England and Wales High Court (Patents Court) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
PATENTS COURT

		July 31, 1997

Mr Justice Jacob:

1. Union Carbide Corporation ("UC") sue BP Chemicals Ltd. (BP) for infringement of two European Patents (UK), Nos. 89,691 and 241,947. They relate to the production of polymers in a fluidised bed reactor. BP is alleged to have infringed by the production of linear low-density polyethylene ("LLDPE") at their Grangemouth plant. BP has carried out such production by two methods, called for the purposes of this case, Marks 1 and 2. Initially the Mark 1 process was used. BP then changed to the Mark 2. When they made the change they sought a declaration of non-infringement, but the claim for this has been superseded by UC's claim for infringement. BP also claim a declaration of non-infringement for a variant of the Mark 2, which was called the Mark 3. BP further counterclaim for revocation of both patents. The subject matter of the patents is what has been called the "condensing mode" of operation of a fluidised bed reactor for making polymers.
2. I should say a word about the history of the applications which I hope may be of assistance to the European Patent Office. The two patents have the same priority date (24^th March 1982) and the same specification. Only the claims (and not even all of these) are different. The application for 691 was made on 23^rd March 1983. In about 1987 947 was divided out of the application for 691. I was told that one examiner said there were two inventions and this was the cause of the division. Whether that is so or not (I gather another examiner was of a different view) it is surely unsatisfactory that the consequence of the division is that the processing of the applications should diverge. As it is, they have, with 691 proceeding well ahead of 947. 691 was granted on 15^th November 1989. So its application took over 6½ years - more than long enough one would have thought, whether or not the applicant was itself dilatory. But 647 took even longer. It was not granted until 28^th October 1992 - 9½ years from application. I cannot see why the decision calling for division should have been allowed to cause that further delay. In my opinion divided out applications should be processed together. Any other way is likely to cause unnecessary duplication of effort and delay. That sort of delay is accompanied by unacceptable uncertainty for European industry.

The Witnesses

For UC

3. UC first called Mr Ryan, a Vice President and their General Manager, Speciality Polyolefins. He gave evidence about commercial matters, UC relying on commercial success and long-felt want to support the inventiveness of their patents. Their main expert (who also gave some factual evidence) was Mr Simpson. He was a Senior Corporate Fellow and currently manager of the Heat Transfer and Fluid Dynamics Group in UC's Engineering Department. He was a specialist in fluid dynamics. He was consulted by process engineers, sometimes by production and R&D personnel. He was not a part of the research team which devised the inventions the subject of the patent. They were separately located. Mr Simpson was only, briefly, consulted, late on in the project. Finally by way of oral evidence Mr Fulks was called. He was now retired but had worked as a production engineer on UC's plants for commercial production. He was concerned with scale up to full scale production. At the time of introduction of the patented process he was in charge (or close to being in charge) of two of the plants concerned. He too was not part of the research team leading to the inventions. In addition to the oral evidence there was a witness statement, accepted as correct, from a Mr Dodson. He reported the results of some measurements taken on UC's Star plant upon which others commented.

For BP

4. BP called as an expert, Professor Gift. He is a very distinguished expert in the general field of fluidised beds. He has many publications in the field and knows many of the others who have also published. After graduation he worked for ICI for a several years, and then taught successively at McGill, Naples, Imperial, Cambridge, and Surrey. He has considerable industrial experience, having been a consultant to a wide range of companies. Such consultancies have included several relating to fluidised beds. But they have not included any consultancy in relation to a fluidised bed for making polymers. He is a man of obviously outstanding ability. But that ability together with the fact that he has never worked in the field does not make him the best witness on the issue of obviousness. Next, BP called M. Raufast. During the 1970s and until 1987 he worked on polymer production in France. Initially his company was called Napthachimie snc and later he worked for BP Chimie snc which had taken over the relevant Napthachimie business. These companies were all effectively part of BP. He was part of the research department and was involved in the adoption of a fluidised bed process by Napthachimie in 1975 (in a reactor called the PZ3A), in work done in relation to that reactor, and in relation to possible further reactors. M. Raufast plainly was in the position of the skilled man at the time. He had a good command of English although there were occasional misunderstandings. Dr Newton came next. He was concerned with the development of BP's Mark 2 process and its spray nozzles. Finally there was Dr Orr who gave evidence about the operation of BP's polyethylene plant at Grangemouth.

Hearsay Evidence

5. In addition to the foregoing witnesses, each side relied upon specified material contained in the other side's discovery.

General Background

"Polyethylene is basically polymerised ethylene or a polymer produced from ethylene and one or more higher alpha-olefins. It comes in three forms which have different molecular structures and physical properties. They are low density polyethylene (LDPE) which was the original polyethylene invented in the 1940's; high density polyethylene (HDPE) and linear low density polyethylene (LLDPE). Each type is used in a variety of different applications based upon differences in their physical properties, primarily density, melt index and molecular weight distribution."

6. The foregoing is an uncontroversial passage taken from the witness statement of Mr Ryan. The starting material ("monomer") of the polymerisation reaction is a short-chain unsaturated carbon compound such as ethylene (also called ethene, C₂H₄), propylene (also called propene, C₃H₆), butene (C₄H₈), hexene (C₆H₁₂). The monomer is caused to react so as to produce long-chained polymers, the details of the reaction not being important for present purposes. Ethylene is generally used in the reaction and can be used with one of the longer monomers. The reaction needs a catalyst (again the details of this do not matter). The reaction requires an elevated temperature and pressure. The monomers are gases at normal temperature and pressure. The longer the monomer the higher its dew point for any given pressure and temperature and, correspondingly, the easier it is to liquefy. Thus, for example, hexene has a higher dew point than ethylene. One can make polymers from a single kind of monomer, or from two kinds (copolymers) or three kinds (terpolymers).
7. There were a number of established methods of causing the reaction which were well-known by 1982. Professor Clift summarises the history thus:

"Polyolefins - primarily polyethylene - were originally produced in gas phase reactors, operated at high pressure (150-300 MPa) and usually stirred. This high pressure process dominated the polyethylene industry for decades. However, it had several disadvantages, the principal one of which was the high investment and operating costs associated with operation at such high pressures. Subsequently, liquid phase polymerisation reactors were introduced, but they had their own problems. One of these was that the liquid had to be removed from the solid product.

The potential advantages of a higher productivity low pressure gas phase route to polyolefins were recognised by the 1960s. By the 1970s, UCC had developed a process which they were both operating themselves and beginning to licence to others. Napthachimie had also developed a process which they were operating themselves. Both of these processes were for the production of High Density PolyEthylene ("HDPE"). There were two aspects to the low pressure gas phase route: the development of high activity polymerisation catalysts which were both inexpensive and did not have to be removed from the polymer product, and the development of a suitable reactor system to allow the reacting "seed particles" to grow in the presence of monomer without fusion or agglomeration. Fluidised beds emerged as the devices with greatest potential due to the rapid particle mixing and temperature uniformity within the bed, lack of moving components and potential for continuous operation."

8. An alternative to fluidised beds for the reaction was a bed which was mechanically stirred - a "stirred bed" system as it was known.
9. UC's process is called "Unipol" and, as Professor Clift notes, it has been widely licensed, commencing in the 1970's.
10. I must say more about fluidised beds. Again I can borrow from Prof. Clift:

"The term fluidisation refers to a general technique, used throughout the process industries (primarily the oil, chemical and mineral sectors), in which a bed of solid material is supported by an upward flow of a fluidising fluid. The solid must be in a suitably divided form: granular particles or powder. The solid particles are supported by the fluidising fluid and are then free to move around, rather than resting on the particles below them, Under these conditions the bed displays many liquid- like properties; this is the origin of the term fluidised bed.

11. The key properties of gas fluidised beds which have led to their widespread industrial adoption (including use as polymerisation reactors) are the excellent bed mixing and heat transfer characteristics. These characteristics are due to the manner in which the fluidising gas moves through the bed. There are two types of gas flow through the bed itself:
12. Interstitial flow: some gas percolates between the particles to exert the drag necessary to maintain fluidisation; Bubble flow: the balance of the gas flow rises through the bed as voids essentially free from particles. These gas voids are known simply as bubbles because they have many features in common with gas bubbles rising through a moderately viscous liquid.
13. The gas voids or bubbles form the bubble phase and the fluidised particles between the bubbles form the particulate phase (sometimes known as the emulsion phase). As the bubbles rise through the bed, they transport some of the particulate phase and thereby maintain rapid circulation and hence good mixing of the solids in the bed. The rapid circulation of the particulate phase also means that there is good transfer of heat both between the solids and the incoming gas and also between the solids and the walls. The combination of good mixing and good heat transfer results in a uniform temperature essentially throughout the bed. In fluidisation practice, it is generally recognised that introducing a gas or a liquid which is significantly hotter or colder than the bed can give rise to a region where the temperature differs from that in the body of the bed, but that this region is localised to the immediate vicinity of the point where the hot or cold fluid is introduced."
14. I must return to the last sentence of this otherwise uncontroversial passage in due course: there is a factual dispute about the degree of localisation of the colder region in BP's plant.
15. Prof. Clift also conveniently describes a fluidised bed by reference to a diagram adopted from a standard work on the subject:



He says:

"The fluidising gas enters a plenum at the base of the process vessel. The gas is raised to the necessary pressure by a blower (3) and may be heated or cooled in heat exchanger 5. From the plenum, the gas enters the bed through a distributor (4) also sometimes known as a grid. The distributor covers the whole cross-section of the process vessel. It normally takes the form of a flat horizontal plate with passages or perforations through which the fluid passes. For some special applications, the distributor may be slightly inclined or conical, departing from horizontal by a few degrees at most."

