presenting
s.seuleanu@rhosonics.com

Raymond MaasChief Technology Officer, lead author
r.maas@rhosonics.com

In Korea

DS LinetechOur partner in Korea, solving difficult concentration measurements with us for the country’s leading display and chip makers.

Talk to a specialist

At ICPT or afterwards. A Rhosonics application specialist will reply.

Bien que nos régions ne soient pas standardisées, nous les utilisons pour vous mettre en relation avec le représentant commercial qui vous convient.
Produit(s) d'intérêt *