Search suggestions
  • ph meter
  • balance
  • conductivity
  • counting platform scale
  • ind570
All Categories
  • All Categories
  • Products
  • Support
  • Expertise
  • Events & Webinars
  • Other
Filter
Filter
Categories
  • All Categories
  • Products
  • Support
  • Expertise
  • Events & Webinars
  • Other
Clear filtersShow results
Page Contents:

FFA™ Amperometric Control Algorithm

Faster, Accurate Water Determination with EVA KF Titrators

Home Expertise Library Understanding the FFA Amperometric Control Algorithm

Faster, Accurate Water Determination with EVA KF Titrators

The FFA™ Amperometric Control Algorithm forecasts titrant addition to shorten Karl Fischer titrations while keeping results accurate. This paper explains how it works and shows seven water-content determinations across oils, foods, solvents and plastics. 

What the FFA™ Algorithm Does

Karl Fischer (KF) titration is the fast, accurate, cost-effective standard for measuring water content. METTLER TOLEDO EVA titrators run the Fast-Forecasting Amperometric algorithm (FFA™), which forecasts titrant addition to cut titration time without sacrificing accuracy.

This White Paper explains the KF technique, then details how the FFA™ controls titrant addition or generation and evaluates the sensor signal. The setup chapter covers the instruments and the key control parameters. The results chapter presents seven water-content determinations: four with the volumetric titrators EVA V1 and EVA V3, and three with the coulometric titrators EVA C1 and EVA C3. A summary closes the paper.

What the FFA Amperometric Control Algorithm Does

How Karl Fischer Titration Works

How Karl Fischer Titration Works

KF titration has been the gold standard for water determination since 1935, valued for specificity, accuracy, reliability and speed. Water reacts with iodine, sulfur dioxide and a base in a suitable solvent; iodine is the titrant.

The reaction consumes water in the presence of iodine, SO2, a base (RN) and a solvent such as methanol. Classic formulations use methanol and imidazole, but many modern reagent combinations exist. Iodine is the titrant; SO2, the base and the solvent are kept in excess so water and iodine react quickly.

Equation 1 — The Karl Fischer Reaction

CH3OH + SO2 + I2 + H2O + 3 RN → (RNH)·(CH3OSO3) + 2 (RNH)·I 

Why the Cell Must Stay Dry

Any water not from the sample distorts the result. Modern instruments use a closed KF cell and remove residual water before analysis (pretitration).

Because even closed cells are not fully tight, incoming water is compensated by controlled iodine addition or generation. Pretitration plus this compensation together are called cell conditioning, which keeps the cell dry until a sample is injected. Both conditioning and sample analysis need control algorithms that read a polarized sensor — either by bivoltametric or biamperometric indication.

Why the Cell Must Stay Dry

Karl Fischer Titration Principles and Performance Optimization

Bivoltametric vs Biamperometric Indication

A polarized double-pin platinum electrode detects iodine, and therefore water. Bivoltametric indication holds a constant current and measures voltage; biamperometric indication holds a constant voltage and measures current. EVA titrators use biamperometric indication.

When water is present, all iodine is consumed, so monitoring iodine is a reliable water indicator. At the cathode iodine is reduced; at the anode iodide is oxidized back to iodine, creating an ionic current that makes the sensor highly sensitive to iodine.


Right figure: Schematics of a polarized double pin platinum electrode with indications of electronic and charge carrier currents.

Figure 1

Bivoltametric Indication (Constant Current, Measures Voltage)

A variable potential holds a constant polarization current (Ipol). With water present, iodine is low and a high potential (mV) is needed; without water, iodine is high and the potential drops. The voltage is a reliable water indicator with a sharp response at low iodine levels.

That sharpness makes it a good endpoint indicator, but slow iodine–water reactions can cause signal oscillations and long titration times.


Right figure: Schematic relationship between iodine concentration (c(I2)) and measured potential (E) in a KF cell.

Figure 2

Schematic Titration Curve for a Typical Bivoltametric Kf Titration

Figure 3

Left figure: typical curve shape for a fast reacting system. 