16. The gas stream leaving the bed entrains some of the finer solid particles ("fines"). These are normally removed from the gas stream before it is passed to the next operation. The first stage of separation is to reduce the gas velocity so that the largest entrained particles fall back into the bed. This is achieved by increasing the diameter of the process vessel above the level of the bed surface, as shown in Figure 2.1. The gas stream then passes through one or more cyclones (8) which may be located inside or outside the process vessel. The cyclones are designed to induce a rotational flow to cause centrifugal separation of the entrained particles. The figure illustrates both options. Where only one cyclone is used, it may be located inside or outside the process vessel. Normal practice is to return the solids collected in the cyclone (or the first cyclone when two stages are used) directly and continuously to the bed as shown in the figure. The amount of solids carried out of the reactor increases as the velocity of the fluidising gas is increased. Ultimately, this sets the upper limit for the fluidising velocity.
17. Some processes require solids to be fed into and removed from the fluidised bed. The figure shows, hypothetically, a screw feeder (9) and a solids take-off (10). For some processes, a liquid or solid/liquid suspension is introduced to the bed. In the figure this is shown as a spray feed (11) onto the surface of the bed. The particular means of introducing the liquid shown in the diagram was probably selected since it is more graphic than the others. However, other means of introducing the liquid were known."
18. In the use of fluidised beds for polymer production the bed consists of the polymer itself, caused to grow from small particles. The monomers are used as part of the fluidising gas (other gases may be included). The catalyst is fed into the reaction chamber. In commercial production the process is continuous, the chamber having an outlet port from which product is drawn. Only a minor proportion of the monomer reacts before the gas comes out of the top of the bed. Unreacted monomer is recycled. The fines coming off with it will include catalyst and so the reaction will or may continue to some degree outside the bed itself. It is important to note that the polymerisation reaction is exothermic. So there is a need to remove heat. Self-evidently the faster the reaction, the greater is there a need of heat removal. At "slow" reaction speeds heat removal is not a problem. The cooling of the recycle gases, together with losses from the reaction chamber are enough. Thus heat removal only becomes a problem once one can get the reaction to go fast enough. The restraint on production rate caused by insufficient speed was called the "kinetic restraint"; the restraint caused by problems of heat removal the "heat removal restraint".
19. Professor Clift explained some of this as follows:

"The rate of production of solid polymer in the reactor is commonly expressed as the space-time yield, i.e. the mass produced per unit volume of reactor per unit time (e.g. measured in kg m-3 hr -1). The space-time yield depends on the operating conditions in the reactor - temperature, pressure and concentration of monomer or comonomers - but these are normally constrained by process requirements. It also depends on the activity and concentration of the catalyst. Increasing the catalyst activity either increases the space-time yield and hence the rate of production which can be obtained from a reactor of given volume or alternatively decreases the concentration of the residual catalyst in the product. The catalyst activity and concentration do not affect the heat of polymerisation. Therefore using a higher activity catalyst to increase the space-time yield has the effect of increasing proportionately the rate at which heat must be removed to maintain constant temperature in the reactor."

Patent 691 and its construction

20. The specification begins as follows:

"This invention relates to a novel method for producing polymers improving the space time yield of an exothermic polymerization reaction conducted in a fluidized bed reactor.

The discovery of the fluidized bed process for the production of polymers provided a means for producing these diverse and widely used polymers with a drastic reduction in capital investment and a dramatic reduction in energy requirements as compared to then conventional processes. The present invention provides a means for even greater savings in energy and capital cost by affording a simple and efficient means for obtaining a substantial increase in production rate in a given size reactor over what was previously possible in a fluidized bed process.

The most common and perhaps universal means of heat removal employed in conventional fluidized bed reactor processes is by compression and cooling of the recycle gas stream at a point external to the reactor. In commercial scale fluidized bed reaction systems for producing polymers such as polyethylene, the amount of fluid which must be circulated to remove the heat of polymerization is greater than the amount of fluid required for support of the fluidized bed and for adequate solids mixing in the fluidized bed. The fluid velocity in the reactor is limited to prevent excessive entrainment of solids. A constant bed temperature will result if the heat generated by the polymerization reaction (which is proportional to the polymer production rate) is equal to the heat absorbed by the fluidizing stream as it passes through the bed, plus any heat removed or lost by other means."

21. The specification then sets out matter which was challenged by BP. It asserts:

"Unfortunately, it has long been believed that the recycle gas temperature could not be lowered any further than to a point slightly above the dew point of the recycle gas stream. The dew point is that temperature at which liquid condensate begins to form in the gas stream. Common practice has been to limit the temperature of the recycle stream at the outlet of the cycle heat exchange zone to a temperature at least about 3 to 10°C its dew point (see EP Specification No. 0 021 605, page 22, lines 8 - 22, [a patent called "Fraser"]). This assumption was predicated on the belief that the introduction of liquid into a gas phase fluidized bed reactor would inevitably result in plugging of the distribution plate, if one is employed; non-uniformity of monomer concentrations inside the fluidized bed and accumulation of liquid at the bottom of the reactor which would interfere with continuous operation or result in complete reactor shut-down. For products, such as those using hexene as a comonomer, the relatively high dew point of the recycle stream has until now severely restricted the production rate.

The primary limitation on reaction rate in a fluidized bed reactor is the rate at which heat can be removed from the polymerization zone."

22. The assertion made is thus that there was a prejudice against allowing the temperature of the recycle stream to fall below the dew point of the recycle gases. The cited passage of Fraser spells this out in more detail. I will quote it later when I deal with anticipation. Whether there really was such a prejudice within UC or within BP, and whether if so it can properly be attributed to the skilled man is a live issue.
23. 691 continues by referring to prior art proposals for heat removal in stirred beds and nothing turns on most of these. I should however mention two acknowledgements.

"In U.S. Patent No. 4,012,573 (Treschman et al.) gases withdrawn from a stirred reactor are condensed to liquid and returned in liquid form to the stirred reactor where the liquid is brought into desired contact with polymer in the stirred bed.

Mitsubishi Petrochemical Co. has proposed the use of liquids or regasified liquids for cooling in a gas phase reactor (J55/045,744/80 and DT 2 139 182). In both of these descriptions the liquid or regasified liquid is injected into the bed rather than entering with the fluidizing gas as in the present invention. DT 2 139 182 is specific to stirred beds rather than fluidized beds. In J55/045,744/80 the liquid is regasified before being injected into the fluidized bed."

24. So the patent is here recognising as prior art the idea of direct injection of liquid (recondensed monomer) into a stirred bed. It also recognises the direct injection of regasified liquid into a fluidised bed. The italicised words are of some importance when one comes to construction. The patent goes on to say:

In a fluidized bed reaction system, as distinguished from stirred or paddle-type reaction systems, uniform distribution of monomer and catalysts in the upwardly moving gas stream is essential to avoid hot spots and resulting polymer chunks. In stirred and paddle-type reactors these problems are overcome by mechanical stirring and agitation. A further requirement of a fluidized bed reactor system is that the velocity of gas flowing through the reactor be adequate to maintain the bed in a fluidized state. The gas velocity required to keep the bed in a fluidized suspension cannot be achieved under normal conditions by mere injection of liquid at the bottom of the bed. Therefore, the direct liquid injection cooling of a reactor, as described by Treschman et al. is not a viable option for a fluidized bed reaction system.

25. The specification goes to the "subject matter" of the invention which is effectively claim 1:

"A continuous process for the production of polymer in a fluidized bed reactor from one or more fluid monomers by continuously passing a gaseous stream through said reactor in the presence of catalyst under reactive conditions, withdrawing polymeric product and unreacted fluids, cooling said unreacted fluids and returning said cooled fluids into said reactor together with sufficient additional monomers to replace those monomers polymerized and withdrawn as product, which comprises:

cooling part or all of said unreacted fluids to form a two-phase mixture of gas and entrained liquid below the dew point and

reintroducing said two-phase mixture into said reactor."

26. The portion before the emboldened words is a fair summary of the well-known prior art as used in the original Unipol and Napthachimie processes The key part of the claim is the emboldened words. There is a dispute as to their scope: is the claim limited to the re-introduction of a two-phase mixture which consists of a two-phase mixture of gas and entrained liquid which is that resulting from the cooling? What if the unreacted fluids are cooled below the dew point, the liquid is separated out completely from the gas, and then the gas is introduced into the reactor separately from the liquid, the latter being introduced either in a two phase mixture made from a freshly introduced make-up gas (Mark 2) or simply injected via an atomising nozzle (Mark 3)?
27. The dispute relates only to the Marks 2 and 3 processes, it being admitted by BP that the Mark 1 process fell within claim 1 (and claim 2) even on the narrow construction.
28. Before resolving this point of construction it is necessary to read on into the patent. I can omit portions which condescend to the various types of polymerisation reaction in which it can be used, pausing only to note that the invention is not limited to any specific exothermic polymerisation reaction. It then says:

"In very general terms, a conventional fluidized bed process for producing resins, particularly polymers produced from monomers, is practised by passing a gaseous stream containing one or more monomers continuously through a fluidized bed reactor under reactive conditions and in the presence of catalyst. The gaseous stream containing unreacted gaseous monomer is withdrawn from the reactor continuously, compressed, cooled and recycled into the reactor. Product is withdrawn from the reactor. Make-up monomer is added to the recycle stream.

The polymer-forming reaction is exothermic, making it necessary to maintain in some fashion the temperature of the gas stream inside the reactor at a temperature not only below the resin and catalyst degradation temperatures, but at a temperature below the fusion or sticking temperature of resin particles produced during the polymerization reaction . This is necessary to prevent plugging of the reactor due to rapid growth of polymer chunks which cannot be removed in a continuous fashion as product. It will be understood, therefore, that the amount of polymer that can be produced in a fluidized bed reactor of a given size in a specified time period is directly related to the amount of heat which can be withdrawn from the fluidized bed.

In accordance with this invention the recycle gas stream is intentionally cooled to a temperature below the dew point of the recycle gas stream to produce a two-phase gas-liquid mixture under conditions such that the liquid phase of said mixture will remain entrained in the gas phase of said mixture at least from the point of entry into the fluidized bed reactor until volatilized or until passage into the fluidized bed. A substantial increase in space time yield results from the practice of this invention with little or no change in product properties or quality. When practiced as described herein the overall process proceeds continuously and smoothly and without unusual operational difficulties.

29. The point of this deliberate use of liquid in a two-phase mixture is to make use of the latent heat of vaporisation of the liquid. Heat is not only removed because the entering gas is colder than the main reactor temperature (as in the prior art) but also because the entering liquid is not only colder but also requires heat to turn it into vapour. Heat removed by actual warming is called "sensible heat": it involves a change in temperature and hence is in principle measurable with a thermometer. Heat removed by vaporisation does not result in a temperature change.
30. The patent continues:

"A primary limitation on the extent to which the recycle gas stream can be cooled below the dew point is in the requirement that gas-to-liquid ratio be maintained at a level sufficient to keep the liquid phase of the two-phase fluid mixture in an entrained or suspended condition until the liquid is vaporized. It is also necessary that the velocity of the upwardly flowing fluid stream be sufficient to maintain the fluidized bed in a suspended condition.

While the liquid content of the two-phase recycle stream can be quite high, as general rule, the quantity of condensed liquid contained in the gas phase should not exceed about 20 weight percent and preferably should not exceed about 10 weight percent, provided always that the velocity of the two-phase recycle stream is high enough to keep the liquid phase in suspension in the gas and to support the fluidized bed within the reactor.

