T 0822/96 27-05-1999

1. The appeal is admissible.

2. In the Statement of Grounds of Appeal, inventive step has been the only issue raised by the Appellant. The Board concurs with the findings of the Opposition Division in the decision under appeal that novelty is given vis-à-vis the cited prior art and that the requirements of Article 123(2) and (3) EPC are met. There is thus no need to consider these matters in further detail.

3. The patent in suit concerns a washing method for a solution dissolving a polycarbonate resin in an organic solvent.

3.1. Such a method is known from E2 which the Board, like the parties and the Opposition Division, regards as representing the closest state of the art.

3.1.1. This citation describes "a method of purification for a polycarbonate solution using an orifice-mixer characterized by comprising steps of contacting a solution of crude polycarbonate containing an impurity matter in an organic solvent with a washing liquid using an orifice-mixer while adjusting the pressure drop in said orifice-mixer to a value within the range of 2 to 12 kg/cm2 and keeping the proportion of said washing liquid to the whole fluid at 5 to 45% by volume." (Claim 1). In its example, E2 additionally describes the separation of the aqueous phase from the polymerisation reaction solution by means of a centrifuge after dilution of the solution with further organic solvent (page 4, last line to page 5, line 2).

3.1.2. This known method aims at an improved, easy and effective method for the purification of polycarbonate resin obtained in a phosgene process, thereby avoiding, first, a complicated sequence of operations, and, secondly, the problem of antinomy brought about by the difficulty of perfect separation of the resin solution and the washing liquid at the end of the purification caused by the vigorous high speed mixing in order to achieve efficient washing (introduction of the description on page 1 and page 2, paragraph 1; Reference Example 2).

3.1.3. The examples and comparative (reference) examples (pages 5 and 6) based on repeated washing steps including (a) mixing by means of an orifice-mixer under controlled pressure conditions and controlled ratios of washing liquid and resin solution and (b) separation by still standing, show that, on the one hand, the purity of polycarbonate resin expressed in terms of light transmittance greatly depends on the mixing conditions, and, on the other hand, mixing by intensive mechanical stirring causes difficulties as regards the separation of the resin solution and the washing liquid.

3.2. In the light of these shortcomings and in line with the introductory statements in the patent specification (page 2, lines 5 to 9, 15 to 23 and 35 to 37), the technical problem underlying the patent in suit may thus be seen as the definition of a process not only more efficient in removing impurities from polycarbonate resin solutions as directly obtained in the phosgene process, which optionally includes a centrifugal separation step prior to the purification, but also more simple so as to rapidly purify solutions with even high concentrations of polymer causing high viscosities.

3.3. According to the patent in suit, this problem is solved by (i) using a line stirring mixer at or above a specific minimum stirring power to prepare a water-in-oil type dispersion containing a limited amount of an aqueous washing liquid dispersed in an organic solution of polycarbonate resin, and (ii) subjecting that water-in-oil dispersion to centrifugal separation, as specified in Claim 1.

3.4. In both tables on pages 5 and 7 of the patent in suit, as well as in the further experiments submitted during examination (received on 29 October 1991), it has been demonstrated that a significantly improved reduction of the contents of inorganic and organic impurities, viz. of sodium and chloride ions and bisphenol A, is achieved by operating in accordance with Claim 1 in comparison to prior art methods or modified prior art methods (see in particular Comparative Examples 5 and 6 which meet the requirements as defined in the claim of E2).

In accordance with the method as claimed, the flow rate of the 12.1 wt.% resin solution in the examples is kept at 38 l/h, compared to 35 and 27 l/h, respectively, in the experiments of E2. The assessment of the polymer concentration for the experiments of E2 results in a concentration in the resin solution used in the purification which is not higher than that in the examples in the patent in suit.

Consequently, the two aspects of the above defined technical problem are effectively solved by the method as defined in Claim 1 of the patent in suit.

4. It remains to be decided whether this solution was obvious to a person skilled in the art having regard to the state of the art relied upon by the Appellant.

4.1. It is evident from the above considerations that document E2 by itself cannot render the claimed invention obvious. Not only it does not provide any incentive to replace the orifice mixer by a line stirred mixer, but it even teaches away from using vigorous stirring said to be generally unsatisfactory (see the top of page 2 and Reference Example 2). Despite the fact that a centrifuge is used to remove the aqueous phase of the reaction mixture (page 5, line 2), the document does not suggest at all to use such a device in the actual separation step of the purification.