Right figure: typical curve shape for a slow reacting system.

Biamperometric Indication (Constant Voltage, Measures Current)

A constant polarization potential (Upol) is applied and the measured current is proportional to iodine concentration — an almost linear relationship. Being less sensitive right at the endpoint, it controls oscillations from slow reactions more easily.

  • Set current ↑ = larger iodine excess 
  • Larger excess = faster titration
  • Higher iodine = darker solvent color
Figure 4 Figure 5

Left figure: Schematic relationship between iodine concentration (c(I2)) and measured current (I) in a biamperometric measurement setup.

Right figure: Schematic titration curve for a biamperometrically controlled titrant addition.

Volumetric vs Coulometric KF Titration

Volumetric KF adds an iodine solution by burette and suits 100 ppm to 100% water. Coulometric KF generates iodine electrochemically and suits trace water from 1 ppm to 5%.

An iodine-containing titrant is dosed into the reaction mixture by a motorized piston burette. Two reagent types are used:

  • One-component: iodine, methanol, SO2 and base in the titrant; usually paired with pure methanol as cell solvent.

  • Two-component: no SO2 or base in the titrant — these are supplied by the solvent instead.

Coulometric KF titration — 1 ppm to 5%

Iodine is generated electrochemically at a generator electrode. The reagent must supply every reaction partner: short-chain alcohol, SO2, a base, and iodide that the electrode oxidizes to iodine on demand. More detail is in the METTLER TOLEDO KF Good Titration Practice guide.

Volumetric vs Coulometric KF Titration

Drift: What It Is and Why It Matters

Drift is the water titrated per minute (µg/min) and is proportional to titrant addition rate, so low drift means a dry cell. Because drift is not sample water, it biases results — worse for small samples and long titrations — but a known, stable drift can be corrected.

The instrument measures and compensates drift automatically during both conditioning and sample analysis. In a well-conditioned cell the drift is low and stable; during analysis it rises, then falls back once all water is titrated. Start and stop criteria are therefore usually tied to drift.

Two sources of drift:

  • Physical drift: water entering the cell from outside or desorbing from cell walls.

  • Chemical drift: water from side reactions in the solvent — e.g. ketones and aldehydes reacting with alcohols to form ketals or acetals plus water.

Drift: What It Is and Why It Matters

Example

icon

Sample Water 1 mg

icon

Drift 10 µg/min

icon

Time 2 min

Measured 1.02 mg (+2%)

Illustrations of the Origin of Drift

Figure 6

Left figure: Physical drift as water entering the Karl Fischer cell from outside or via resorption from the cell walls.

Right figure:  Chemical drift as water produced in side-reactions of the sample with the Karl Fischer solvent.

What Controls Titration Time

Titration time depends on how much water is added, the endpoint approach, and the iodine–water reaction rate. The FFA™ drives the cell to a low, stable drift as fast as possible.

importance of centralized robotic automation strategy

There is a maximum rate of iodine addition or generation. In coulometric KF the maximum water neutralized per minute is about 2 mg, so larger samples take longer.

  • Coulometric: inject < 2 mg water
  • Volumetric: inject < 20 mg water

Approaching the endpoint, iodine must be added carefully to avoid overtitration — often the longest phase, and the one with the most time-saving potential. Reaction rate also matters, e.g. with fresh one-component reagents.

What Controls Titration Time

The Reaction Rate of Water Consumption

The water-consuming reaction is a time-limiting factor. Its rate is proportional to the concentrations of iodine, sulfur dioxide and water:

Equation 2 — KF Reaction Rate


−d[I2] / dt = k · [I2] · [SO2] · [H2O]


At low iodine or SO2, water consumption is slow. This is common in volumetric titrations with one-component reagents, especially in the first few runs when cell SO2 is very low.

Benefits of the FFA™ Control Algorithm

FFA™ reads the dSens M143 digital double-pin platinum electrode at a constant Upol and recalculates the titrant to add at 10 Hz, compensating incoming water in real time. The result: substantially shorter titrations with accurate results.

importance of centralized robotic automation strategy

Incoming water reacts with iodine and perturbs the system; the FFA™ computes the titrant needed to restore balance ten times per second.