31. The 20% liquid content of the two-phase recycle stream is by weight, not by volume. Mr Simpson explained that in practice the two phase system took the form of a fine mist of liquid:

"To explain what is actually thought to be happening in the recycle line, condensation occurs on the walls of the tubes of the cooler and the condensed liquid is stripped off by the gas stream in the form of very fine mist. The gas velocity in the tubes of the cycle gas cooler is typically in the range of 45 to 55 ft/sec (14 to 17 m/sec). Any liquid remaining as a film is re-entrained at the ends of the tubes of the cooler where the two-phase mixture flows into the outlet head of the cooler. The mist particles would be extremely fine and would not be individually visible to the eye."

The specification then says:

"The entry point for the two-phase recycle stream should be below the fluidized bed (polymerization zone) to ensure uniformity of the upwardly flowing gas stream and to maintain the bed in a suspended condition. The recycle stream containing entrained liquid is introduced into the reactor at a point in the lower region of the reactor and most preferably at the very bottom of the reactor to ensure uniformity of the fluid stream passing upwardly through the fluidized bed."

32. Now thus far the description has entirely been confined to a two-phase stream, a stream resulting from the intentional cooling to below the dew-point. The liquid "remains entrained" in the gas phase of the mixture. That is why the specification says you should not have more than 20% by weight of entrained liquid - an observation that only makes sense if you are considering the two-phase mixture of the cooling process. Similarly the language speaks of a "primary limitation" on the cooling. You must not cool so much that you get too much liquid to be entrained. And the entry point spoken of is the entry point of the "two phase recycle stream".
33. The specification continues still on the basis that one is dealing only with the recycle two phase stream - the "mist" produced when the cooling is taken down below the dew point.

"The two-phase mixture can be divided into at least two separate streams at least one of which is introduced into the reactor, e.g. at a point below the fluidized bed and one or more of said streams is optionally introduced into said fluidized bed, provided that the velocity of all gases introduced below said fluidized bed is at least sufficient to support said bed and to maintain it in a fluidized condition. In all cases the composition of the gas stream is kept essentially uniform and flowing in a manner such that there are no dead spaces in the bed where unremovable solids can form.

34. All this is saying is that you can divide the mist into two parts, one part of which is injected directly into the bed and the other below (i.e. into the plenum). There is a warning that you need enough below to keep the bed fluidised. This passage is the foundation for claim 11 which reads:

"A process according to claim 1 wherein

said two-phase mixture is divided into at least two separate streams at least one of which is introduced into the reactor at a point below the fluidized bed and

one or more of said streams is optionally introduced into said fluidized bed,

provided that the velocity of all gases introduced below said fluidized bed is at least sufficient to support said bed and to maintain it in a fluidized condition."

35. Claim 2 is for the case where the mixture is all introduced below the bed:

"A process according to claim 1 wherein said two-phase mixture of gas and entrained liquid is introduced into said reactor at a point below the fluidized bed."

36. I now come to a passage on which UC rely:

It will be apparent that if desired, it is possible to form a two-phase fluid stream within the reactor at the point of injection by separately injecting gas and liquid under conditions which will produce a two-phase stream. Little advantage is seen in operating in this fashion due to the added and unnecessary burden and cost of separating the gas

and liquid phases after cooling. It may however, be desirable to inject make-up monomer into the reactor in this fashion. The injection of liquid or gaseous make-up monomer at the point of entry of the two-phase recycle stream or elsewhere in the reactor or in the recycle stream is contemplated by this invention. Additional quantities of one or more monomers can be added to replace monomers polymerized and withdrawn as product are introduced into the recycle stream prior to entry of the said two-phase mixture into the reactor."

37. UC rely on the italicised words to say that it is here made clear that the "said two phase mixture" of claim 1 need not be just the mist resulting from the cooling. They say that the skilled man would realise that that the critical aspect of the invention is the cooling to below the dew point and use of liquefied monomer to aid cooling. He is taught he can use such monomer provided it is somehow entrained in the gas when it enters the bed. It does not matter if it stays entrained after cooling or is separated out and re-entrained. Nor, if it is re-entrained, does it matter that it is re-entrained by gas other than recycled gas. They submit that BP's construction is over-literal.
38. BP say that the language of the claim is limited to the case where it is essentially the mist of the cooling process which is reintroduced into the reactor. It is that the said two-phase mixture which must be reintroduced. They say that to regard the introduction of separated out liquid dissociated from its gas phase as re-introduction of "said two phase mixture" is more than language can bear. Moreover they say that it is their construction which is consistent with the inventor's purpose to be gathered from the body of the specification. That is why he refers to the "primary limitation", that is why uses the expression "remain entrained", that is why he distinguishes Mitsubishi and in so doing says entering with the fluidizing gas, that is why he refers to the "entry point of said two-phase stream". Separate introduction of the gas and the liquid is consistent with none of these things.
39. BP further say that the italicised passage goes no further than to say in effect, "it makes no difference if you separate out and then re-entrain in the gas before introduction. You can do this in the reactor but there is not much point". The reference to the "reactor" does not mean the bed at this point. And this is made clear by claim 13 which reads:

"A process according to claim 1 wherein said additional monomers added to replace monomers polymerized and withdrawn as product, are introduced into the reactor in liquid form, below the fluidized bed and at or near the point of entry of said two-phase mixture into said reactor."

40. This is the only claim which refers to the introduction of liquid at all. It is limited to the case where the make-up monomers are added below the bed - in effect added to the two-phase mixture. I should say that UC's final argument reads claim 13 as "below ..or at or near the point of entry". A small word makes a big difference.
41. I think BP are right. This inventor's purpose is to reintroduce the two phase mixture produced by cooling below the dew-point. It is at least part of that mixture which is reintroduced as the fluidising gas. The inventor did not envisage and did not seek to encompass a separation of the liquid such that the two-phase mixture was not reintroduced. He would not have said any of the things referred to by BP in their argument (e.g. the "primary limitation") and he would not have said, as he did at the end of the patent:

"Advantages of this invention over other heat removal methods which use condensed liquids include the simplicity of this invention, good liquid distribution, maximum heat removal, and uniform gas stream composition in the polymerization zone as a result of introducing the two-phase recycle stream in the lower regions of the reactor. No liquid collection, separation, or injection equipment is required and the liquids are dispersed in the stream entering the fluidized bed over the entire cross- sectional area, avoiding temperature gradients and high local concentrations of liquid around injection means.

42. It is the two-phase recycle stream which achieves the inventor's purpose. And no injection equipment is required so that that there are no local concentrations of liquid. BP Mark 2 (and 3) has both of these.

Infringement of 691

43. Infringement of claims 1 and 2 by the Mark 1 process stands admitted. I think noninfringement by the Marks 2 and 3 follows from my construction of the claim. The process is confidential. However I think I can say enough of the detail to explain what is happening without disclosing any confidential matter. The unreacted gases go first to a cyclone which separates the fines. These are returned to the reactor. The gases are cooled to below the dew-point so as to condense the monomer (4MP-1 or hexene-1 as the case may be). Make-up monomer, and, where necessary, hydrogen and nitrogen, are added. The mixture goes to a separator (accepted by UC for the purposes of this action only) to be 100% efficient. The gas (dry) enters the plenum (i.e. below the grid) and is the fluidising gas. The liquid enters via a nozzle 66cm above the plenum. It is atomised within the specially designed nozzle, which also introduces fresh ethylene as both a make-up and assist gas. The nozzle is specially designed and patented. The atomised liquid and ethylene does not have the same constitution as the original mixture.
44. UC say there is infringement because gas and the liquid are both reintroduced. Moreover the liquid is entrained and, when it meets the gas, in effect the two-phase mixture is re-created. However I do not think that what is going here can fairly be described as the "re-introduction of said two-phase mixture" (of gas and entrained liquid formed by cooling).
45. As for the Mark 3 process, BP propose the use of an atomising nozzle without the use of the assist ethylene. The liquid is sprayed directly into the bed. If I am right about the Mark 2, this cannot infringe and BP are entitled to the declaration sought.

Patent 947 and its construction

46. The body of the specification is practically the same as that of the first patent. I can go straight to claim 1, which, it is agreed, is the only claim of significance. It reads:

A method for controlling the temperature of a gas fluidized bed during the production of polymer in a fluidized bed reactor by an exothermic polymerization reaction, which comprises:

continuously introducing into the bed a stream comprising unreacted monomer gas cooled to below the maximum desired temperature within said bed and

simultaneously or separately introducing a stream of liquid into said reactor under conditions such that an essentially uniform mixture of said gas and liquid is introduced into

said bed at a level below the region of maximum desired temperature within said reactor wherein said liquid is vaporised.

47. There was no dispute over the meaning of the emboldened words. They refer to the essentially isothermal region of the fluidised bed. There was a factual dispute as to whether or not the 66cm height of the BP Mark 2 and 3 nozzles is high enough to be within that region.

Batch Process?

48. So far as construction is concerned, BP raise what I very strongly suspect to be a sterile argument. They contend that the claim covers a batch as well as continuous process. UC say the claim only covers a continuous process. BP wish to argue that to the extent that the claim covers a batch process it covers some Solvay prior art and thus lacks novelty. But BP in fact carry out a continuous process. So, if they were right on construction and the claim were invalid accordingly, much the most likely result of the point would merely be that UC would be given leave to amend down to a continuous process with minor (on the scale of this case) consequences as to costs.
49. The argument runs thus. The only reference to "continuous" within the claim is to the continuous introduction of recycled monomer. This happens in both batch and continuous processes. So the claim covers both. The argument is reinforced by reference to claim 3 (which it is agreed has the same scope as claim 1 of the 691 patent). Claim 3 reads:

"A process for producing polymer from monomer by an exothermic polymerization reaction in a gas fluidized bed reactor having an upper polymerization zone containing a bed of growing polymer particles and a lower gas diffusion zone, which comprises:

(1) continuously passing a gaseous stream comprising monomer through said polymerization zone with an upward velocity sufficient to maintain said particles in a suspended and gas fluidized condition;

(2) continuously introducing a polymerization catalyst into the reactor for introducing it into said polymerization zone;

(3) withdrawing polymer product from said polymerization zone;

(4) continuously withdrawing a stream of unreacted gases comprising monomer from said polymerization zone, compressing and cooling said stream to a temperature below the dew point of said stream to form a mixture comprising a gas phase and a liquid phase; and

(5) continuously introducing said mixture into said polymerization zone wherein said liquid phase is vaporized."

50. DP contrast the repeated use of "continuously" in this claim with its single use in claim 1.
51. UC say that the single use of the word "continuously" is enough to make it clear that the claim is restricted to a continuous process. Moreover, they say, the whole specification clearly describes, and is concerned only with, a continuous process. The problems addressed by the patent are to stop reactor shutdown, something which happens anyway in batch process. I think UC are right. The skilled man, reading the claim purposively, would be astonished to think it related to a batch process. He would read the word "continuously" as meaning a reference to a steady-state introduction of recycled monomer, not the introduction until the batch was finished.

