Application Bulletin 137/6 e

Coulometric water content determination according to Karl

Fischer

Application Bulletin 137/6 e

Coulometric water content determination according to

Karl Fischer

Branch Example for iodine: 2 I- - 2 e- → I2

All branches 126.9 g iodine is released by 96485 A in 1 s – or 1.315 mg

iodine is generated by 100 mA in 10 s.
Requirements for coulometric titrations are the following:
Keywords
• The process must take place with 100% current
Titration; Karl Fischer titration; coulometric; KFC; water efficiency.
content determination; ASTM E 1064
• No side reactions must occur.
• Oxidation or reduction must lead to a defined oxidation
Summary state.
This Application Bulletin gives an overview of the coulometric Metrohm KF Coulometers meet these requirements with
water content determination according to Karl Fischer. commercially available reagents. The instruments work
Amongst others, it describes the handling of electrodes, according to the galvanostatic principle, i.e. with a constant
samples, and water standards. The described procedures current.
and parameters comply with the ASTM E 1064. The same chemical processes as in a volumetric KF titration
take place, i.e., 1 mol H2O consumes 1 mol of I2. As the
iodine is generated electrolytically from the iodide in the KF
Introduction
reagent, the coulometric water content determination is an
Coulometric water content determination is primarily used for
absolute method and a titer determination is not necessary.
the determination of small amounts of water. Metrohm KF
In order to generate iodine at the anode of the generator
Coulometers work in a determination range of 10 µg –
electrode, Metrohm KF Coulometers work with variable
200 mg water. However, larger amounts of water (> 10 mg)
current strengths and pulse lengths. For generator electrodes
require a lot of time and/or may exceed the water capacity of
with diaphragms following current strengths are used: 100,
the KF reagent, which could lead to incorrect results.
200 and 400 mA. For cells without diaphragms a constant
Instead of a buret, the electric current is used to generate the
current of 400 mA is applied. Higher current strengths have
iodine needed for the Karl Fischer reaction. The current
not been able to establish themselves – side reactions occur
releases a stoichiometrically corresponding amount of iodine
and heat is produced. Working with variable pulse lengths
from the iodide-containing reagent.
allows precise «iodine addition» even in the region of the
Faraday’s law applies and can be used to calculate the water endpoint.
content of the sample.
The instrument measures the time and current flow that is
required to reach the titration endpoint. The product of time
M×Q and current Q is directly proportional to the amount of iodine
m= (1)
z×F generated and therefore to the amount of water determined
(see formula 1).
m Mass of converted substance in g The endpoint is indicated voltammetrically by applying an
M Molar mass in g/mol alternating current of constant strength to a double Pt wire-
Q Measured amount of charge in As electrode. This results in a voltage difference between the Pt
z Number of exchanged electrons (equivalence wires. The voltage drastically decreases in the presence of
number, charge number)
minimal quantities of free iodine. This fact is used to
F Faraday constant, 1 F = 96485 coulomb/mol
determine the endpoint of the titration.
(1 coulomb = 1 C = 1 ampere second = 1 As)

Page 1 of 5
 Application Bulletin 137/6 e
Coulometric water content determination according to Karl
Fischer

Instruments Parameters
Table 1: Method parameter for a coulometric KF titration
• Titrator with a mode for coulometric Karl Fischer
titration Parameter Setting
General I(pol) 10 µA
Generator current* 400 mA
Electrodes
Control parameter EP at 50 mV
Double Pt wire-electrode (indicator electrode for
Dynamics 70 mV
coulometric Karl Fischer titration)
Max. rate maximum
Generator electrode with or without diaphragm
µg/min
Min. rate 15 µg/min
Reagents
Stop criterion rel. drift
Special coulometric KF reagents are available from various
Relative stop drift 5 µg/min
manufacturers (e.g. Sigma-Aldrich, Merck, …). The reagents
Conditioning Start drift 10 µg/min
have been optimized for the use with generator electrodes
*For a generator electrode with a diaphragm the parameter
with and without diaphragms and also for special applications
Generator current is set to «auto»
(e.g. water content determination in ketones).
If ketone reagents are used, then the endpoint, start and stop
Additionally, liquid water standards are available. These
drift must be adjusted, as the ketone reagents suppress the
standards can be used to check the system (recovery of
side reactions, but do not prevent them completely.
added water, see formula 2).

