England and Wales High Court (Patents Court) Decisions

Processes to make powders easier to handle are a major and profitable business all over the industrial world. These processes are particularly useful in the pharmaceutical, detergent, food and fertiliser businesses. One of the processes used for this purpose is fluidization. Fluidization is the operation by which solid particles are transformed into a fluid-like state through suspension in a gas or liquid. The device in which particles are fluidized is called a fluidized bed. In the particular kind of fluidized bed with which I am concerned, a liquid spray material is sprayed into a stream of particles. Depending on the conditions within the fluidized bed, the liquid either coats the particles ("coating") or makes them stick together ("agglomeration"). The patent in suit claims a product and a process for both.

The Claimant ("Vector") claims that the patent is invalid on the ground that the claimed invention is not new; or, if new, is obvious in the light of common general knowledge and cited prior art.

The patentee ("Glatt") applies to amend the patent. The application is opposed both on substantive and discretionary grounds. Logically, it seems to me that the first issue I should consider is that of amendment, since it is the patent as amended on which the attack has been focussed in the evidence and argument. However, before doing so, I will describe the technical background, the patent in suit and the cited prior art.

Mr Richard Davis appeared for Vector; and Mr Douglas Campbell appeared for Glatt. I heard expert evidence from Professor Derek Geldart for Vector; and from Professor Michael Hounslow for Glatt.

Professor Geldart is (among other things) Emeritus Professor of Powder Technology at Bradford University (where he was appointed to a personal chair in 1988). He had had a long connection with the nuclear fuels industry; and it was while working in that industry that he built his first fluidized bed. He is the author of many publications; and the recipient of a gold medal for his outstanding contribution to technology over twenty five years. He had little practical experience, however, of coating or granulation. Professor Hounslow is Professor of Chemical Engineering in the University of Sheffield. He is the author of a number of papers on fluidized bed granulation; and leads a team advising Unilever on that topic. He has more practical experience of this field than Professor Geldart. I found Professor Geldart's evidence on the matters chiefly in issue more conclusionary and less fully reasoned than Professor Hounslow's. Professor Hounslow was also more methodical in his approach both to the patent in suit and to the pleaded prior art. He was, in short, better at explaining things; and at giving cogent reasons for his conclusions. In general terms, where they disagreed, I prefer Professor Hounslow's evidence.

Common general knowledge is the mental equipment that is necessary for competency in the field of endeavour in question. Common general knowledge in this sense is knowledge over and above the general knowledge which an ordinary member of the public would possess. Before coming to the patent in suit, I first set out what would have been common general knowledge in this sense at the priority date (1991).

A fluidized bed is formed by passing a fluid, usually a gas, upwards through a bed of particles supported on a distributor. The basic components of a fluidized bed are: a container; a gas distributor; a powder and a source of gas.

The container is typically cylindrical (although it may be square or rectangular). Almost always they flare out into a section with a larger diameter at the top, whose function is to reduce the velocity of the gas and consequently to reduce the entrainment of particles. This section is usually referred to as an expansion chamber.

An important component of a fluidized bed is the means by which the gas is introduced. In most cases it is important that this be done uniformly, so that the velocity of the gas does not vary greatly across the bed. When uniform distribution of gas is required, it is usually achieved by a gas distributor and a plenum chamber. The gas distributor is a [strong enough] gas-permeable solid barrier immediately beneath the bed of solid. This description of a gas distributor is Professor Hounslow's; but Professor Geldart agreed with it, subject to the qualification in square brackets. It typically takes the form of a fine screen (e.g. a woven metal fabric or a solid plate through which many holes have been drilled). It must be strong because it must take the weight of the solid when no gas is present; and it must be able to withstand the differences in pressure on its upper and lower surfaces. The function of the gas distributor is normally to ensure that the fluidizing gas is distributed over the entire cross-sectional area, while at the same time preventing particles from entering the plenum chamber. The gas distributor is the most difficult part of a fluidizing bed to design. The plenum chamber is a space beneath the distributor that allows room for the gas to flow laterally in order to achieve a uniform distribution of velocity. This description was also Professor Hounslow's. Professor Geldart agreed with it.