Right figure: Schematic illustration of the FFA™ KF control, which defines the amount of iodine that is added or generated to the chemical system based on the current measured by the dSens M143 sensor.

Benefits of the FFA™ Control Algorithm
figure 8

Handling Slow Reaction Rates

With one-component reagents, low SO2 and base can slow the reaction. Using the control focus "Accuracy", the FFA™ classifies the cell chemistry as fast or slow from a calibration done during conditioning, then adjusts control parameters to dose the exact titrant needed efficiently.

For slow reactions the FFA™ predicts the titrant needed for completion instead of waiting, thanks to the near-linear biamperometric response. This can cause a brief apparent overtitration; the current then settles back to the endpoint once the reaction finishes.

Left figure: Titration curve with biamperometric indication obtained for slow reactions with the control focus “Accuracy” using an EVA

Two Speed Gains Follow

  1. The titrator does not wait for the slowly evolving signal.

  2. Higher iodine concentration speeds up the reaction (per Equation 2).

Note: EVA coulometric titrators have no control foci, because slow reactions are unusual with coulometric reagents.

Drift Stability Start/Stop Criteria

Classic criteria start and stop a titration on absolute or relative drift thresholds. EVA titrators add a drift-stability criterion that checks whether the drift is stable regardless of its absolute value, with automatic drift correction ensuring accuracy.

Illustrations for Start and Stop Criteria That Are Used Classically in Karl Fischer Titrations

Figure 9

Left figure: maximum start drift

Right figure: relative drift stop

The titrator begins a stability check when drift changes by less than 2 µg/(min/s). Within the check, drift must change by less than x during y seconds to terminate:

Figure 10

Because corrections can be applied, drift stability can matter more than the absolute drift value. Its key advantage over a relative drift stop: it does not depend on absolute limits — useful for samples with unavoidable side reactions (ketones, aldehydes).

Left figure: Illustration of the drift stability criterion. The drift is considered stable if its changes remain in a certain margin for a certain time.

Key Advantages of the FFA™

  • Biamperometric indication forecasts titrant amount, especially for slow one-component-reagent systems.

  • Adjustable set current works at higher iodine concentrations, raising reaction rate and cutting titration time.

  • A dedicated "Speed" control focus is optimized for faster volumetric titrations.

  • Higher analog resolution (burette drive and generator electrode) gives a more accurate drift representation.

  • Drift-stability start/stop criteria shorten waiting time and can enable determinations impossible with classic criteria — or that would need frequent solvent exchange.

System Setup and Method Parameters

System Setup and Method Parameters

All examples use EVA titrators with a dSens M143 electrode and a dPump KF solvent pump. Volumetric runs use EVA V1/V3 with a 5 mL burette; coulometric runs use EVA C1/C3 with a generator electrode.

EVA V1 / V3 Volumetric Systems

The volumetric system comprises the titrator with a detachable terminal, the dDrive burette drive with burette, a titration kit with the KF cell, a dSens M143 double-pin platinum electrode, and a dPump KF solvent pump for automatic solvent exchange:

  • Online drift stable & < 25 µg/min after conditioning  
  • Titrant via 5 mL burette
  •  Fresh molecular sieve in all drying tubes

The signal is digitized by the dSens M143 and processed by the titrator; solvent and waste tubes stay mounted during analysis.

Right figure: EVA V1 titration with titrator, terminal, KF cell, dSens M143 electrode, dosing unit and solvent pump dPump KF.

EVA V1 / V3 Volumetric Systems

EVA C1 / C3 coulometric systems

EVA C1 / C3 coulometric systems

The coulometric system comprises the titrator with terminal, a generator electrode (with or without diaphragm), a titration kit with the KF cell, a dSens M143 electrode, and a dPump KF solvent pump.