"Essentially uniform mixture of said gas is introduced"

52. The other dispute of construction was more important. Do these words mean that the mixture must be formed before it is introduced? BP say yes - for if you introduce the liquid directly into the bed, you could not be introducing an essentially uniform mixture of gas and liquid. This must be right. Further there is the question of what constitutes the mixture. Can it be of liquid and any gas, or must it be liquid and recycled gas?
53. BP argue that in practice to fall within the claim the mixture would have to be formed in the plenum. 1 do not see why - provided you have enough gas coming through the plenum I do not see why a mixture could not be introduced above. The evidence does not support the proposition that the only way to get a uniform mixture introduced is via the plenum. But I do think the patentee only contemplates the introduction of an already mixed two-phase mixture, the gas of that mixture being unreacted monomer. I think the real question is what does the claim mean when it speaks of unreacted monomer? For all monomer is, by definition, unreacted. What I think the skilled man would understand by the qualification is that this is monomer which has had the chance of reacting but did not react - i.e. recycled monomer. That makes entire sense with the general teaching of the patent which is all about cooling (actually to below the dew point though claim 1 is not so limited) the recycled unreacted monomer. Actually the patentee is fairly explicit about equating unreacted monomer with the recycle stream. He says, for instance^[1]:

"The portion of the fluidizing stream which does not react in the bed constitutes the recycle stream"

54. All the teaching about a primary limitation and so on which applies to the first patent must also apply to this claim. And the word must have some meaning - one cannot just disregard it.
55. So, in my opinion, the essentially uniform mixture of said gas and liquid is a mixture of liquid and recycled monomer gas. It does not extend to a different monomer gas.

Infringement of 947

56. It is accepted that the Mark 1 process infringes claim 3, which is by common consent the same as claim 1 of 691. For the reasons I have already given, neither the Marks 2 or 3 infringe that claim. As to claim 1, the Mark 1 process clearly infringes too.
57. Turning to the Mark 2 process, on my construction of claim 1 BP do not introduce a uniform mixture of said gas and liquid. They introduce a different gas - something the patentee had not contemplated. It is no good saying it does not matter whether the gas is recycled or not: that ethylene is ethylene is ethylene. So it is, but the patentee's vision did not extend that far and he limited his claim to the use of "said gas" because that was as far as his purpose extended. As for the Mark 3 process, this cannot infringe - no mixture of gas and liquid is introduced at all because the liquid goes in as such.
58. There is another, formidable, reason why the Mark 2 process does not infringe claim 1. As I have said the point of introduction from the nozzle is 66cm above the grid. I find on the evidence that this is well above the point where the bed becomes isothermal. That point is about 30cm - a foot - above the grid. I say this for a number of reasons.

(1) The patent itself expects this point to be at about this height. It says:

"The temperature of the bed is controlled at an essentially constant temperature under steady state conditions by constantly removing the heat of reaction. No noticeable temperature gradient appears to exist within the upper portion of the bed. A temperature gradient will exist in the bottom of the bed in a layer of about 15 to 30 cm (6 to 12 inches), between the temperature of the inlet fluid and the temperature of the remainder of the bed."

This passage is not and does not purport to be limiting the height of the gradient. It is stating what the patentee believed it to be.

(2) It is here reflecting the expectation of the ordinary skilled man. Such an expectation was to be found in basic chemical engineering works such as Kirk Othmer and Perry, works which every chemical engineer has at his side. Perry, for instance, speaks of a distance of 1-3 inches below the isothermal region.

(3) Dr Newton of BP who designed the Mark 2 had this expectation. He was eminently sensible about this. He was working with a reactor which had a thermocouple at 25cm. The next thermocouple was at 6m. What is significant is that within 25cm the temperature rise recorded is 97% of the rise to the isothermal temperature as measured at 6m. Clearly there is nothing like a gradual rise going on and it would be wrong to assume that the further 3% occurs linearly or even approximately so.

(4) Mr Simpson was of the opinion that the height was much greater. He considered that this bed was unlike other beds because the rush of gas was so much greater. He agreed with Professor Clift that immediately above the grid holes there would be jets of fluid unmixed with particles. These jets would essentially be at entry temperature. He says the jets are much longer than does Prof Clift, who says they are about 12cm. Above that says Mr Simpson there is a chaotic region which is 75% gas by volume. So, he says, it will take longer before bubbles form. It is the bubbles which are involved in heat exchange and are responsible for the isothermic nature of the bed. Further, because there is so little solid matter present (25%) there is not enough to cause heat exchange - the particles will cool down. Mr Simpson relied on a couple of academic papers in relation to his estimate of jet length. Otherwise he pointed to no published material to support his conjecture.

(5) In my opinion Prof. Clift demolished this conjecture of Mr Simpson. He had considerable knowledge of the academic papers concerned, and explained that the authors, whom he knew, had made an error in thinking that their particular method of observation of jet length (involving a glass plate within a jet) did not itself extend that length. Moreover the results of the two papers did not agree. Prof. Clift got his estimate of jet length from a correlation devised by Merry and gave forceful evidence that this in practice worked. He further showed from an X-ray video that one could see the jet length approximately, a length which corresponded to that predicted by Merry's formula. He also thought you could see bubble formation above the jets. Whether that was so or not, he further pointed out that Mr Simpson, whilst right in saying that there was 75% by volume of gas, had overlooked the fact that the solid present - particles - were of the order of 30 times as dense. So there was much more hot particle material by mass than gas, leading to fast warming. He pointed out that the particles were in no way localised to the region of introduction. On the contrary they were fast moving into and out of it, so an individual particle would not cool down. Finally Mr Simpson's conjecture was simply inconsistent with the very sharp rise in the first 25cm.

(6) Mr Simpson also sought to rely upon some results from the UC Star plant in the USA. This had thermocouples in much the same places as BP. But the results did not to my mind support his thesis. He put in some graphs joining the measured temperature at 25cm with that at 6m. This point joining was a useless exercise. It could only mislead if intended to show the actual temperature variation with height. The graphs also mislead in that only the last few degrees are shown on the temperature axis, thus masking the very sharp rise in the first 25cm. Prof. Clift thought that the Star results supported him. In any event the Star reactor has different dimensions and operating conditions. Moreover BP had no means of checking or verifying the Star results even though the experiments were done for the purpose of this action. I think they were of no value.

59. I think the evidence is overwhelming to the effect that BP's height is well above the point where the bed goes essentially isothermal. There is not a mere failure on UC to prove infringement. BP have proved non-infringement. The latter point is significant so far as BP seek a declaration of non-infringement of the Mark 3 process.

Anticipation of 691

Lack of Novelty over Fraser

60. This is attacked on the basis of anticipation of claims 1 and 2 by Fraser. It is agreed that claim 3 of 947 will fall also if the attack succeeds. The attack applies on BP's narrow construction of the claim.
61. Fraser is a UC patent, published in January 1981. It relates to the Unipol process and in particular the manufacture of high tear strength ethylene-based hydrocarbon terpolymers. These are made in a fluidised bed. An express object of Fraser is:

"to provide a continuous process for making such polymers under gas phase conditions with one or more C₅-C₈ comonomers which must be recycled without being allowed to condense out."

62. The general idea is to make a terpolymer rather than copolymer because a lower dew point is involved. The details do not matter. What matters is the following statement:

"Attempts to make film grade resins having optimum physical properties in the form of copolymers made from ethylene and hexene-1 in the gas phase fluid bed polymerization processes disclosed in these recently filed United States patent applications however, have encountered problems, when C₅ - C₈ comonomers are employed. The polymerization reaction involved in an exothermic reaction. In order to conduct the reaction continuously in a commercially feasible manner it is necessary to recirculate the unreacted monomers and to remove excess heat (of reaction) from the recycled monomers in a heat exchanger. Where C₅ to C₈ comonomers are used these monomers tend to condense out in the reactor system and lead to

(i) the production of sticky or tacky polymers which are difficult to handle and process,

(ii) the plugging of the gas distribution plate used in such fluid bed reactors,

(iii) the fouling of sensing elements inserted in the reactor system for the purpose of monitoring and controlling the reaction conditions, and

(iv) reactor temperature control problems due to heat capacity variations."

63. It is said, broadly, that this passage in effect teaches condensing mode. True it counsels strongly against it, but nonetheless the idea of condensing liquid out in the reactor system is there. Putting it another way, you could only get the problems referred to (though it is not admitted that you do) if you condense in the system and in particular (to get plate plugging) condense below the grid.
64. A prior document is novelty-destroying if it contains "clear and unambiguous instructions to do what the patentee has invented" (the language of General Tire^[2]). The EPO uses equivalent language "whether this feature may be derived directly and unmistakably", e.g. Scanditronix/Ion chamber.^[3] Assume for the purpose of argument that Fraser actually recited claim 1 but added what it does say about the four difficulties for C₅-C₈. It would in effect be giving clear and unambiguous directions not to do what is claimed. I do not think that such a direction is novelty destroying. The information given by a direction not to do X because it will have adverse consequences is not equivalent to a direction to do X because it has beneficial consequences or does not have the supposed adverse consequences. As Lord Westbury put it in Hills v Evans^[4] the information given by the prior publication must be equal to that given by the subsequent patent. The information in those two statements is not equal. Invention can lie in finding out that that which those in the art thought ought not be done, ought to be done. From the point of view of the purpose of patent law it would be odd if there were no patent incentive for those who investigate the prejudices of the prior art.
65. In so holding I am in complete agreement with the Technical Board of Appeal in relation to the opposition to 691. They said:

"In the Board's view, however, novelty would only be destroyed if the continuous operation in the condensed mode as claimed in the patent in suit were unambiguously disclosed in [Fraser] as a feasible option. The indication in [Fraser] that C₅ to C₈ comonomers tend to condense out and the warning against the consequences of such a condensation is not equal to the description of an operational process as disclosed and claimed in the patent in suit. [Fraser] discloses in this respect only that condensation in the reactor system should be avoided and also some reasons on which this warning is based, but there is no indication that a continuous operation in the condensed mode was a feasible option."

66. I think there is another reason for rejecting Fraser as an anticipation. The claim requires that a two-phase mixture of gas and entrained liquid be formed and reintroduced into the reactor. Fraser is not explicit about the formation of an entrained two-phase mixture. Saying the monomers "tend to condense out" is not the same thing. The liquid may or may not be entrained.

Lack of Novelty over Resine (UK Patent 1,110,566)

67. BP only allege anticipation by this patent (published in 1968) if claim 1 covers the separation of a two-phase mixture and the separate injection of the liquid anywhere in the reactor provided the liquid is re-entrained somehow. Since I have held the claim to be narrower than this strictly I need not consider this point. However I will do so because argument was directed to anticipation on the basis of UC's construction and because Resine must also be considered in the context of alleged obviousness of claims 1 and 2 of 947.
68. Resine discloses a small-scale fluidised bed reactor. His diagram is as follows



Gas leaving the reactor is cooled at E1 - but not so there is condensation. The flow is split into two streams (3,4). Stream 3 is compressed by P1 and reintroduced, as a fluidising gas. The other stream (4) is cooled by E2 and enters separator S. The gas from S goes back into the main recycle gas loop. The liquid is fed into R2 and formed into a slurry with catalyst. The slurry is pumped into the reactor via 11.