Fluidized beds are particularly attractive for granulation and coating of particles. Because they give rise to substantial motion in the solids, they allow the solids to pass repeatedly through the region where the liquid is supplied, without remaining there too long. This is important for two related reasons. First, it facilitates the application of the same quantity of liquid to each particle. Second, it minimises the risk of over-wetting of particles. If this occurs then particles may stick together in an uncontrolled and undesirable way.

In order to distribute the liquid evenly among the particles, it is necessary to break it up into droplets. It is also important that the droplets themselves be sufficiently separated in space so that they are distributed on several particles in the vicinity of the spray; and are not deposited at high concentration on a few particles. If liquid is delivered in high concentrations to only a few particles at a time then:

i) It is very difficult to obtain uniform coating;

ii) Those particles that receive large concentrations of liquid will be very wet and sticky which promotes granulation in an uncontrolled manner;

iii) Collisions between particles involving one or more particles that are over-wet will cause the bed to "quench". If the whole of the bed "quenches" it leaves a sticky mass of solid sitting on the gas distributor which at least will mean that that batch of material is useless; and at worst will damage the fluidized bed processor itself. This phenomenon is described in Gas Fluidization Technology (edited by Professor Geldart) p. 266:

"The main potential problem with fluid bed granulation is an operational one. If the ratio of liquid to solid becomes too high the particles will rapidly form large agglomerates the size of golf balls and fluidization will rapidly be lost. The plant must then be shut down and the bed dug out. The operator gets only a few minutes warning of the onset of this "wet quenching" process."

It is therefore desirable:

i) To separate the particles in space;

ii) To separate the droplets in space; and

iii) That the droplets are smaller than the particles.

It is for the latter two reasons that the droplets are usually sprayed into the fluidized bed processor.

A spray of liquid droplets is usually formed by pumping the liquid through a nozzle, and atomising it by means of a gas stream. A stream of liquid issuing from an opening or orifice will naturally tend to break up into droplets. The faster the liquid flow, the smaller will be the droplets. A nozzle that operates with only one flow (the flow of liquid) is called a single phase nozzle. One drawback of the single phase nozzle is that the droplets tend to be relatively large and uncontrolled in size. This can be alleviated by the use of an additional stream of gas. The use of an additional stream of gas reduces droplet size by breaking up the droplets and also by entraining further gas thus decreasing the concentration of the droplets. This is atomisation. Since there are two flows (liquid and gas) this kind of nozzle is called a two-phase nozzle. The two-phase nozzle has been in use for many years.

One of the problems associated with nozzles is that, particularly in the case of a two-phase nozzle, the spray pattern is easily disrupted by eddies which can cause the spray pattern to be blown back onto the nozzle. This can lead to the build up of deposits on the nozzle which can, in turn, block the gas or liquid orifices, or both.

In addition, the injection of a jet of high velocity fluid (such as the gas issuing from a nozzle) entrains fluid in the surrounding area. This gives rise to a diminution in pressure and thus a flow of surrounding fluid towards the jet (the venturi effect). The direction of the flow is naturally towards the source of the spray. This can have the effect of causing particles to block the nozzle.

One way of overcoming problems associated with the blockage of nozzles is the use of a spouted bed. This is a fluidizing bed in which gas is introduced by one centrally placed nozzle. The nozzle creates a central zone in which particles are transported upwards and return to the bottom in an annular downward moving zone. One advantage of a spouted bed is that it is not necessary to use a gas distribution plate, which is liable to become scaled and plugged.