  • Online drift stable & < 10 µg/min after conditioning  
  • Iodine generated at the generator cell
  • Volume levels equalized with dried air

Right figure: EVA C3 titration system with titrator, terminal, dSens M143 electrode, KF cell, generator electrode and d Pump KF solvent pump.

Method Parameters Reference

EVA titrators expose several parameters that tune accuracy and speed. Volumetric control parameters are in Table 2, coulometric in Table 3, and shared start/stop criteria in Table 4.

importance of centralized robotic automation strategy

Coulometric KF Control Parameters Affecting Result Quality

ParameterWhat it does
Potential (Upol)Fixed polarization potential (mV). Higher potential can reduce overtitration risk for some reagents but may lengthen titrations.
Set currentTargeted current (µA) in the conditioned cell. Typical: 2–5 µA.
Cautious modeLimits the initial titrant rate so small water amounts are measured accurately.
Generator modeSets whether the iodine-generation pulse amplitude is fixed or chosen automatically. Automatic mode adjusts both pulse height and length, spanning maximum to very low iodine generation.

Start/Stop Criteria Shared by Volumetric and Coulometric Methods

ParameterWhat it does
Max. start driftMaximum absolute drift (µg/min) below which analysis may start. Higher settings increase the drift’s influence — more so for small water contents and long analyses.
Drift stop relativeTerminates when drift falls to (start drift + relative value). Larger values give secure endpoints and shorter analyses but raise potential error.
Drift stabilityTerminates on drift stability instead of an absolute value. Stable = change < 40 µg/min/s over 10 s ("accuracy") or < 60 µg/min/s over 10 s ("speed"). Independent of absolute drift; ideal when side reactions raise drift on injection.
Fixed timeTerminates after a set number of seconds. Mainly for automated gas-phase (oven) extraction.

Results: Seven Application Examples

Results: Seven Application Examples

Seven FFA™ determinations — a Ph. Eur. suitability test, edible oil, tea leaves, gummy bears, dry acetone, silicone oil and PPS plastic — all show high repeatability and short analysis times.

PH. Eur. 2.5.12 Suitability Test (Volumetric)

Result

A 1% certified water standard met all four Ph. Eur. 2.5.12 acceptance criteria: recovery 99.37%, slope 0.994, y-intercept error −0.29%, x-intercept error 0.28%.

Ph. Eur. 2.5.12 determines the sample’s water content first, then adds five portions at 50–100% of that amount. The instrument runs a linear regression and checks recovery, slope, and intercept errors.

importance of centralized robotic automation strategy

Equation 3

r = 100 · (W1 / W2)

% recovery; W₁ = water added, W₂ = water found (mg)


Equations 4 and 5

e1 = 100 · (a − M) / M      e2 = 100 · (|d| − M) / M

intercept % errors; a = y-intercept, |d| = |x-intercept|, M = water found (mg)

Setup

  • Sample: 1% certified water standard (10 mg H2O/g).
  • Solvent: HYDRANAL™ Methanol Dry (Fluka).
  • Titrant: HYDRANAL™ Composite 5, titer 5.218 mg/mL (set with HYDRANAL™ Water Standard 10.0, certified 10.03 mg/g).
  • Control focus "accuracy", Upol 100 mV, set current 30 µA, max start drift 50 µg/min, relative drift stop 50 µg/min, cautious mode off.

Linear Regression Applied on Measured Water Contents for a Sustrainabliliy Test for Volumetric Karl Fisher Water Determination According to PH

Figure 13

Recovery of 99.37% sits close to the ideal 100% and inside the limits; the 0.994 slope and small intercept errors also comply with Ph. Eur. 2.5.12.

Water in Sunflower Oil (EVA V1/V3)

Result

Ten samples averaged 1.941 mg/g water at srel 1.86%, mean titration 40.4 s. Low water and poor methanol solubility make edible oils challenging — yet repeatability stayed high.


Setup

 ~1 g oil injected into 30:70 dry methanol/chloroform; mass by back-weighing. Titrant CombiTitrant 5 (Merck), titer 5.0496 mg/mL. Control focus "accuracy", Upol 100 mV, set current 30 µA, cautious mode on (low water expected), relative drift stop 50 µg/min, max start drift 50 µg/min.

