Automatic KF water determination

Regina Schlink, Claudia Dengler, Dr Peter Kirschenbuehler and Dr Karl-Heinz Surborg discuss the use of a new Karl Fischer oven method to determine the water content of drugs

Many active pharmaceutical ingredients and adjuvants contain water in an adsorbed form (surface water) or bound as a hydrate (water of crystallisation.

The water content of drugs strongly influences their quality, shelf life and stability as well as the release of the active substances. The determination of water therefore assumes great importance in pharmaceutical analysis.

The European Pharmacopoeia, 4th Edition (2002), describes various methods for determining the water content of pharmaceuticals. By far the most important method is the Karl Fischer (KF)titration.

Normally the titration is carried out volumetrically (semi-micro determination). For substances with a very low water content a coulometric KF titration (micro-determination) is performed.

traditional methods

Many substances release their water only slowly or at high temperatures. They are therefore not suitable for a direct Karl Fischer titration. An additional problem is the low solubility of certain samples in alcohols. In these cases traditional methods recommend the use of toxic solvents to promote dissolution or alternatively extensive sample preparation procedures. Other substances undergo side reactions with the KF reagents, thus falsifying the result.

The European Pharmacopoeia specifies that these types of pharmaceuticals are not to be analysed by Karl Fischer titration but by determining the loss on drying in a drying cabinet or desiccator (under vacuum if necessary).

However, with this method all volatile components released at the particular temperature (e.g. impurities) are determined and not specifically the water content of the substance.

By using the KF oven method the above-mentioned problems can be avoided. The substance under investigation is heated in a tube oven and a carrier gas transfers the water released to the titration cell where it is determined by Karl Fischer titration. As only the water enters the KF cell and the sample itself does not come into contact with the KF reagent, this means that unwanted side reactions and matrix effects are eliminated.

Metrohm's 774 Oven Sample Processor allows for automation of the KF oven method. In contrast to the conventional Karl Fischer drying oven, the samples are no longer introduced into the oven using a sample boat, but rather by using a vial technique. The substances to be analysed are weighed directly into sample vials, which are then sealed tightly and placed in the rack of the processor.

process advantages

For the analysis the sample vessel is moved by the turntable to the appropriate position above the oven and then lowered automatically into the heating block. At the same time a double hollow needle pierces the septum of the vial. Via the inlet needle a stream of dry carrier gas (air or inert gas) is passed through the heated sample.

The carrier gas, loaded with the released moisture, then flows through the outlet needle and a heated transfer tube directly into the titration cell, where the Karl Fischer water determination takes place. Depending on the sample's water content, the determination is carried out either volumetrically or, at trace levels, coulometrically.

The automation of the Karl Fischer water determination using the processor brings decisive advantages:

• Strictly reproducible analysis conditions for all samples as demonstrated by the significantly improved repeatability of the results;

• Considerably increased sample throughput and therefore improved efficiency;

• Manual sample preparation is reduced to a minimum;

• Considerable savings in time;

• No contamination of the oven and titration cell; consequently there are no carryover and memory effects;

• Much lower reagent consumption as the titration solution requires changing only at infrequent intervals;

• Improved water release from the sample as the carrier gas does not just pass over the sample but directly through it.

In addition, the oven allows temperature gradients to be run. Using the recorded water-release curve, it is possible to determine the optimum analysis temperature for the particular sample. The curve also allows statements to be made about the kinetics of water release as a function of temperature.

product analysis

Using the KF oven method, some 40 pharmaceuticals from the European Pharmacopoeia were analysed. The analyses were carried out with the 774 Oven Sample Processor in combination with a 756 KF Coulometer.

The investigated pharmaceuticals were substances with defined water content, some of which undergo side reactions with the KF reagents and therefore cannot be analysed by direct Karl Fischer titration. As stated above, according to the Pharmacopoeia the water content of such substances must be determined by loss on drying in a drying cabinet or desiccator (under vacuum if necessary).

The procedure and reagents are set out in table 1.

When selecting the temperature to be used for driving off the water, the thermal stability of the particular pharmaceutical must be taken into account as well as the fact that water is released at a sufficiently rapid rate only at temperatures above 100°C.

This means that the oven temperature should be chosen as high as possible to ensure short determination times, but still be 20 to 30°C below the decomposition temperature.

The analysis temperatures are determined on the basis of the water-release curves that were recorded for all the investigated pharmaceuticals in the temperature range 50;250°C. Figure 1 shows such a water-release curve for metamizole sodium.

In addition, all the pharmaceuticals were examined by means of a Kofler microscope and their melting points were determined. This instrument allows the substance to be closely observed during the heating-up and melting phases; any alterations such as colour changes, sublimation or decomposition reactions can be easily recognised.

Metamizole sodium decomposes at 220 to 221°C. Water determination by direct volumetric or coulometric KF titration is not possible as the substance is oxidised by iodine.

The water-release curve shown in figure 1 was recorded using a heating rate of 2°C/min. The red curve corresponds to the absolute amount of water released, while the blue curve shows the associated drift.

Both the surface moisture and the water of crystallisation are released within the time interval 0-1600s (50-103°C); this is indicated by the continuous increase of the red curve in this region as well as by the occurrence of the 'drift peak'.

water content

The drift then falls to its original value of approximately 10µg/min and remains virtually constant for 3800s. Starting at 5400s (230°C) both curves show a steep increase. Evidently water is released by decomposition from this temperature onward.

A temperature from the central region of the plateau of the red curve (150°C) was selected as the oven temperature for determining the water content of metamizole sodium. This ensures that the water is released quickly and completely without product decomposition.

The complete results of the KF determination for 15 of the substances among the 40 investigated pharmaceuticals are shown in table 2.

By way of example, figure 2 shows the titration curve for morphine hydrochloride at an oven temperature of 180°C. The red curve again corresponds to the absolute amount of water released and the blue curve to the associated drift. The three 'drift peaks' show very clearly that morphine hydrochloride is present in the form of its trihydrate.

It can also be seen from both curves that the substance stops releasing water after about 180s. The subsequent slight increase of the red curve can be attributed to the low background drift (blank consumption). It is generally recommended to use an extraction time of, for example, 5-10min to ensure that the determination is not stopped too soon, especially if the sample releases its water of crystallisation slowly.

The water contents determined with the new procedure all lie within the ranges specified in the European Pharmacopoeia. This usually gives a very wide recovery range for the loss on drying. In the case of quinine hydrochloride, for example, a range between 66.2 and 110.3% is specified, based on the theoretical (calculated) water content.

In contrast, the oven system yields an excellent recovery of 96.8% for this substance. When all the investigated pharmaceuticals are considered, the recovery using the KF oven method lies between 90 and 110%.

excellent repeatability

The repeatability of the values obtained with the oven system is also excellent. This can be seen from the relative standard deviations, which lie between 0.3 and 2.0%.

When compared with the drying cabinet method, the KF water determination using this method offers determinations, which normally take several hours, being performed in 10 to 12min. In addition, the analytical procedure is completely automated.

There is a further bonus point of the specificity of the described method, in which only the water released by the substance is determined and not all the other volatile components released at the particular temperature.

Finally, the small amount of substance required is an advantage, particularly when analysing very expensive pharmaceuticals. Whereas the determination of the loss on drying normally requires sample weights of 1g, the KF oven method in combination with the KF Coulometer requires only 15 to 30mg.

Moreover, as the substance is not destroyed during the water determination it can be used again for further investigations.