69. Resine regards his process as dry. He says:

"We have now developed a method of carrying out the polymerization of normally gaseous olefins and other vinyl-type monomers in the complete absence of a liquid phase, thereby eliminating all the drawbacks due to the use of liquid reaction media."

Thus although he introduces some liquid in the slurry he does not regard this as a significant amount. He recognises the cooling effect of the slurry, saying:

"Owing to the higher temperature and lower pressure the liquid suspension medium having the catalyst suspended therein evaporates on entering the reactor Rl and absorbs heat."

And

"As mentioned above the liquid suspension medium entering Rl evaporates leaving the solid polymer particles with the fresh adsorbed catalyst in an extremely finely sub-divided condition, suitable for growth.

Evaporation of the liquid improves the heat control in the region of the inlet for the catalyst, that is, at the portion of the reactor where the reaction is most exothermic.

The quantity of the inert liquefiable gas and its ratio to the other inert gases which may be present and to the olefin monomer to be polymerized may be varied within fairly wide limits.

An essential requirement in this connection is for the quantity of liquefiable inert gas to be such that a sufficient volume of liquid is available for preparing the catalyst suspension in R2."

70. I do not think that Resine falls within the claim even on the wide construction. There is no disclosure of entrained liquid at all. The slurry is just let into the bed. Moreover the amount of liquid let in is obviously very small. Although there is no specific lower limit set by the claim (and condensation of only part of the unreacted fluids is contemplated) I doubt whether the skilled man would regard the trivial amount of liquid here as within the claim.
71. There is no anticipation by Resine even on UC's wide interpretation of the claim.

Anticipation of 947

Fraser

72. The principal reason for rejecting Fraser as novelty-destroying of 936 applies to claims 1 and 2 of 947. Moreover it is impossible to read into Fraser even a specific direction not to introduce an essentially uniform mixture of gas, nor as to where that should not be introduced. The case on anticipation by Fraser, whatever the construction, fails.

Resine

73. If I am right on my construction of 947, Resine clearly does not disclose the introduction into the bed of a "uniform mixture of said gas and liquid". And this is so even on UC's construction. No mixture of gas and liquid is introduced. Anticipation fails in relation to claims 1 and 2.

Solvay

74. In 1974-5 three UK patents of Solvay, a well-known Belgian chemical company, were published under Nos. 1,363,566, 1,370,078 and 1,398,695. They all relates to the same work for the production of PVC. It is common ground that, because of internal cross-references, they can be read as one.
75. Solvay discloses only a batch process. The specific examples are on a laboratory scale. Thus if I am right on my construction of claim 1 of 947, there is no anticipation. I should here note that BP's argument that scale-up of Solvay would inevitably lead to adoption of a continuous process is, to my mind, irrelevant on the question of anticipation which is confined to what Solvay actually disclosed. Likewise there is no anticipation of claim 3 which is admittedly to a continuous process.
76. If I am wrong about claim 1, then Solvay must be considered further. Since it is relied upon also for obviousness, this I will do. The relevant diagram (from 078) is as follows:



Gaseous monomer (vinyl chloride) is fed along 5 to the fluidised bed 2. Liquid monomer is stored in 14 and introduced to the bed via 8. Ex. 1 of 078 describes what happens:

"The reaction begins immediately and the resulting temperature increase is recorded by gauge (21) and transmitted to regulator (10) which operates gate valve (9) so that it opens. This results in the temperature being reduced back to the required level by injection of the liquid monomer mixture from reservoir (14) through pipe (8) into the middle of the fluidised material. The mixture of liquid vinyl chloride and ethylene vaporises immediately in the reaction medium resulting in the removal of the polymerisation heat."

77. "The middle of the fluidised material" is a positive direction to feed into the isothermal zone, it being common ground that the low temperature region is only near the bottom of the reactor. So there is no introduction of anything (still less a mixture of said gas and liquid) at a level below the region of maximum desired temperature. There is no anticipation by 078.
78. BP therefore turn to 965. This is about nozzles for injection of the liquid monomer into the bed. The patent also deals with where the injection is to be. It says:

"In order to obtain optimum thermal control of the polymerisation reaction, the at least one sprayer should be placed when operating during polymerisation in the polymerisation reactor such that the liquid monomer is sprayed from a height which is always between 0 and 75% of that of the fluidised solid material in the reactor and preferably between 15 and 70% of this height."

79. The argument is that the reference to 0% height is a reference to grid level which is, it is common ground, below the specified level. So also is a small portion above zero of the 0-75% range. Thus it is said there is injection of monomer at the specified level within the much larger range of Solvay. Thus, the argument runs, there are clear and unambiguous directions inter alia to come within that range. This is so even though Solvay's preferred range (15-70%) is outside the claimed range.
80. I think this argument is right. A prior disclosure of a range should normally be regarded as disclosing each and every part of the range. However if there is something special about a later claimed part of the range, there may be room for invention - an invention along the lines of a selection invention. But there must be something special about the later range. Here, all that 947 says about the point of introduction I will re-quote:

"The entry point for the two-phase recycle stream should be below the fluidized bed (polymerization zone) to ensure uniformity of the upwardly flowing gas stream and to maintain the bed in a suspended condition. The recycle stream containing entrained liquid is introduced into the reactor at a point in the lower region of the reactor and most preferably at the very bottom of the reactor to ensure uniformity of the fluid stream passing upwardly through the fluidized bed."

81. This says no more than the point of the low introduction is to ensure uniformity and maintain a suspended condition. This is fairly self-evident and is because in 947 the entering material aids in fluidisation. Solvay sets his limit of 75% for another pragmatic reason - ensuring thermal control. I think Solvay discloses injection at the level of the claim.
82. UC say there is no anticipation for another reason: that Solvay is not concerned with "A method for controlling the temperature ... during the production of polymer". They say that these words should not be equated with what is going on in a continuous process. I do not see why. Solvay is explicitly concerned with temperature control of the whole bed. He says, in 965:

'Thermal control of exothermic polymerisation reactions using the principle of internal refrigeration, i.e. by injecting the monomer or monomers in liquid form, is known, the vaporisation of the monomer or monomers providing the cooling necessary for removing the polymerisation heat."

83. I am not sure where he believed it was known from - perhaps his earlier applications (965 was the last of the three). And the passage I first cited is all about the best way of getting thermal control. If I were wrong about the construction of claim 1 of 947 then the suggested difference is not in fact a difference.

Obviousness - General

The skilled man

84. In considering obviousness it is first necessary to consider the attributes of the notional "person skilled in the art" postulated by s.3 of the Patents Act 1977 (corresponding to Art. 56 of the EPC). There was not much of a dispute about this. Mr Simpson said:

"I believe that the problems with which the patents are concerned would certainly have been addressed by a team of individuals whose collective knowledge and experience included the practical application of fluidised beds to for the manufacture of polymers."

85. That is obviously right. And it is reflected in the actual teams that each side employed. Central to the teams is a research department containing process engineers. More peripherally, but definitely to be consulted, would be engineers concerned with production (such as Mr Fulks), fluid and heat exchange experts (such as Mr Simpson) and outside consultants (such as Prof Gift). The evidence discloses that all these types of person were in fact involved on the respective sides. I say there was not much of a dispute because UC appear to have misunderstood a passage in Prof. Clift's report where he introduces fluidisation generally. All Prof Gift is saying is that a competent chemical engineer in general would have some familiarity with fluidised beds of the sort that is taught to students. He was not purporting to define the "skilled man".
86. The reason for the misunderstanding is not hard to see: UC wished to emphasise what they called the "prejudice" against allowing fluid to form in the system - the prejudice identified in the Fraser patent to which I have referred. One of my tasks is to decide whether there would have been such a prejudice in the mind of the ordinary skilled man.

Common General Knowledge

87. Patent law attributes to its notional skilled man a background technical knowledge which goes by the jargon "common general knowledge." Such knowledge is "part of the mental attitude of a well constructed representative of the class whom the specification is addressed" (British Ore Concentration Syndicate -v- Mineral Separation^[5]) or "generally regarded as a good basis for further action" (General Tire^[6]). It is not good enough to show that a matter was known to some but not to others and in particular it is not good enough to show that knowledge (or a prejudice) was confined to one or a limited class of suggested exemplars of the skilled man.
88. In the present case UC accept that the following matters formed part of the common general knowledge:

(1) that polymerisation reactions were exothermic and that, in a continuous process, one parameter of operation was that the temperature in the reactor was maintained below the sintering temperature of the polymer.

(2) that the removal of heat was not the only way of obtaining desired products. This could also be achieved by adjusting pressure or the chemical nature of the product - e.g. terpolymers.

(3) that the concept of the removal of heat by evaporative cooling was part of common general knowledge in the sense of being known to skilled persons working in the field. Thus Perry lists 5 options for the removal of heat from a fluidised bed, one of which includes the addition of volatile liquid so that the latent heat of vaporisation removes excess energy. However UC say that the fact that it is a basic option as a general concept does not make it common general knowledge in the context of the cooling polymerisation reactions in a gas fluidised bed.

89. UC assert that although the general concept that the introduction of liquid onto or into a medium to effect cooling was common general knowledge, the skilled team with knowledge of gas fluidised bed reactors would have, as part of their mental attitude, the fact that this was to be avoided in the case of a polymerisation reactions being undertaken in a gas fluidised bed. The fact that it was known to introduce liquid into other sorts of fluidised bed reactors - e.g. dryers - and stirred bed reactors, does not serve to detract from what was thought by a skilled team building on their experience with polymerisation reactions in fluidised beds. They point to what Mr Simpson said in his witness statement:

" the common general knowledge of a team working in the field of fluidised bed polyolefin production would have included a belief that subject to the exception of carriers for catalysts and prepolymers the introduction of liquids into fluidised bed polymerisation reactors should be avoided. I believe that this would have been taken as self-evident."

The reasons for this belief are, in short, that the introduction of fluids would have one or more of the effects referred to by Fraser.

90. Whether or not Mr Simpson said was correct forms an important part of this case. I must return to it in more detail. Pressing on with what UC accept as part of the common general knowledge:

(4) that of the other 4 Perry options, adiabatic cooling was not considered practical, solids circulation was a possible candidate (although rejected by UC), as was internal cooling (e.g. by water cooled coils.