Table 2: Control and titration parameter for ketone reagents

Filling the generator electrode Parameter Setting
It is assumed that the generator electrode is dry and has been General I(pol) 10 µA
assembled according to the Instructions for Use.
Generator current* auto mA
Generator electrode without diaphragm:
Control parameter EP at 20 mV
Handling the 6.0345.100 Generator electrode without Dynamics 70 mV
diaphragm is simple. It only needs one reagent and is quickly
Max. rate maximum
ready for use (no moisture deposits in the diaphragm!). Only
µg/min
reagents that are specially intended for use with generator
electrodes without a diaphragm must be used. About 100 mL Min. rate 15 µg/min
of the chosen reagent is filled into the titration cell and the Stop criterion rel. drift
ground joint opening is closed with the stopper. Relative stop drift 20 µg/min
Generator electrode with diaphragm: Conditioning Start drift 60 µg/min
Reagents for coulometric water determination with generator *For a generator electrode with a diaphragm, the parameter
Generator current is set to «auto»
electrodes with diaphragms consist of an anode solution
(anolyte), which is filled into the titration cell and a cathode
solution (catholyte) which is filled into the generator
Conditioning and drift
electrode.
The titration cell must first be dried before the determinations
About 100 mL anolyte is filled into the titration cell (anode can be started. The conditioning is started to remove water
chamber) and about 5 mL catholyte into the generator contained in the reagent and water on the surfaces of the
electrode with diaphragm (cathode chamber). We equipment (titration cell, electrodes, …). The water content
recommend that you use an injection syringe for filling in the determination of the sample should only be started once a
catholyte. The filling level should be about the same or
low and stable drift is reached.
2-3 mm lower than that in the anode chamber.
A constant drift equal or lower than 4 μg/min is acceptable.
Lower values are certainly possible. If higher, stable values
occur then the results are normally still good as the drift can
be compensated.

Page 2 of 5
 Application Bulletin 137/6 e
Coulometric water content determination according to Karl
Fischer