The Wurster device (invented by Dr Dale Wurster in 1959) is a gas suspension technique used in a fluidizing bed. The outer container contains an inner cylinder half its diameter. The inner cylinder is known as a partition. At the base of the outer container is a fine screen, and an air distribution plate. In the centre of the plate is a nozzle positioned to spray upwardly. The holes in the plate beneath the partition are larger in diameter than those in the surrounding ring. Gas passes through the plate at a high volume and velocity; and pneumatically transports particles vertically through the partition and coating zone. This is called the upbed. The coated particles leave the partition and begin to decelerate in an expansion chamber immediately above the partition. Once the velocity is such that the particles can no longer be entrained, they drop down between the partition and the outer container and can cycle through the spray zone once again. This is called the downbed. Where a Wurster insert and a gas distribution plate is used in combination with a spouted bed (which does not need a gas distribution plate) it is called a spout fluid bed.

Initially this technique was not taken up by the pharmaceutical industry, although there was better take up in other industries. By the 1990s however, it was used in the pharmaceutical industry as well.

Common general knowledge may also include the contents of standard textbooks that the skilled addressee would have at his elbow, and would regard as a readily available and reliable source of information. Although there was some disagreement between the experts about the precise publications that would fit this bill, they both agreed that the skilled addressee would have (or have ready access to) Kunii and Levenspiel's book: Fluidization Engineering.

Chapter 4 of Kunii and Levenspiel deals with distributors. It illustrates four kinds of perforated plates. These include "grate bars" which consist of a series of parallel bars, saw-toothed in cross-section (and hence without a uniform cross-section). They explain that perforated plate distributors cannot be used under severe operating conditions. Under those conditions, it is necessary to use what are called "tuyere designs". However, tuyere designs were not used in the fields of granulation and coating.

The problem of agglomeration is caused by the particles becoming too wet. Although it was known that the spray was most concentrated at the nozzle port, it was not known that that concentration was the cause of over-wetting. How wet the particles are is a balance between the rate at which the liquid is sprayed into the processor and the rate at which it evaporates and solidifies, combined with the position of the particles themselves.

It was also known that there was a limit to the rate at which spray could be introduced to the processor; but it was not known what the rate limiting factors were. Over-wetting was one possibility; but others were the drying limits; and adverse trajectories of particles within the processor. Moreover, an increase in the spray rate will increase humidity within the processor, and hence diminish the drying rate.

The claimed invention was devised by Mr David Jones, whose evidence was largely unchallenged. It evolved from the need to improve performance and productivity over that obtainable from traditional Wurster processors alone. In evaluating the Wurster process Mr Jones considered the rate limiting factors governing the rate at which spray could be introduced into a fluidized bed. The first was the capacity of the nozzle to atomise the liquid at the desired spray rate and droplet size, using a sufficient mass of air and atomising air velocities that are reasonable and do not cause attrition of the fluidizing substrate.

Mr Jones explained that the developing spray pattern has a high concentration of liquid close to the nozzle port. This results in a strong potential for over-wetting of the substrate in this area. He called this area "the wet zone". He realised that the wet zone limited the spray application rate. This was a second rate limiting factor. The consequence of this is that the drying capacity of the processor remains largely unused. This led him to incorporate a spray nozzle surround, which keeps the substrate away from the spray pattern until it is more fully developed (i.e. more diffuse).

In short what Mr Jones hit upon was the cause of the problem; and what to do to solve it.

The patent was filed in October 1992 with a priority date of 28 October 1991. It was granted on 5 March 1997 (EP 0 570 546). In the conventional way the patent begins by describing the field of technology involved. It then sets out the problem that it intends to solve. It describes the problem thus:

The spray material consists normally at least to a substantial part of a liquid. The nozzle forms a liquid jet that breaks into small droplets. These droplets are distributed in a cone-shaped cloud or spray pattern. The known processor may experience turbulent [air] flow immediately above the spray nozzle. Particles to be coated or agglomerated may enter the spray pattern produced by the spray nozzle very close to the spray nozzle and before the spray pattern has fully developed.