Going on with common general knowledge, BP assert (I think correctly and without dispute) that the following formed part of that knowledge:

(5)that fluidisation polymerisation reactors have the following features common to all fluidised bed reactors, namely

(i) no moving mechanical parts, thus giving them an advantage over stirred beds;

(ii) that they can be operated continuously and on a large scale;

(iii) that they have extremely efficient mixing ability if the fluidising velocities are significantly (e.g. 5 times) above the minimum value; (iv) that they have very good heat transfer properties from gas to solids and so are substantially isothermal save for a small entry zone at the base.

(6) that stirred bed reactors remove heat of polymerisation by injection of liquid.

(7) that in addition the use of dry polymerisation fluidised reactors) as in the original Unipol and Napthachimie processes was well known for the manufacture of HDPE and there was no heat removal problem for such a process having regard to the reaction speeds of the time (which of course depended on catalyst efficiency). That all this is common general knowledge is implicit in the cases of both sides.

(8) BP assert that an increasing pressure from prospective licensees to produce LLDPE at commercial rates with longer monomers such as hexene and 4MP1 formed "by the early 80s"

91. That there was such pressure (and also pressure on BP and UC) by then is not in dispute. But so far as UC is concerned that pressure came earlier. Mr Ryan gave unchallenged evidence that in 1979 Mobil (a Unipol licensee) wanted to produce a material equivalent to Dow's "Dowlex" film, a C₈ copolymer. He said:

"We were unable to offer Mobil a higher alpha-olefin copolymer at that time because we could not raise our production rates to a level anywhere near the level of production for ethylene- butene copolymer because of the dew point problem. In those days we used to talk about ethylene-hexene copolymer running at half the rate of ethylene-butene production."

92. As a consequence, Mr Ryan said that UC went on to consider terpolymers. I think this evidence is important, throwing light on a need to make LLDPE at an earlier date than that accepted by BP. There is some confirmation in this on the BP side since it too investigated terpolymers at about the same time, 1979. They say they did not go to condensing mode then for technical reasons which I must consider later. The point at this stage is that there was a desire, some two years before the priority date, to make LLDPE commercially.

The "prejudice "

93. I can turn to the disputed part of the common general knowledge - namely whether the skilled man, even though he knew in principle that liquid could be introduced into a fluidised bed to remove heat, would not think that suitable for a polymerising fluidised bed of the Unipol or Napthachimie type.
94. BP say that such an idea was peculiar to UC and does not form part of the common general knowledge. The reason for this, say BP, is that at a very early stage (i.e. in the early 1970's) UC did away with a cyclone. The consequence was that fines were carried over. These contained catalyst particles which, being particularly small, were significant. So, in the Unipol process as then practised (dry) there were sometimes disasters. Polymer could form on the grid, causing it to block up. On occasions the blockage was so great that there was insufficient gas to maintain the bed and it simply solidified. There are graphic accounts in the evidence of operatives having to get inside the reactors at UC to cut through a vast mass of polymer with chain saws. BP say that is what is likely to happen if you do not use cyclones. This would greatly increase your fears about the operation of the bed and in particular doing anything to change it. You particularly would not want to introduce liquid for fear of forming a sludge with fines thus increasing the chance of a disaster. Certainly that was Mr Fulks' view: he accepted his general attitude was "if it works don't fix it." BP add this: that the risk of a major problem if you do not use a cyclone is significant even if you do use condensing mode. And there is certainly there is evidence of such a problem in 1985 with a condensing mode plant of UC's or their licensees.
95. BP go on to say that they never had a problem of this sort because they used (and still use) a cyclone. Certainly the evidence shows that there is indeed very little carryover from the cyclone in relation to the current process. UC say that cyclones are not perfect and that fines will get through - especially the very small catalyst particles.
96. BP put forward M. Raufast as exemplifying the skilled man at the heart of things. His oral evidence was to the effect that at Napthachimie/BP condensing mode was an obvious alternative at all times. It was not adopted in the only reactor they had at the time (the PZ3 A) for two reasons. First, when it was built, there was no problem and no need. It was built for HDPE where one would not want or need to use condensing made. There was no requirement to make LLDPE and no need to use the longer monomers which gave rise to the possibility of condensation. Moreover heat removal was not a problem because the limiting factor on production rate was the reaction rate (the "kinetic restraint") not heat removal. In due course, say BP condensing mode was considered but rejected. It was rejected not because of a fear of plate pluggage or the like but because, with existing plant, there would have been trouble outside the reactor - "in the loop" as M. Raufast put it. His evidence went like this:

Q. What would be the complication of condensation?

A. Condensation would have been an issue for the main loop of the reactor either for the compressor or the compressor, you know, we did not want to have a liquid as a section of the compressor or to have any other part of the loop as I mentioned earlier, some difficulty with the low point.

The reference to "the low point" was a reference to what he had described as "dead zones".

97. BP emphasised the "non-fear" of liquid at BP by reference to a proposal which M. Raufast made in 1977. Under French law there was a procedure known as an "enveloppe soleau". One could deposit with the French patent office a written proposal which then preserved one's rights later to do what was described. One did this even if one did not want a patent oneself.
98. M. Raufast had learned of the Solvay work via a paper published by them. This broadly (but not in all respects) described what is in the cited Solvay patents. M. Raufast thought something similar could be done for the manufacture of polypropylene. His proposal was filed in an "enveloppe soleau" in February 1977. His idea was take the unreacted gases from the top of the reactor and divide them into two parts. One part is recycled uncooled back to the grid to form part of the fluidizing gas. The other part is taken to a separator and the olefins (vinyl chloride specifically) condensed out. The gas from the separator is allowed to escape. The liquid monomer is either put straight into the bed, or is taken down to join the fluidising gas. M. Raufast said in his witness statement:

"I had no misgivings about the principle of introducing liquid into a fluidised bed by either means. It was necessary to separate out part of the recycle gas before condensation of that part for reasons of control. Since the recycle gas consists mainly of propylene, once the dew point is reached the transition between gas and liquid is very rapid. If the whole of the recycle gas is cooled, it is difficult to control the extent of condensation so as to leave enough gas remaining to fluidise the bed."

As to the manner of introduction direct into the bed, M. Raufast said he would not have simply put the liquid in. It would have been atomised somehow:

"the idea was to spray the liquid and certainly not to have the liquid dripping in the bed."

The point of entry into the bed is not dealt with explicitly by the enveloppe soleau. But it appears from the diagram to be at about the same level as the point of polymer withdrawal and so would have been into the isothermal region, i.e. above the thermal gradient region at the base.

99. There was some criticism of the practicality of the examples of the enveloppe soleau -which M. Raufast answered. I do not think that matters - its significance is that BP/Napthachimie thought that atomised liquid introduction was, in some circumstances, practical. It may be noted however, that the direct injection proposal would not have been within claim 1 of 947 on anyone's construction. Nor would it have been within claim 1 of 691 as I have construed it: there is no re-introduction of "said two phase mixture" As to the through-the-grid alternative that is not so clear. There is no deliberate cooling of the unreacted monomer to below the "maximum desired temperature within said bed" but there may have been enough natural cooling to count. And it is doubtful whether having the fluid join the unreacted monomer before it enters the reactor is within the claim, a point I am not disposed to consider further.
100. I accept that the enveloppe soleau is some evidence of a non-fear. But it did not proceed far (apparently because of catalyst difficulties) and I must also have regard to what happened later. The best view of this must be the contemporaneous documents rather than M. Raufast's memory 15-17 years later. They, to my mind, do show at least a degree of inhibition about putting liquid in - even when there was no existing constraint caused by the need to use the existing plant which could not readily be adapted to liquid use. Mr Thorley took me through a number of documents. It is sufficient to refer to a few though there were a number of others to much the same effect.
101. In December 1980 a report considering the manufacture of LLDPE by M. Raufast and another contained the following passages of which I have underlined some.

For PEbd [i.e. LLDPE] this difference [i.e. the difference between the entry temperature and the reactor temperature] is reduced by the two ends:

higher temperature of the gas entering the reactor to avoid any risk of condensation in the lines

temperature of the gas leaving the reactor, therefore temperature of polymerization, lower than for PEhd.

These are severe constraints. On a new facility they will impose evacuation of the heat in two stages: one upstream of the compressor and the other downstream to recover the heat due to the compression.

102. Furthermore, at many points in the gas circuit it will be necessary to counteract the risk of condensation (pressure tappings, associated gas, dead lines ...).
103. 3 Injection of the comonomer
104. The introduction in the liquid state and in massive quantity (150-220 litres per tonne of polymer) poses a delicate question. It is possible to envisage injection into the fluidized bed, into the cold gas, or into the hot gas. Although the fluidized beds are reputed to tolerate large quantities of liquid (drying in the fluid bed), it would appear wise to base ourselves, at least a priori, on a solution avoiding the direct introduction in the fluidizer. With the same objective of reducing the solid - liquid contact (pasty state ...) it would be more worthwhile not to introduce the liquid comonomer into the cold gas, because this would be too close to saturation. Direct introduction into the hot gas at the outlet of the reactor would therefore have some advantages by taking advantage at the same time of the heat available for the evaporation and the intensity of mixing achieved by the cyclones."
105. It is not really possible to reconcile this document with a freedom from concern about the introduction of any liquid - note the passages I have underlined And it is to be noted that a new plant was under discussion and hence there was no restraint arising from a pre-existing design.
106. In January the next year, M. Raufast did, but with considerable reservation, consider some form of liquid injection. His document reads:

POSSIBLE TECHNOLOGIES

With the PEBD process the processing of the liquid in the fluidization loop may have two origins:

a) the addition of comonomer (1-butene or higher homologue) which compensates for the quantity polymerized

b) the partial condensation of the comonomer present in the fluidization gas at the time of eliminating the heat of reaction in the cycle gas exchanges.

The make-up of comonomer in the liquid state is already done on PEBD. The comonomer is injected into the fluidization gas and vaporises totally there. On PEBD the quantity of make-up comonomer is greater but should remain compatible with this method of introduction insofar as the fluidization gas remains a dry vapour. This restriction imposes constraints on the operating conditions (temperature, composition ...). It is possible that the search for a particular polymer, the kinetics or the optimization of the process lead to operating conditions which prevent the total vaporization of the makeup comonomer and even involve the partial condensation of the fluidization gas during its cooling.

With the 1-butene, once the dew point is reached, a reduction of 1°C of the temperature of the cooling causes condensation of 2% of the gas phase, i.e. approximately 12,000 kg/h of additional condensates! With a heavier comonomer (but more rare than the gas phase) the increase of condensates would be less rapid but would still be large.

This great sensitivity of the flow condensed at the cooling temperature obliges us to design a liquid injection device in the fluidizer capable of operating with heavy flows.