When working with an oven, a drift equal or lower than Sample size
10 μg/min is acceptable. The drift depends on the gas flow The sample size should be small so that as many samples as
(the smaller the gas flow the lower the drift). possible can be titrated in the same electrolyte solution and
the titration time is kept short. However, take care that the
sample contains approximately 50 μg H2O. The following
Sample addition
table provides guidelines for the sample weight.
Generally the coulometric titration cell should never be
opened to add samples. The influence of the humid air
Table 3: Water to be determined and water content with
entering the titration cell would falsify the results.
corresponding sample size
Liquid samples Water content of Sample size H2O to be
Liquid samples are added with the aid of a syringe. Either a sample determined
syringe with a long needle is used with the needle being 100000 ppm = 10% 50 mg 5000 µg
immersed beneath the surface of the reagent during injection.
10000 ppm = 1% 10 – 100 mg 100- 1000 µg
Alternatively a short needle can be used, with the last drop
1000 ppm = 0.1% 0.1 – 1 g 100- 1000 µg
being drawn back into the needle. The best way of
determining the actual sample weight is by weighing the 100 ppm = 0.01% 1g 100 µg
syringe before and after injection. 10 ppm = 0.001% 5g 50 µg
Volatile or low-viscosity samples should be refrigerated
before the sample is taken, in order to prevent handling
Tips and tricks
losses. In contrast, the syringe itself should not be directly
refrigerated as this could cause the formation of condensate. Reagent exchange
For the same reason aspirating air into a syringe which has In the following cases, the electrolyte solutions should be
been cooled by taking up a refrigerated sample should be exchanged:
avoided. • When the titration vessel is too full.
Highly viscous samples can be warmed to lower their • When the capacity of the reagent is exhausted.
viscosity; the syringe must also be warmed. The same goal
• If the drift is too high and shaking the cell does not
(lower viscosity) can be reached by dilution with a suitable
result in any improvement.
solvent. In this case the water content of the solvent must be
determined and deducted as a blank value correction. • If a two-phase mixture is formed in the titration vessel.
In this case only the sample phase can be aspirated off.
With samples containing a lot of water, care must be taken
that the needle is not introduced into the measuring cell • If during the determination the error message "check
through the septum before the determination has been generator electr." appears.
started, otherwise the drift and therefore the result of the For the generator electrode, with diaphragm the catholyte
analysis could be falsified. should be exchanged approximately once a week. Extended
With samples containing only traces of water the syringe use may cause darkening of the catholyte and yellow
must be thoroughly dried beforehand. If possible the syringe precipitation in the cathode compartment. An unpleasant
should be rinsed with the sample solution by taking up the smell also indicates the need for catholyte exchange.
sample solution several times and then discarding it. Indicator electrode
Solid samples A new indicator electrode may require a certain running-in
Whenever possible solid samples should be extracted or period for the formation of the surface. This may cause
dissolved in a suitable solvent and the resulting solution unusually long titration times and measurement results which
injected. A blank value correction should be made for the are too high. These phenomena vanish after a short period of
solvent. use. In order to speed up the running-in of a new indicator
electrode the Coulometer can be conditioned overnight, for
If no suitable solvent can be found for a solid sample, or if the
example.
sample reacts with the Karl Fischer solution, a drying oven
should be used. A polluted indicator electrode can be carefully cleaned with
an abrasive cleansing agent (aluminum oxide (6.2802.000
Polishing Set) or toothpaste). After cleaning it should be
rinsed with ethanol.

Page 3 of 5
 Application Bulletin 137/6 e
Coulometric water content determination according to Karl
Fischer

The two Pt wires of the indicator electrode should be as Handling of the liquid water standard
parallel to one another as possible. Check on insertion. 1 Open the ampoule containing the standard as
Cleaning recommended by the manufacturer.

The electrolyte solution can normally be exchanged without 2 Aspirate approximately 1 mL of the standard into the
any special cleaning of the parts. If cleaning is necessary then syringe and then eject the standard into the waste.
care should be taken that the Pt grid of the generator 3 Aspirate the remaining content of the ampoule into
electrode is not damaged. the needle (in case air is aspirated, eject the air out
Generator electrode with diaphragm of the syringe).

• Resinous deposits on the diaphragm  Hang the 4 Remove excess liquid from the outside of the needle
generator electrode vertically from a support rod, fill with a paper tissue.
with conc. HNO3 and allow standing overnight. Then 5 Place the needle on a balance and tare the balance.
rinse with water followed by methanol/ethanol*. 6 Then start the determination and inject a suitable
• Pollutants containing oil  Clean with a solvent (e.g. amount of standard (see table 1 to 3) through the
hexane) and then rinse with methanol/ethanol*. septum into the titration vessel. Do not inject the
• Salt-like deposits  Clean with water and then rinse whole content of the syringe! Please take care that
with methanol/ethanol*. the standard is injected into the reagent and not at
the electrode or the wall of the titration vessel. This
Cleaning (rinsing) the diaphragm  Fill the cathode
leads to unreproducible results.
compartment of the generator electrode with
methanol/ethanol* and allow the filling to drain out. Repeat 7 After injecting the standard, place the syringe again
the process 2-3 times. This process should also be carried on the balance.
out when the electrode has been cleaned as described 8 Enter the injected sample weight in the software.
above. Repeat step 4 to 8 at least three times. If the complete content
Generator electrode without diaphragm: of an ampoule has been injected, the needle can be filled with
• Pollutants containing oil  Clean with a solvent (e.g. fresh standard (same batch). In this case the needle does not
hexane) and then rinse with methanol/ethanol*. need to be rinsed again. Start directly with step 4.