This results in uncontrolled droplet formation upon those particles that enter the spray pattern too soon. This gives in turn the drawback that the effectiveness of the processor is reduced and results in a relatively extensive processing time. The fact that particles may enter the spray pattern [too] soon may result also in excessive agglomeration of the particles. This is particularly disadvantageous if the particles are only to be coated, but may also be unfavourable if the particles shall be agglomerated. If the particles to be wetted have abrasive properties, they may moreover move with sufficient force to cause erosion of the nozzle tip." (Column 1 _35-55)

The patent next describes the object of the invention:

"Starting from the prior art known from [Naunapper, one of the pieces of prior art], it is an object of the invention to provide a fluidized bed processor comprising a spray nozzle and a method for granulating or coating a product which avoid that the granulating or coating is negatively influenced thereby that the particles enter the spray pattern too soon. Moreover, it shall be avoided that abrasive particles cause erosion of the nozzle tip." (Column 2 _6-13)

The invention is described as follows:

[T]he processor ... comprises a means which is positioned adjacent said spray nozzle and adapted to form in operation a gas stream that surrounds the spray nozzle so that said means said gas stream shields in operation the initial spray pattern developed by said nozzle against the entrance of particles moving through the fluidized bed.

This means and gas stream forms a shield or barrier. This ensures that particles, disposed in the product container section outwardly of the shield or barrier, are prevented from entering the spray pattern before the spray pattern is sufficiently developed. This allows the droplet density to decrease before the droplets contact the particles of the fluidized bed. Accordingly, the particle surface will be more evenly wetted preventing excessive particle agglomeration. The liquid contact with the particles can be more precisely controlled. The processor and the method according to the invention give particularly the advantage that particles to be coated are prevented from entering the spray pattern until such time as the droplet density of the spray pattern has been substantially reduced. This enables to use higher spray rates with less agglomeration." (Column 2 _17-41)

The patent goes on to explain that the fluidized bed is preferably configured as a Wurster type processor. The means is configured to shield the nozzle within the upbed and allowing a flow of air within the shield but around the nozzle. The shield may be adjustable and "utilized to alter the associated spray pattern". The upflow of air is itself described as a "columnar shield". The overall effect is described as follows:

"There may be provided a coating zone within the upbed of a Wurster type coating processor whereby the coating zone can be protected by a surrounding column of upwardly moving air in order to allow the coating zone to more fully develop and the liquid droplet density thereof to be substantially reduced prior to entry of particles into the coating zone." (Column 3 _9-15)

The patent then introduces two figures, each of which is described as a "fragmentary schematic vertical sectional view".

The patent describes the figures as follows:

"Referring now more specifically to the drawings the numeral 10 generally designates a typical Wurster type coater, modified in accordance with the present invention. The coater includes a product container section 12, an expansion chamber 14 into which the upper end of the product container section 12 opens, and a lower plenum chamber 16 disposed beneath the product container, separated therefrom through the utilization of an air distribution plate or screen 18. The upper end of the expansion chamber 14 may open into a filter housing (not shown) disposed thereabove including the necessary air filter structure and air outlet. The air distribution plate or screen 18 defines a plurality of air passage openings 20 through which air or gas from the lower plenum chamber 16 may pass into the product container section 12. As is conventional in Wurster-type fluidized bed systems, the holes in the plate 18 in the area outside the cylindrical partition 22 are smaller in diameter than the holes beneath the partition 22. This results in higher air volumes and velocities in the central area than in the downbed area. Although only a single plate 18 is depicted, it should be understood that the plate 18 is typically formed of two plates, an outer annular plate and an inner replaceable plate whereby the air flow ratios between the upbed and the downbed may be changed by substituting the inner plate with a plate having different air opening configurations....