It seems therefore that the intermediate solution consisting in condensing a little (some m³/h) to reintroduce into the fluidizer a moderate quantity of liquid and thus to retain the fluidizer as it is, is for the moment hardly credible. This opinion would have to be reconsidered if the heavy comonomers (hexene, octene, ...) were employed at "homeopathic doses".

This leads us to consider, in summary only the two limiting cases:

no condensation: the fluidizer functions as at present on PEHD (and PEBD on the pilot plants) in a rigorously dry phase. However, this operation may necessitate adaptation of the reactor peripheral units.

Condensation in considerable quantities: the flows mentioned show that it is necessary to adapt the fluidizer so that it accepts the corresponding liquid flows. This leads us to a "wet" fluidizer.

WET FLUIDIZER

We know that the fluidizers accept considerable liquid flows. This is the case for example with the fluid bed dryers. Professor Zenz tells us that we can evaporate up to 0.5 1 of liquid/h x 1 of a fluid bed......

If the fluid bed is potentially suitable to vaporise the liquid, it is necessary in advance to ensure good dissipation of it to avoid any phenomenon of sticking and hot points. Two approaches are possible:

(a) Fluidization + mechanical agitation (type BASF in its polypropylene process)

(b) Modification of the fluidiser sole plate to obtain a "spouted bed type behaviour close to the point of introduction of the liquid.

Solution b) consists in creating a very turbulent zone close to the liquid inlet point into the fluidizer.

107. M. Raufast drew out a device which shows what he meant by a "spouted bed". It was not a true spouted bed because the bed was maintained by gas coming up through the grid. His drawings are as follows:



He went on to describe this as follows:

"The device appears as a standard sole plate but equipped with a few nozzles (for example 4) for introducing the gas-liquid mixture. The gas-liquid mixture is given a speed of a few dozen m/s at the neck of a mixer head. This mixture behaves like the vector gas of an ejector and "sucks" the powder of the fluid bed into a mixer (type "tube draft" or Venturi). The jet at the outlet of the mixer has considerable kinetic energy (density and high speed) and also has a large diameter (a few centimetres). We therefore have locally behaviour typical of the "spouted-bed". The jet must therefore be very penetrating which favours the mass transfer (dispersion of the liquid) and the transfer of heat. A draft Soleau envelope built on this principle is enclosed.

DRY FLUID MEDIUM

In the absence of constraints, the question does not arise and it is of course the dry fluidized medium which should be preferred (technical credibility).

108. Now BP say that this shows that M. Raufast did come up with the idea of condensing mode for LLDPE. But I am not really sure it can fairly be said that he was here acting as no more than an ordinary man skilled in the art. Firstly he seemed to think it lacked "technical credibility" and second it all depended on his special injection technique (which I do not think really equates with the relatively minor adaptations to a baffle device (called a Chinese hat) which UC had to make to ensure even distribution and no pooling. I think M. Raufast was being somewhat ingenious in his special injection device. He was cross-examined about the document. Some of the cross-examination went like this:

Q. I do not quite understand that, Mr. Raufast, because this is a paper where you are setting out possible approaches and you say two approaches are possible. You do not say two amongst many are possible. These must have been your preferred methods, the ones that you wanted your employers to consider seriously. Is that not correct?

A. One was related to a mechanical device, generally speaking, having a kind of agitation system as it was with the BASF process. Something we did not want to consider further. The other possibility was with what was already available in the system meaning gas and liquid just out of the fluidizer bed to see if there was any possibility to achieve a good dissipation of the liquid. Here was just an example of what could be done.

And, a little further on:

Q. Reading on .... Can you explain why did you not at this stage simply refer back to your Enveloppe Soleau of 1977?

A. The Enveloppe Soleau of 1977 mentioned the two possibilities to inject liquid either below the grid or above the grid without any specific consideration of the details of the system. We were convinced that the solution was feasible but it was not the time to start any detailed works to see the nitty-gritty or the apparatus to do that.

Q But you proposed this rather complex solution. Would it not have been easier to refer back to your 1977 solution and say "of course that could be used"?

A. Yes. As I mentioned, this was one possibility amongst several others. It turned out that we gave up this solution because it was deemed not necessary to have condensation in the reactor. Therefore, there was no need to continue to work on this issue.

I do not think that at this time there was a realisation that the practical way forward was via a process of the claims of either patent.

109. In July 1981 M. Raufast prepared a memorandum recommending the use of nitrogen. He pointed out that for LLDPE.

"The concentration in the 1-butene become sufficiently important to reveal a preoccupation with the dew point which is quite unknown on HDPE."

He considers the consequences:

"The disappearance of hydrogen in favour of a denser gas like 1-butene causes an increase in density of the gas which is not easily compatible with the existing equipment (compressor of the PZ 3 A for example). Hence the lowering of the total pressure to eliminate this limitation, at least during the tests. By the same means we reduce the dew point temperature and especially the risks of condensation. For a new facility it would be necessary to follow the same route if only to reduce the risks of condensation.

Production

All other things being equal, a reactor producing PEBDL at only 8-10 bar total pressure and 80°C as suggested above would find its production potential reduced to half of that obtained on PEHD because of the disappearance of the possibilities of eliminating the heat.

For a new facility of high capacity, with such a pressure an unacceptable reactor size would result. With the technology and the process data of today, there is practically one way of remaining within a reasonable framework: increasing the heat evacuation power of the gas by adding an inert gas superposing its pressure on that of the process gas."

It seems to me that he was missing the real point - that you could use the latent heat of vaporisation to remain "within a reasonable framework". I do not know why he did so - after all he had proposed the enveloppe Soleau. It may be that he did not think his complex solution would really work. Whatever the reason, the documents do not show that harnessing the latent heat was seen as the way forward. I do not go to further documents. Mr Thorley referred me to them and they continue the same story.

110. I conclude that in both UC and BP the common perception was that, however theoretically it might be useful, condensation was a thing to be avoided. This is an important conclusion because it would colour how the skilled man would read the specific prior art relied upon.
111. In so holding I do not overlook Mr Simpson's reaction to the idea of condensing mode. As I have said he was not part of the team which made the invention. But shortly before the idea was to be tried (in December 1981) he was asked his views. Mr Jones' (one of the inventors) notes say:

"(1) Larry Simpson foresees no problems with liquid in the cycle gas (2) date from a previous high rate run at G-1759 indicates condensing mode has occurred."

And in a fuller note:

"LLS [i.e. Mr Simpson] does not expect liquid separation and knock-out in the bottom head.

The presence of liquid in the cycle gas is not expected to cause additional plate pluggage. In fact, LLS expects the condensed liquid will make the system cleaner than with current operation."

112. Mr Simpson said that in fact he thought condensing mode was a risky operation but that there was nothing in the laws of physics which was against it. It is not clear whether Mr Simpson was told that in fact, by accident, the plant had been operated in condensing mode with no problem. It seems to me clear that Mr Simpson himself had no great prejudice against operating in that mode - more a reservation.
113. On other hand others within UC were more circumspect, as appears from the documents of the time which I was taken through by Mr Waugh, some samples of which I should mention. Thus when the "accident" of running in condensing mode happened, Mr Beret's (one of the inventors) notes read:

"No operating problems. Nothing. Amazingly trouble free run. Will quit tomorrow."

And later his notes show a fear of chunk formation. And a Mr Engle is reported as saying:

"Risk include the possibility of increased distributor plate plugging and the possibility of production of low molecular weight resin at the point of condensate injection due to locally high hexane levels in the bed."

At a slightly earlier time, 1978, Mr. Beret wrote:

"If during the test, the reactor inlet temperature approaches the dew point, several operational problems are foreseen. Reiterating some of these expected problems:

a. butene control problems

b. distributor plate ΔP

c. cooler ΔP

d. large agglomerated particles

e. wide variations in resin properties f product discharge problems"

114. These documents do demonstrate the fear. They also show (in places I have not cited) the perception that it was heat removal which was limiting the space-time yield - a perception in the 1978/9 period which confirms that it was earlier than suggested by BP.
115. It is to the particular citations I must now turn.

Obviousness: the approach to the citations

116. I will try to adopt the structured approach of the Windsurfing^[7] case:

"There are we think four steps which require to be taken in answering the jury question, The first is to identify the inventive concept embodied in the patent in suit. There after the court has to assume the mantle of the normally skilled but unimaginative addressee in the art at the priority date and to impute to him what was, at that date, common general knowledge in the art in question. The third step is to identify what, if any, differences exist between the matter cited as being known or used and the alleged invention. Finally, the court has to ask itself whether, viewed without any knowledge of the alleged invention, those differences constitute steps which would have been obvious to the skilled man or whether they require any invention."

117. The first step is to identify the inventive concept. This is not always an easy task. As I said in Unilever v Chefaro^[8].

"It is the "inventive concept" of the claim in question which must be considered, not some generalised concept to be derived from the specification as a whole. Different claims can, and generally will have, different inventive concepts. The first stage of identification of the concept is likely to be a question of construction: what does the claim mean? It might be thought that there is no second stage - the concept is what the claim covers and that is that. But that is too wooden and not what courts applying Windsurfing stage one have done. It is too wooden because if one merely construes the claim one does not distinguish between portions which matter and portions which, although limitations on the ambit of the claim, do not. One is trying to identify the essence of the claim in this exercise."

Inventive concept of claims 1, 2 and 11 of 691

Claim 1

118. UC submit that it is:

"to operate a continuous process for the production of polymer in a gas fluidised bed in which the recycle gas stream is cooled so as to condense out at least a part of the gas stream which can be used to cool the polymerisation reaction by the re-introduction of the two phase mixture of gas and entrained liquid (condensate) into the reactor - i.e. operating in "condensing mode.""

119. BP say it is:

"The entrainment of liquid in the recycle gas which has been formed by cooling the recycle gas which has been formed by cooling the recycle gas stream below its dew point and reintroducing such entrained gas/liquid mixture back into the reactor below the grid."

120. Curiously UC's formulation is not as wide as the construction for which it contended. The formulation seems to be limited to the re-introduction of the two phase mixture. Not that it matters, for I have rejected that construction. The differences between the parties are thus that UC emphasise the continuous process and BP add below the grid. UC are clearly right about the continuous process. But what about below the grid? Plainly not all the mixture need come in below the grid (cf claim 2) but I think BP are right to this extent: that it is implicit that the mixture (or at least part of it) is the fluidising gas. This follows from the fact that the claim is for operation of a fluidized bed.
121. There is no difficulty about the concept of claim 1 - it adds the requirement that all the mixture comes in below the grid. Nor is there any problem about claim 11. It adds the requirement that the mixture is split, one fluidising stream coming in below the grid and the other stream(s) introduced into the bed, or not.