Water in Tea Leaves by External Extraction (EVA V3)

Result

Six Rooibos samples averaged 6.65% water at an exceptional srel 0.08%, mean titration under 90 s. External extraction handles solids that will not dissolve in KF solvent.

Crushed leaves are extracted in dry methanol; the supernatant is titrated and blank-corrected.

importance of centralized robotic automation strategy

Equation 6

R(%) = [100 / (100 − wtot)] · [ wtot · (msol/mext) − B · (msol/mext) ]

  • w_tot = supernatant water (%)
  • m_sol = solvent (g); m_ext = sample (g)
  • B = blank water (%)


Setup 

50 mL dry methanol + 3–4 g crushed leaves, sealed, 3 h with occasional shaking; blank from methanol only. Solvent HYDRANAL™ Methanol Dry (Fluka); titrant HYDRANAL™ Composite 5 SC (Fluka). Control focus "accuracy", Upol 100 mV, set current 30 µA, cautious mode on, relative drift stop 25 µg/min. Sample mext 3.8213 g; solvent msol 40.303 g.

Tea-Leaf Samples by External Extraction

SampleWater [%]Time [s]
16.65488.1
26.65897.1
36.64881.1
46.64394.1
56.64785.1
66.64993.2
Average6.65089.8
s0.005
srel [%]0.08

Water in Gummy Bears (EVA V3)

Result

Six samples averaged 14.274% water at srel 0.97%, effective titration ~0:25. A high-speed disperser and formamide fully release and dissolve the water from sugar-rich candy.


Setup

Pieces < 25 mg, ~100 mg per run; dispersed at 24,000 rpm (80%) for 2 min, then 15 s settle. Solvent 1:1 HYDRANAL™ Methanol Dry (Fluka) / EMSURE® Formamide (Supelco); titrant HYDRANAL™ Composite 5, titer 5.163 mg/mL. Control focus "speed", Upol 100 mV, set current 80 µA, cautious mode on, relative drift stop 50 µg/min, max start drift 50 µg/min. Disperser: POLYTRON® PT1300 D (Kinematica), PT-DA 12/2EC-E123 aggregate, homogenizer adapter 30869296.

Water in Dry Acetone (EVA C1/C3)

Result

Dry acetone averaged ~24.5 ppm water at srel 5.8%. Despite rising drift from ketone side reactions, all samples ran without reagent exchange and with no visible effect on results.

Ketones and aldehydes react with methanol to release water, raising drift and often forcing reagent replacement. K-reagents suppress but do not stop this. The drift-stability start criterion lets the titration begin even at elevated drift.

Setup

Acetone (CH3-CO-CH3) dried over molecular sieve overnight; ~2 mL injected per run. Anolyte HYDRANAL Coulomat AK; catholyte HYDRANAL Coulomat CG-K; generator electrode with diaphragm. Set current 2 µA, cautious mode on, drift-stability start, relative drift stop 10 µg/min.

Water in Dry Acetone

SampleSize [g]Water [ppm]
11.6275125.3
21.6458325.9
31.6437522.2
41.6180426.2
51.6096418.9*
61.7621423.9
71.5914323.5
Average—24.5
s [ppm]
1.4
srel [%]
5.8

*Sample 5 Treated as an Outlier and Removed from the Statistics.

Water in Silicone Oil (EVA C1/C3)

Result

Six silicone-oil samples averaged 60.63 ppm water at srel 2.51%. Low water content underpins silicone oil’s use as a lubricant, thermic or hydraulic fluid.


Setup

~2 mL injected per run. Anolyte HYDRANAL Coulomat AG + 20% dichloromethane; catholyte HYDRANAL Coulomat CG; generator electrode with diaphragm. Set current 5 µA, cautious mode on, drift-stability start, relative drift stop 2 µg/min.