• Salt-like deposits  Clean with water and then rinse There are two possibilities to add liquid standard. It can be
with methanol/ethanol*. injected with the tip of the needle above the reagent level. In
*Please make sure the ethanol does not contain any ketone
this case the last drop must be aspirated back into the
additives. syringe. Otherwise it is wiped off at the septum and might not
be determined although the weight of it is taken into account.
Dry all parts thoroughly after cleaning. A hot-air blower can
be used for this. If the parts are dried in a drying oven take If the needle is long enough, it can be immersed in the
care that the temperature does not exceed 70 °C (plastic reagent directly. In this case there is no last drop and the
components!). needle can be pulled out of the titration vessel without
aspirating back any liquid.

Checking the instrument

Commercial, certified water standard solutions with a water
content of 1.00 ± 0.003 mg/g and/or 0.10 ± 0.005 mg/g
should be used for checking the instrument as a fully
integrated measuring system.

Table 4: Recommended sample sizes:

Liquid standard 1.0 mg/g 0.2 – 2.0 g

Liquid standard 0.1 mg/g 0.5 – 5.0 g

Page 4 of 5
 Application Bulletin 137/6 e
Coulometric water content determination according to Karl
Fischer

Troubleshooting Titration times too long

Drift too high • Wait until drift during conditioning becomes stable.

• Depots containing water in the titration vessel  shake • Amount of water too large
titration vessel. • Set stop drift higher.
• Reagent exhausted or contaminated  exchange. • Set control range smaller.
• Moisture penetrating into titration vessel: • Set maximal titration rate higher.
 Molecular sieve exhausted?
 Septum pierced? Literature
 Seals not OK? • Metrohm Monograph water determination by Karl
 Ground joint sleeves not smooth? Fischer Titration. 8.026.5003 – 2003-09
• Generator electrode diaphragm polluted or moist. • HYDRANAL Multi Media Guide, Sigma-Aldrich
• Sample matrix consumes iodine. Change reagent more • HYDRANAL Manual, Eugen Scholz, Reagents for Karl
often. Fischer Titration, Sigma-Aldrich
• When working with Oven/Oven Sample Processor: • Merck Apura Analytical Application Notes Finder: Karl
 Molecular sieve of Oven/Oven Sample Fischer (Merck Webpage)
Processor exhausted? Metrohm Application Bulletins
 Gas flow too high? • AB 141 Analysis of edible fats and oils
 Allow to run overnight. • AB 209 Coulometric water determination according to
 Screw seals tight? Karl Fischer in insulating oils as well as in hydrocarbons
and their derivates
Drift unstable
• AB 280 Automatic water content determination using
• Poor stirring  Stir so, that mixing is efficient, but
gas extraction
without the formation of air bubbles.
• AB 357 Determination of water in gases and liquefied
• Reset the control parameters to standard values.
gases with the 875 KF Gas Analyzer
Result too high • AB 421 Automated coulometric Karl Fischer titration
• Titration vessel not properly conditioned  shake and
wait until drift has stabilized.
Author
• With the generator electrode without diaphragm  set
generator current to 400 mA. Competence Center Titration

• Sample contains substances which can be oxidized. Metrohm International Headquarters

• Set stop drift higher.

• Drift correction too small, e.g. with unstable drift or with
manual drift correction.

Result too low

• Drift correction too large, i.e. the drift was too high at
the start or unstable drift.
• Stop drift too high.
• Minimal titration rate too low.
• Sample releases iodine.

Results are widely scattered

• Inhomogeneous sample? Poor reproducibility of sample
addition?
• Drift unstable.

Page 5 of 5