Obviousness of 691 over Mitsui (Japanese Patents 69577/80 and EP Appn. 40,992)

122. Although two Mitsui applications are cited, it is agreed they can be taken together, using the EP application with the drawing from the Japanese application. Apparently the EP application was originally published with the wrong drawing.
123. Mitsui's principal concern is nothing to do with the patents in suit. It is about being able to adjust the molecular weight distribution and/or chemical composition distribution of the final polymer. For this purpose two reactors are used, the product of the first being slurried with liquid hydrocarbon and fed to the second. The apparatus is shown in the only diagram:



124. The reactors may be of any sort ("fluidised beds, stirred beds, stirred fluidised bed, a tubular reactor etc."). BP concentrated on example 5 which is to a fluidised bed. The 3 lines into reactor A supply catalyst (suspended in butane - an inert gas for present purposes), ethylene/4MPl and hydrogen. Device D is a condenser and E a separator (a "knock out pot"). Mitsui says:

"The condensed liquid containing butane and 4MP1 .... Was partly fed in the liquid state into the polymerisation vessel from the drum E to utilise it for removal of the heat of polymerisation by its latent heat of evaporation."

125. It will be seen that Mitsui does not show how or where the liquid (shown by an arrow from E) is introduced into A. He does not bother with this at all.
126. The polythene made in A is taken off drum C. To that drum is added (inert) butane and a slurry is made. It is fed to reactor B and is shown as being fed to the top. Mitsui says:

"Butane transferred from the re-slurry drum C to the second-step gas-phase polymerization vessel was wholly gasified by the heat of polymerisation in the same way as Example 1, and was utilized so as to remove a part of the heat of polymerization."

BP do not rely upon this passage (which after all is concerned only with the inert butane) as with the next passage, which is to do with unreacted gases of B. They go to a condenser I and a knock-out pot J. Mitsui says:

"Butane and 4MP1 discharged from the polymerisation vessel together with the recycle gas, cooled at the heat exchanger I and recovered as a condensate at J were partly sent in the liquid state to the polymerization vessel where they were gasified for utilization in removing the heat of polymerisation by its latent heat of evaporation."

127. Mitsui is not explicit as to which vessel the liquid goes. As in the case of knock-out pot E all he shows is an arrow coming out. Probably it would go to reactor B, and the evidence^[9] is that one would guess that it is added at the top, along with the slurry.
128. I have already identified the common general knowledge {Windsurfing 2) and can go on to Windsurfing 3, the differences:

(1) Mitsui does not make use of a gas/liquid entrained mixture;

(2) Mitsui does not reintroduce the two phase mixture into the reactor at all;

(3) Mitsui does not use any such mixture as the fluidising gas. He uses a dry gas.

129. Turning to the final question, obviousness, I have come to the clear conclusion that Mitsui does not render the claim obvious. I think the skilled man would regard the document as somewhat obscure, as not concerned with any problem of space-time yield and of no assistance. BP's argument, which I reproduce below, is a hindsight argument. They say:

"The only point not specified is precisely where the liquid condensate is returned to reactors A or B - it being left to the engineer to work out for himself. He would reason as follows:

Given that liquid comonomer is being returned to the vessel clearly one would want the liquid to enter the bed in such a way as not only to remove some of the heat of polymerisation but also so that it is uniformly dispersed in the bed in order to copolymerise with the ethylene.

Schematically the butane catalyst suspension is illustrated as being introduced via line 1 at the top of the bed of reactor A - but given the fine particle size of the catalyst (22 microns) this would not be good engineering practice as the fine particles would be more likely to be carried over and hence not uniformly distributed throughout the bed.

Likewise schematically the butane polymer slurry is illustrated as being introduced via line 9 at the top of the bed of reactor B - but the volatilised butane would not perform its function of heat removal from the bed unless introduced sufficiently far into the bed to allow the liquid to be properly dispersed."

Actually the argument is based in part on a fallacy, a fallacy about the fines. The fines in Mitsui are likely to be in a slurry.

130. Accordingly I reject the allegation of obviousness based on Mitsui.

Obviousness of 691 over Fraser

131. This was pleaded but not pursued. Quite apart from the problem of Fraser pouring cold water on the concept of condensing mode, he does not disclose the concept of entrainment at all and could hardly serve as the basis of an obviousness attack.

Inventive Concept of 947, claims 1 and 2.

Claim 1

132. UC say the concept is:

"to control the temperature of a continuous process for the production of polymer in a gas fluidised bed by the introduction of a stream of liquid into the reactor as an essentially uniform mixture at a level below the region of maximum desired temperature."

133. BP say:

"it covers both entrainment and injection, but is limited by the fact that the uniform mixture of recycle gas and liquid is to be introduced into the bed below the region of maximum desired temperature."

134. UC's formulation is, consistently with their submission on construction, extremely wide, covering any injection of liquid. I have rejected this and held that the claim, and therefore concept is limited to the introduction of a uniform mixture of recycled gas and liquid monomer, the monomer being recycled monomer. Direct injection into the reactor is within the concept only if it a uniform mixture with those components which is introduced. The concept is limited to a continuous process.

Claim 2

135. This adds little to claim 1. I heard little about it and propose to say no more.

Obviousness of 947 over Mitsui

136. I can go straight to Windsurfing 3. The differences are:

(1) Mitsui does not make use of a uniform mixture of gas and liquid;

(2) A fortiori he does not use recycled (said) gas as part of a phase mixture;

(3) Mitsui does not introduce below the region of maximum desired temperature.

137. As for Windsurfing 4, I do not see how the skilled man would readily take Example 5 of Mitsui and adapt it so as to take those three steps. The argument is a hindsight argument and I reject it.

Obviousness of 947 over Solvay

138. Again I can go straight to Windsurfing 3. The differences are:

(1) It is a discontinuous process;

(2) It is on laboratory or pilot scale;

(3) The injection is by "spraying". It is not a spraying of a mixture of said gas and liquid, however, as I construe the claim.

139. I have held that there is no difference as regards the point of injection, Solvay recommending from 0 to75% of the bed height.
140. Windsurfing 4. Thus the differences are not great. BP contend that a skilled man would see that Solvay is the first stage of development of a commercial process and that the next stage, scale-up to a commercial or pilot plant, would surely involve a transition to a continuous process. I agree, and find confirmation from this in that both Solvay (in a later paper) and M. Raufast (when proposing his polypropylene plant described in the enveloppe soleau) considered the natural thing was to use a continuous process. Thus if the only differences were (1) and (2), which they would be on UC's construction of the claim, the claim would be obvious. I do not find this surprising. On any view it is not a claim limited to cooling the unreacted gases to below the dewpoint. Nor is it a claim limited to having any liquid in the loop. All that can be dry. The claim covers injection of liquid directly into the bed and on UC's construction the only difference between the claim and Solvay is that the injection is of a mixture of gas and liquid. That would be the self-evident way to inject because of the danger of forming agglomerates if the liquid was simply introduced in an unatomised stream. And after all, Solvay indicates atomisation by the use of the word "spray". However I have construed the claim more narrowly - as being limited to introducing the liquid using the cooled recycle gas to form the mixture. That idea is not in Solvay. Nor did their later paper or M. Raufast fasten on to it. I do not think that idea is shown to be self-evident from Solvay. Hence I do not consider claim 1 (and hence claim 2) of 947 obvious.

Obviousness of claims 1 and 2 of 947 over Resine

141. This was only lightly pressed and did not even feature in BP's written argument. I have already identified the differences and it suffices to say here that any obviousness case must be weaker than one based on Solvay.

Obviousness: Commercial Success

142. UC seek to support the patents on the basis of commercial success of their own process. One must always be chary of such a plea, for commercial success may be the product of many other factors. Moreover even where there is commercial success, it is often important (but not essential) to show that there was a prior need for the invention - long-felt want as it has been called in the cases. Furthermore the plea is seldom likely to be relevant if the prior art relied upon was published only shortly before the date of the invention: there is unlikely to have been enough time for the ordinary skilled man to have missed the point. And finally it is necessary to show that the success is due to the invention - the features of the claim in question. One must compare the features of what has achieved success with the features of the claim. If the claim is wider than that for which success is claimed, then normally a plea of commercial success is simply irrelevant. Here, for instance, I think, the claimed commercial success cannot support claims 1 and 2 of 947. They are much wider (all the more so if UC are right on construction) than the actual process which led to UC's commercial success.
143. Bearing in mind all those caveats, I nonetheless think that UC are entitled to rely upon commercial success to support claim 1 of 691. It relates essentially to using the entrained mixture as the fluidising medium coming up through the grid. That is what UC do. It is technically and commercially a success and plainly is widely used by UC and its licensees. Moreover there is evidence of a need for such a process in 1979 in the evidence of Mr Ryan to which I have referred.

Insufficiency

144. BP plead that both patents are insufficient, that they do not disclose the invention clearly enough and completely enough for any of the claims to be performed by a person skilled in the art.
145. The plea is in the nature of a squeeze. BP are not really saying the skilled man would not know how to operate in condensing mode. They say that he would have to do significant engineering design work to make it operate efficiently. In this regard they point to work done at UC concerning the design of a baffle used in the plenum to ensure uniform gas flow and the non-formation of liquid pools. The patent says no more than this:

"A baffle or similar means for preventing regions of low gas velocity in the vicinity of the recycle stream entry point may be provided to keep solids and liquids entrained in the upwardly flowing recycle stream."

146. The drawing schematically shows no more than a cone in the middle of the plenum above the point of entry (at the bottom) and well below the grid. Obviously in real life some work would have to be done on the baffle. It was, over a few months, as was put to Mr Simpson. It is not that condensing mode would not work without this effort, it is that by design of a baffle you could put more liquid through without actual pools of liquid forming in the plenum. I do not think the evidence as to this work remotely establishes insufficiency. Nor do I think it makes a dent in the plea of commercial success. The suggestion was that the commercial success was really due to this design work and not the invention. Likewise I do not accept that this design work is really the difference between the pleaded prior art and what UC do, that the patents in suit teach no more than that which is in the prior art.

Conclusion

147. In the result I hold that the BP Mark 1 process infringed claim 1 of 691 and claims 1 and 3 of 947 together with some subsidiary claims which do not matter. The BP Mark 2 process as carried out at Grangemouth now and the possible variant of unassisted injection (the Mark 3) do not infringe. Both patents, on my claim construction, are valid. I will hear counsel.

Note 1    p.6-    [Back]

Note 2   [1972] RPC 457    [Back]

Note 3   OJ 1990. 188    [Back]

Note 4   [1862] 31 LJ Ch 457    [Back]

Note 5   (1909)26 RPC 124 at 138    [Back]

Note 6   [1972] RPC 457 at 482-3    [Back]

Note 7   [1985] RPC 59 at p.73    [Back]

Note 8   [1994] RPC 567 at 580    [Back]

Note 9   Prof. Clift report §4.18    [Back]