Impurity Analysis (Quantitative) II. A Legal Case Study

Course Material Index  Section Index  Previous Page  Next Page 

A Legal Case Study

The case study used here is taken from an actual court case that took place in March 1997 in the Patents Court, London. The plaintiffs owned a patent on the manufacture of an agricultural product commercially sold under the trade name Virkon. The defendants were accused of patent infringement by manufacturing a similar product sold under the trade name Virucidal Extra. Both were listed under UK Statutory Instrument 1997 as approved disinfectants under the "Diseases of Animals (Approved Disinfectants) (Amendment) Order 1997" (see links). A comparison of the raw materials used in the manufacture of each product is given in the table below:

The composition of the two products was clearly similar with the notable exception that plaintiffs' product was based on a water-soluble inorganic halide, which in practice was NaCl, while the defendants' product was based on sodium dichloroisocyanurate. In both cases, the virucidal action is due to the presence of hypochlorite ions in solution. However, patent EP-0 260 293 B1 specified (a) a "dry, water-soluble biocidal composition"; and (b) "0.01 to 5 parts by weight of water-soluble inorganic halide". The claim brought by the plaintiffs was that the defendants' product contained impurity sodium chloride within the limits covered by the patent in the dry solid product.

There were many scientific arguments in this case, one of which concerned the exact meaning of parts by weight when used in the form of a range. However, the main case rested on whether sodium chloride was present in the dry mixture. Although many wet chemistry experiments were performed, the key experimental tool in this case was X-ray powder diffraction.

The lower limit of 0.01 part by weight poses a severe problem for XRPD since it raises the question of sample sensitivity. In this particular instance the defendants were fortunate in that the diffraction intensity from sodium chloride is distributed in relatively few peaks due its cubic symmetry. The defendants' product was a mixture of many phases: so in order to check for crystalline sodium chloride, each phase could be analysed separately since the relative contribution of each phase to the mixture was not contested.

The plaintiffs' case rested on the existence of sodium chloride being present, as a manufacturing impurity, in the one chlorine-based component, i.e. the sodium dichloroisocyanurate. The scientific argument for this was based on the following (slightly hilarious) evidence: an X-ray powder diffraction pattern of the sodium dichloroisocyanurate was taken and the d spacings of the peaks were compared to those listed in the PDF database for sodium dichloroisocyanurate. An intense non-matching peak with a d spacing of 3.239 Å was then indexed as a 111 NaCl reflection, while a very weak peak with a d spacing of 2.825 Å was indexed as the NaCl 200 reflection!!! The exclamation marks are well-deserved (as I am sure anyone who has calculated the intensity ratio of these two reflections will appreciate). Moreover, certain comments in the PDF file for this compound (31-1883) should have set alarm bells ringing for any expert in XRPD: amongst other things, the PDF data was unindexed and of unproven quality. (Conversion of the PDF data back to its original 2θ revealed an instrumental zero error of approximately 0.16°.)

While it was easy to rubbish the above evidence as merely an example of poor scientific practice in the field of XRPD, a good expert witness in a court case has to examine the data without bias towards the party paying the expert witness fee. A quick data set of sodium dichloroisocyanurate was collected over a wide range of scattering angle using a well-calibrated diffractometer. In addition, the 111, 200, and 220 diffraction peaks of NaCl were measured. This provided additional checks on both diffractometer alignment and PDF data on sodium chloride. A long scan of sodium dichloroisocyanurate revealed no measurable peak in the region of the 220 peak of NaCl (see figure).

The initial experiments demonstrated the absence of NaCl within the detection limits of PXRD. However, a much stronger case is presented if all of the peaks in the pattern are indexed and identified. It is good laboratory practice to measure data on a material of unknown behaviour at least twice: a repeated data acquistion on sodium dichloroisocyanurate showed the diffraction pattern changing with time, probably due to hydration of the anhydrous powder. (These results instantly explained many of the unattributed peaks in the plaintiffs' data.) It soon became clear that the sample was hydrating while it was ground prior to mounting on the diffractometer.

Data were collected on a sample that had been ground and allowed to stand in a sample holder in the open air. This powder diffraction data was indexed with some difficulty, because the peaks were displaced slightly as a result of sample expansion during the hydration process which took place during the data collection period. (A similar effect is evident in the PDF data on "sodium dichloroisocyanurate".) Whole pattern fitting (to be explained in the next section) demonstrated that all peaks in the pattern could be accounted for in terms of an orthorhombic unit cell, a = 8.0175(4) Å, b = 14.5121(8) Å, and c = 6.6336(3) Å, with possible space group P2₁2₁2₁.

The indexed unit cell proves that there are no peaks from the hydrated sodium salt at the position (45.447°) of the 220 sodium chloride peak: the nearest sample peaks, 322 and 410, are at 45.296° and 45.654°, respectively. By contrast, there the 122 reflection (at 31.665°) of the hydrated sodium dichloroisocyanurate occurs at almost the same angle as would the NaCl 200 reflection (at 31.691°). This demonstrated that great care must be taken in the interpretation of any peak intensity in this region of the diffraction pattern.

Concerning the question of sensitivity of the technique, the best procedure for demonstrating this is by the "spiking method". Small amounts (typically 10 to 20 mg) of pre-ground sodium chloride were precisely weighed out (±10 µg) and added to the hydrated sodium dichloroisocyanurate (see figure). Note that it is important to spike the sample with low levels; the large (2.2%, 16.5%, and 49.2%) levels used initially in the plaintiffs' experiments were open to riducule. The spiking method clearly demonstrated that the level of NaCl in the sodium dichloroisocyanurate was less than 0.1%, from analysis of both plaintiffs' and defendants' data. Since sodium dichloroisocyanurate only contributes 5% by weight to the final product, the concentration of NaCl impurity in the final product due to use of sodium dichloroisocyanurate is less than 0.005%.

In order to be exceedingly thorough, the defence team also tested the stability of the hydrated sodium dichloroisocyanurate, since the PDF database made reference to a dihydrate (31-1884). This form was readily produced by leaving a sample in very humid conditions, but under ambient conditions the sample slowly reverted to the stable intermediate.

The case had two further scientific twists in it. Firstly, there were three manufacturers of sodium dichloroisocyanurate worldwide. The product used in the above experiments contained less than 0.1% NaCl. By contrast, two other manufacturers' products obtained by the plaintiffs were both contaminated with uncontestably high levels of NaCl. However, neither of these were used, by chance, in the manufacturing process! Secondly, sodium chloride had been used in the manufacture of the surfactant, thus accidently introducing dissolved chloride ions into the "solid" product. This was ruled not to be sodium chloride since the powder diffraction pattern of the surfactant failed to show any peaks due to crystalline NaCl.

There were many other legal arguments in this case. The outcome was generally in favour of the defendants. However, the whole case could have been avoided if the defendants had simply used different quantities of sulphamic acid and sodium hexametaphosphate in their product in the first place. This step alone would have avoided any possible infringement of the plaintiffs' patent in the first place; and provided no opportunity for scientific experts to testify in court!

Course Material Index  Section Index  Previous Page  Next Page