The product container 12 has a cylindrical partition 22 supported therein in any convenient manner having open upper and lower ends, 24 and 26, the lower end 26 being spaced above the air distribution plate or screen 18. The partition 22 divides the interior of the product container section 12 into an outer annular downbed area 28 and interior upbed 30. A spray nozzle assembly, referred to generally by the reference number 32, is mounted at or through the air distribution plate 18 and preferably projects upwardly into the interior of the cylindrical partition 22 and the upbed 30 defined therein. The spray nozzle 32 receives a supply of air under pressure through an air supply line 34 and coating liquid under pressure through a liquid supply line 36, as is known in the art.

A shielding means, such as an inner cylindrical partition 40 is disposed about the upwardly projecting spray nozzle assembly 32. The inner cylindrical partition 40 has its lower open end snugly and telescopingly received within a tubular collar 42 secured to the air distribution plate or screen 18 about the opening 44 through which the spray nozzle assembly 32 is secured. The tubular collar 42 is secured to the air distribution plate or screen 18 through the utilization of suitable fasteners 46. The collar 42 includes circumferentially spaced axial slots 48 in which mounting studs 50 projecting radially outwardly of the inner cylindrical partition 40 are slidingly received. The mounting studs 50 have threaded nuts 52 threadedly engaged therewith whereby vertically adjustment of the inner cylindrical partition 42 relative to the tubular collar 42 may be enabled. The spray nozzle assembly 32 discharges a spray pattern 56 of air and coating liquid. Some of the air introduced into the lower plenum chamber 16 passes upwardly through the openings 20 formed through the air distribution plate or screen 18 below the inner cylindrical partition 40. The partition 40, as well as the tubular column of upwardly rising air about the spray nozzle assembly 32, shields the lower, beginning portion, of the spray pattern 56. The particles 60, passing upwardly through the upbed 30, are not drawn into this spray pattern. The annular column of air thereby allows spray pattern 56 to substantially develop, and the liquid droplet density of the spray pattern is substantially reduced before particles 60 enter into the spray pattern 56. By adjusting the height of the inner cylindrical partition 40 relative to the tubular collar 42, the height of the upper end of the inner cylindrical partition 40 relative to the upper extremity of the spray nozzle assembly 32 may be adjusted."(Column 3 ₃₂ - 4 ₅₄)

Figure 1 is reproduced below. The arrows within the areas containing blobs represent the movement of the particles up the upbed and down the downbed. The arrows within and emanating from partition 40 represent gas flows.

Having described the general arrangement of the invention, the patent continues with the results of certain experiments. The details do not matter; but the upshot was that, using the claimed invention, it was possible to increase the spray rate markedly; in some cases as much as threefold.

Claim 1 as granted (with proposed amendments in brackets) is:

"Fluidized bed processor having a product container section (12) opening downwardly into a plenum chamber (16) through a gas distribution plate or screen (18) having openings (20) formed therethrough for upward flow of fluidizing gas, particularly air, from the plenum chamber (16) into the product container section (12) for fluidizing particles (60) therein so as to form a fluidized bed, the product container section (12) including a spray nozzle (32), characterised thereby, that a means is provided which is positioned adjacent said spray nozzle (32) [and/or] adapted to form in operation a gas stream that surrounds the spray nozzle (32) so that said means [and/or] said gas stream shield[s] in operation the initial spray pattern developed by said nozzle (32) against the entrance of particles (60) moving though the fluidized bed."

Both of the "and/or"s are to be replaced by "and"; and the final "s" on "shields" is to be deleted, if the amendments are allowed.

Claim 2 as proposed to be amended is:

"Processor according to claim 1, characterised thereby, that the product container section (12) includes a substantially cylindrical partition (22) with an upbed (30) defined therein and said means is provided within the upbed (30)."

Claim 3 as proposed to be amended is:

"Processor according to claim 2, characterised thereby, that said means is a shielding means which comprises a cylindrical inner partition (40) within the upbed (30), which shielding means is positioned adjacent said spray nozzle (32), and adapted to form in operation the gas stream that surrounds the spray nozzle (32), and allows said gas stream to flow up within the cylindrical inner partition (40) around the spray nozzle (32)."