4.2. Document E1 relates to a process for the removal of impurities from highly viscous polymer masses, in particular, of organic and inorganic residues from highly viscous polymeric solutions and melts (page 2, paragraph 1). According to page 3, paragraph 1, these impurities are present directly in the polymer mass or dissolved in droplets of an aqueous phase which are in turn suspended or emulsified in the viscous polymer mass.

4.2.1. In a list of polymers that can be treated accordingly (see the paragraph bridging pages 4 and 5), polycarbonate resin is also mentioned. The polymer solutions or melts preferably have viscosities of about 200 to 7000 cP (mPas).

4.2.2. According to page 4, lines 4 to 18 and Claim 1 of the citation, a polymer mass containing impurities dissolved in droplets of an aqueous phase, which are in turn dispersed in the viscous polymer mass, is at first passed through a coalescing filter bed, then the coalesced aqueous phase is removed therefrom. Thereafter the polymer mass is admixed with an aqueous stabiliser (page 8, line 3 et seq.) and the stabilised polymer mass is washed with water under high shear while maintaining the polymer mass as the continuous phase. Finally, the purified polymer mass is continuously removed. According to page 5, lines 14 to 17, the coalesced aqueous phase can be rapidly separated from the polymer mass by settling under gravitation.

It has not been a matter of dispute between the parties that the additional stabilisation step was a difference between the methods of E1 and the patent in suit.

4.2.3. If the impurities are dissolved directly in the polymer or the organic solvent phase, a washing step has to be carried out prior to coalescing the dispersion, wherein high shear is applied by a suitable device such as a line-type-impeller agitated mixer or an ultrasonic vibrating blade mixer. In this washing step, the amount of aqueous washing liquid has to be limited so as to maintain the organic polymer solution as the continuous phase. For any given mixer design, at a fixed shear rate, the energy input is determined by the effective viscosity (page 6, lines 12 to 25).

4.2.4. For further purification and removal of residual impurities, the polymer is thereafter subjected to high shear washing whereby the polymer is again maintained as the continuous phase. For this step a rotating disc contactor, Rushton or Scheibel columns are recommended, while most (other) countercurrent washing units have been found to be unsatisfactory (page 9, paragraph 2). When using the rotating disc contactor in this step, the polymer mass can be maintained as the continuous phase at a wash water/polymer mass ratio of 2 or 3 or more (page 10, lines 8 to 11); when using a countercurrent column, the said ratio in the feed streams can be less or more than about 1:1, the higher ratio being preferred (page 10, lines 21 to 23). The two phases are separated in settling zones (page 10, line 27).

4.2.5. Thereafter, one or more high shear washing steps in autoclaves or using a high shear line mixer in combination with a coalescing device can additionally be carried out (page 11, paragraph 2).

4.2.6. Only Examples VIII and IX deal specifically with the purification of polycarbonate resin. In Example VIII the aqueous phase originating from the polymerisation is first decanted after settling, the polymer solution is then pumped through a coalescing bed and the effluent is allowed to settle by gravity, and finally the aqueous layer is removed ("essentially 10% by weight of the total effluent"). In Example IX, after the water present in the polymerisation mixture has been decanted and the polymer has been washed with water and thereafter stabilised, the polymer solution is additionally treated with water in a vibrating blade ultrasonic mixer. The resulting emulsion is then broken by coalescing filtration as in Example VIII. 87.7% of the impurities are removed in this way.

4.2.7. This degree of removal of the impurities is clearly below the figures (at least 96%) obtained in the examples of the patent in suit; it rather corresponds to those in the comparative examples reported in Tables 1 and 2 of the patent in suit.

4.2.8. The document by itself thus does not provide any incentive to dispense with the pretreatment, which includes a coalescing filtration and separation, and with the stabilisation step, to select a line stirring mixer and to use a centrifugal separation as the final purification step in order to improve the purification efficiency, including an improved removal of impurities.

4.3. E3 relates to a further method to remove impurities from a polycarbonate solution in methylene chloride as obtained from the phosgene process, the impurities being incorporated with a small amount of water in the form of an emulsion (column 1, lines 15 to 22). According to Claim 1, the known method comprises the steps of (1) mixing the crude polycarbonate solution with a certain amount of aqueous washing liquid to form a water-in-oil dispersed phase, (2) adding thereto a further amount of aqueous washing liquid to cause a phase inversion of the dispersion obtained in step (1), (3) then separating the dispersed organic phase, i.e. the purified polycarbonate solution, from the continuous aqueous phase, i.e. the diluted aqueous washing liquid containing the impurities.