Water in Silicone-Oil Samples

SampleSize [g]Water [ppm]
11.2557061.59
21.2930161.88
31.3834860.17
41.6796558.11
51.7822962.06
61.5938959.96
71.5914323.5
Average—60.63
s [ppm]
1.52
srel [%]
2.51

Water in PPS Plastic (Inmotion Kf Pro + EVA C3)

Result

Six PPS samples averaged 449.9 ppm water at srel 2.3%, mean transfer-plus-titration time 9:20. Oven extraction handles plastics that cannot dissolve in KF solvent.

Polyphenylene sulfide (PPS) is a high-performance thermoplastic whose water content affects properties, processing and durability. Because PPS will not dissolve in KF solvent, water is driven off by heating in an oven autosampler.

Setup

Six vials ~1.5 g, weights auto-transferred from the balance; two blank vials plus a drift vial. Reagent HYDRANAL Coulomat AG-Oven; InMotion KF Pro 10 mL Oven Autosampler with EVA C3; generator electrode without diaphragm. Oven 220 °C, max start drift 10 µg/min, relative drift stop 3 µg/min with 5 s delay.

Water in PPS Samples

SampleBlank [µg]Water [ppm]Size [g]Time [min:s]

Blank 1

98.3

—

—

—

Blank 2

96.9

—

—

—

11.38348

445.4

1.5013

9:53

21.67965433.7

1.6139

8:38

31.78229457.3

1.3821

9:20

41.59389453.8

1.5081

9:04

51.59143463.3

1.3535

8:54

6—446.1

1.5545

10:12

Average97.6449.9

—

9:20

s [ppm]
10.5

srel [%]
2.3

Repeatability of ~2.3% shows the method stopped cleanly once all sample water had transferred and titrated.

Conclusions

Scientist working in a lab

EVA volumetric and coulometric KF titrators deliver fast, accurate water determination. Biamperometric indication lets the FFA™ optimize titrant addition — faster analyses at high accuracy.

Across sunflower oil, tea leaves, gummy bears, acetone and silicone oil, results were highly repeatable with short analysis times. The Ph. Eur. 2.5.12 suitability test met every acceptance criterion, and PPS plastic was determined accurately with the InMotion KF oven autosampler and an EVA C3. Overall, the EVA KF titrators are a reliable, efficient choice for Karl Fischer water determination.

Glossary of Karl Fischer terms

Conditioning Keeping the KF cell dry and ready for titration. Includes a pretitration.
DriftWater titrated per minute in the KF cell, in micrograms of water per minute (µg/min).
EndpointTarget sensor signal showing a titration is nearly complete. A second, drift-related criterion (e.g. a relative stop drift) usually confirms completion.
EVA METTLER TOLEDO titration platform. Stands for Efficient, Versatile, Accurate.
FFAFast-Forecasting Amperometric — the algorithm EVA titrators use to control titrant addition orgeneration.
KFKarl Fischer, inventor of the KF titration for water-content determination
KF cellClosed vessel in which the Karl Fischer titration takes place.
Overtitration Passing the endpoint, leaving an excess of titrant in the cel
PretitrationRemoving water from the KF cell so that all water found during a titration comes from the sample.

References

  1. Angew. Chem., vol. 48, pp. 394–396, 1935.
  2. E. Scholz, Karl-Fischer-Titration, Springer-Verlag, 1984.
  3. METTLER TOLEDO, GTP Brochure: Good Titration Practice in Karl Fischer Titration.
  4. Ph. Eur. 2.5.12, "Water: Semi-micro determination".

Disclaimer

This report presents one or more examples of analysis, conducted with utmost care using the instruments described. Results were evaluated to the current state of our knowledge; other laboratory conditions may require slight method changes. When chemicals, solvents and gases are used, follow the general safety rules and the manufacturer’s or supplier’s instructions.

METTLER TOLEDO Group · Local contact: www.mt.com/contacts · For more information: www.mt.com

Subject to technical changes. © 10/2025 METTLER TOLEDO. All rights reserved. Rewritten for web/LLM readability from RITM1438115.

I want to…
Need assistance?
Our team is here to help you achieve your goals! Speak with our experts.