Claim 4 as proposed to be amended is:

Processor according to any one of the claims 1 to 3 characterised thereby, that the spray nozzle (32) produces said initial spray pattern by means of atomization air, and that said gas stream is other than the atomization air.

Claim 5 as proposed to be amended is:

Processor according to claim 3, characterised thereby, that the cylindrical partition (40) of the shielding means surrounds the spray nozzle (32) and one end of that cylindrical partition (40) is disposed at a level adjacent the outermost extremity of the spray nozzle (32).

Claim 6 as proposed to be amended is:

"Processor according to claim 2, characterised thereby, that said substantially cylindrical partition (22) is a substantially cylindrical outer partition (22) and is spaced above the distribution plate or screen (18) for defining, in addition to said upbed area (30), an outer downbed area (28), that the spray nozzle (32) is adapted to discharge upwardly, that the said means comprises an upstanding inner partition (40) mounted adjacent to the gas distribution plate or screen (18), that the upper end of the inner partition (40) is open and disposed at least as high as the upper extremity of the spray nozzle (32) and that the lower end of the inner partition (40) is open for receiving gas from the plenum chamber (16) having passed through the plate or screen (18)."

Claim 13 (as proposed to be amended) is:

"Fluidized bed processor having a product container section (12) opening downwardly into a plenum chamber (16) through a gas distribution plate or screen (18) having openings (20) formed therethrough for upward flow of fluidizing gas, particularly air, from the plenum chamber (16) into the product container section (12) for fluidizing particles (60) therein so as to form a fluidized bed, the product container section (12) including a spray nozzle (32) and a substantially cylindrical partition (22) with an upbed (30) defined therein, characterised thereby, that a means is provided within the upbed (30) which . . means is adapted to form in operation a gas stream that surrounds the spray nozzle (32) so that ...said gas stream shields in operation the initial spray pattern developed by said nozzle (32) against the entrance of particles (60) moving through the fluidized bed."

Claim 14 is the method claim. It is, I think, common ground that the method claim incorporates all the elements of the product claims; so that it is not necessary to set it out in full.

I have already mentioned the meaning that both experts attached to the expression "distributor" or gas distribution plate. As the patent in suit makes clear, the gas distribution plate may be formed of more than one section; and may consist of two stacked plates. The patent also explains (in a passage proposed to be deleted) that although an air distribution plate or screen is depicted, gas flow may originate into the processor "through other structural means". Mr Davis argues from this that when the unamended patent speaks of a "gas distribution plate or screen" (as it does in claim 1), it also includes other structural means that do not amount to a gas distribution plate or screen.

I am prepared to accept that claim 1 may well embrace a distributor of the grate bar type as shown in Kunii and Levenspiel (which is not strictly speaking a "plate"). It may even extend to the tuyere design, which is not a plate either, although this is unlikely since tuyere designs were not used in this kind of processor. But I do not consider that the expression goes any further. In particular I do not consider that the expression can be stretched to include something that the skilled addressee would not recognise as a plate or screen at all. Mr Davis argued that the expression encompassed anything within a fluidized bed processor that could fulfil the function of a gas distribution plate. The argument was developed with a firm eye on the pleaded prior art (notably Larson). However, in my judgment it was an argument based on the so-called "doctrine of equivalents" which does not have a place in our law. I do not accept it.

The patent makes clear that the deflector or shield can be formed "integral with the spray nozzle itself". Nevertheless, in my judgment the patent is plainly distinguishing between the nozzle and the shield. The latter must add to the former, even if it is integral with it. First, it is the spray pattern emerging from the nozzle that requires shielding. Second, Naunapper (which is cited prior art and is said to be the starting point for the claimed invention) has both two fluid and three fluid nozzles. The three fluid nozzle would already have a stream of air surrounding the spray. Third, the claim requires the means to be adapted to form a gas stream that "surrounds" the nozzle. Fourth, Figure 1 clearly distinguishes between the means and the nozzle.