4.3.1. In column 1, lines 41 to 68, reference is made to problems which occur when either aqueous washing liquid is added in small amounts to form a water-in-oil emulsion, in particular under alkaline conditions and at high solids contents and viscosities of the organic solution, or when it is added in large amounts to form a oil-in-water dispersion. In the first case, the time for separating the continuous organic phase from the dispersed aqueous phase is said to be too long to be suitable for an industrial operation, in the second case, the elution of impurities from the continuous organic phase is said to be so slow that it is difficult to attain a satisfactory washing effect (see also column 3, lines 41 to 45).

4.3.2. According to the first two paragraphs in column 2, the method of E3 does not only allow to effectively remove impurities from polycarbonate solutions of high concentrations and viscosities which may even be alkaline "by the combination of certain steps", but "a satisfactory result can be attained even through the use of a simple apparatus such as a combination of stirring vessel and a settler separator".

The known method requires a thorough mixing of the two organic and aqueous solutions in both steps (1) and (2) by using e.g. a stirring vessel, a multi-plated tower or other vessels or towers having a stirrer (column 3, lines 46 to 50 and 64 to 68).

Due to the phase inversion in step (2), the phase separability of the mixture is improved, whereby the phases can be easily separated by a conventional method using a settler separator, tower type separator or a centrifugal separator (column 4, lines 1 to 7).

In the drawings as well as in the examples and references, stirring vessels (equipped e.g. with a turbine blade mixer) and settlers are shown and used, respectively.

4.3.3. In the comparative examples ("references") of the document, the unsatisfactory results of embodiments without phase inversion are demonstrated (see e.g. Table 1, References 1 and 2). These poor results in the prior art have been confirmed by Examples 7 and 8 as submitted on 20 October 1991 during the examination proceedings (Document (1) referred to in these examples corresponds to E3). Moreover, there is no hint that by modifying the method, let alone by departing from the explicit teaching of E3, even further improved results would be obtained, as shown by the above additional examples.

4.3.4. It is evident that the gist of the method of E3 is the combination of a washing step in water-in-oil dispersion and its phase inversion before the separation of the organic and aqueous phases rather than the selection of particular mixing and separation devices. It is clear from the teaching in E3 that it had to be expected that dispensing with the phase inversion would result in a very poor removal of impurities. In any case, E3 on its own does not provide any incentive to dispense with the phase inversion and to select a line stirring mixer at a specific power input to improve the purification efficiency.

4.4. Even a combination of the documents relied upon by the Appellant would not render obvious the claimed subject-matter.

4.4.1. From the above discussion it is evident that the known processes are each characterised by a combination of specific features and that in each case the desired level of purification can only be achieved when all the required conditions are fulfilled. A given feature, whether it is a device used in a particular step of the purification method or the physical state of the system, is thus only important within the framework of a specific technical context. In the absence of any indication, a skilled person would thus have no reason to depart from the corresponding teaching and, consequently, would not consider an isolated feature from one process to combine it with another method.

4.4.2. It is not disputed that the solution proposed in the patent in suit is based on well-known technical features and that a combination of features appropriately selected from the various disclosures could result in a process within the terms of the method as claimed. As pointed out by the Respondent, however, the question is not whether a skilled person could have considered such a combination at the priority date of the patent in suit, but whether this skilled person faced with the above defined technical problem would have done so with a reasonable expectation of success (cf. T 2/83, OJ EPO 1984, 265).

4.4.3. In the Board's view, the Appellant clearly failed to demonstrate that there was an incentive to modify the process known from E2 in accordance with the requirements of Claim 1 of the patent in suit in order to simultaneously improve the efficiency of purification and simplify the method. In the absence of such a link between the technical problem and the features of its solution, the Appellant's arguments can only be made with the benefit of the contribution of the patent specification and amount thus to hindsight analysis.

4.5. It follows that the method as defined in Claim 1 would not be obvious to a person skilled in the art having regard to the state of the art relied upon by the Appellant, whether considered in isolation or in combination and, therefore, involves an inventive step.

5. Claims 2 to 7, which relate to preferred embodiments of the method according to Claim 1, are supported by the patentability of the main claim and thus also allowable.