The claim says that the means and the gas stream shield "the initial spray pattern". The specification says that the partition as well as the upwardly rising air shields "the lower beginning portion" of the spray pattern. It is true that the patent does not specify precisely what it means by "initial". But it must mean sufficient of the spray pattern as it emerges from the nozzle for the claimed invention to have some effect. As Professor Hounslow put it: "If we were speaking only of the first millimetre or so of the spray pattern, the point at which it emerges, then simply shielding that would be of no value at all."

Larson is a US patent (US 3 110 626) filed in 1961 and granted in 1963. It describes an apparatus for coating discrete solid material. The patent describes a coating chamber which is connected at its base to a tube. The tube has a constriction in its upper portion, which the patent refers to as a "venturi". The venturi causes a drop in pressure and an increase in velocity of the gas passing through it. The patent describes the benefits of the venturi as follows:

i) It provides a means of supporting discrete particles to be coated with a minimum pressure loss;

ii) It eliminates the need for a screen;

iii) It avoids the existence of dead regions.

The atomiser assembly can be placed within the throat of the venturi. Because the particles are supported above the mouth of the venturi, this can be done without risking plugging of the nozzle assembly.

The principal embodiment of the invention is described in Figure 1. Figure 1 is partially reproduced below.

Fluidization gas enters the apparatus through a transverse duct, passing through a gas heater (20). It passes into a space which the patent calls a "duct" (24). The gas then rises past a lozenge shaped element which the patent calls "a gas foil guidance element" (48) which divides it into two streams. Thence it passes through the throat of the venturi (32). A nozzle projects above the throat of the venturi. The spray issues from the nozzle; but the sides of the nozzle project to some extent above the outflow of the spray. The nozzle (as shown in Figure 3) has two feeds: one for liquid and one for gas. The gas entering the assembly is in turn divided into two streams. One stream atomises the spray. The other creates a gas wall. The specification describes this as follows:

"Intake 22 leads to an atomizing gas reservoir 64 which serves as an equalizer chamber for gas pressure. This equalizer chamber is formed from two parts of the atomizer, to wit, the atomization nozzle body 62 and a hexagonally shaped fluid nozzle 63. Fluid nozzle 63 is equipped with twelve exit ducts, six of which ducts (indicated by reference numeral 65 in FIG 3) lead to an equalizer reservoir 71 which the other six (indicated by reference numeral 66 in FIG 1) lead to air nozzle 67.

Air nozzle 67 fits over fluid nozzle 63 and atomization nozzle body 62. Threaded retaining ring 68 secures air nozzle 67 to fluid nozzle 63. Slipped over the outside of air nozzle 67 is an inner nozzle protector sleeve 69. An outer nozzle protector sleeve is formed by the uppermost edges of the gas foil guidance element 48 itself. Equalizer reservoir 71 is thus provided which insures that a continuous gas wall will initially circumscribe the atomized coating material at it issues forth from exit port 58."

Figure 3 is reproduced below.

Suzukawa is a Japanese patent application (JP47-7442). It is concerned with a method for granulation in a fluidized bed and a device for implementing that method. The method and device are said (page 2 paragraph 6) to overcome defects in the prior art and in particular (paragraph 4) to be suitable for the manufacture of small "grains" at low cost. It works by:

"causing a forced circulation motion of the solid material, liquid material, or solid powder material in the fluidized bed by means of a gas jet stream introduced so that it points upward from the floor of the fluidized bed."

"Gas stream" is defined thus:

"Gas jet stream means the flow of the gas that is significantly faster than the converted superficial linear velocity within the fluidized bed and does not pass through a baffle plate."

The experts agreed that a "baffle plate" is the same as a gas distributor. The patent goes on to explain that gas streams ejecting from the floor of the fluidized bed are not the only means of the forced circulation in the fluidized bed.

"Gas jet streams for the selective removal of granular material, gas jet streams for carrying raw materials, and gas jet streams ejecting from the perimeter of the liquid material atomizing nozzles can also be used as gas jet streams for creating forced circulation. In order to attain still stronger forced circulation in the fluidized bed, it is also possible to use two or more gas jet streams in combination.

In this invention, liquid material is supplied from a triple-walled tube, described below. Thus, a liquid material atomizing gas jet stream is ejected from around the mouth of a liquid material ejector, and further a gas jet stream is ejected from around that. The effect of this gas jet stream will be described below.

Since a gas jet stream sprays from the perimeter of the liquid material atomizing nozzle the spatial density of the solid particles in the vicinity of this nozzle is more diffuse than that of other areas, and has the effect of promoting diffusion of the liquid material within the bed.

Moreover another feature of this invention that should be emphasized is that the aforesaid liquid material ejector port tip is disposed within the fluidized bed. Thus, according to the dispersion method within the bed of the liquid material that is described in the invention, even though the ejector port is inserted inside the fluidized bed, it is protected by the ejection of the gas jet stream from its vicinity and the ejector port tip does not get clogged by powder or other material adhering to it."

The working of the invention is further described as follows:

"Specifically, using triple-walled dispersion device 14, the gas jet stream sent through the outside tube 15a induces a forced circulation motion in the solid particles and solid powder material in the vicinity of liquid material ejector port 14b, while a void forms in the vicinity of liquid material ejection port 14b makes it possible to increase dispersion velocity and to expand the dispersion region."

The experts agreed that the "triple-walled tube" is a three fluid nozzle of a kind known in 1991; and that "dispersion velocity" means the rate at which liquid may be added.

Kurita is another Japanese patent application (JP2-90957). It is concerned with a spray nozzle. It is a three fluid nozzle. The problem described by Kurita is that:

"if the movement of particles in the area around the spray nozzle is poor, large particles will easily be formed in the granulation operation, and the particles will easily conglomerate in the coating process. Furthermore, the inventors have also discovered that liquid will solidify and particles will adhere to the area around the spray port of the spray nozzle which may cause plugging of the spray nozzle"

The invention consists of a spray nozzle for coating granules. It comprises a first fluid spray channel, a second fluid spray channel which sprays a second fluid and a third fluid spray channel which sprays a third fluid. The first and second spray channels are used for liquid and gas respectively.

The function of the third spray channel is defined by the patent as follows:

"This third fluid spray channel 10 protects the fine mist produced by the first fluid spray channel 2 and the second fluid spray channel 5 and sprays a gas (secondary gas) for activating the behaviour of the particles which are in proximity to the spray direction of the spray nozzle 1, as shown by the double dot line."

Naunapper is a German patent (DE3323418). It is referred to in the patent in suit as the starting point for the claimed invention. This patent describes a device for detecting the occurrence of "sticking" during fluidized bed coating or granulation; and in the event that sticking is detected further proposes how corrective action can be taken. It does not claim to be able to prevent the occurrence of "sticking" in the first place. As it explains:

"The task of the current invention is to provide a device ... which enables more reliable monitoring of operational disruptions, particularly from reduced product movement within the production zone."

Naunapper includes a Figure 7. It uses monitoring equipment with sensors which are similar to those used in Figures 1 to 6. The left nozzle in Figure 7 is described as being a multi-substance nozzle 5a, which sprays liquid (arrow Pf3) as well as air (arrow Pf4). This nozzle has an "additional air casing". The right nozzle does not have any additional air casing. The patent does not draw any particular attention to the nozzle on the left, other than to mention it. Nor does it indicate what the purpose of the air casing is.

Section 75 of the Patents Act 1977 gives the court power, in proceedings in which the validity of a patent is put in issue, to allow the proprietor of the patent to amend the specification of the patent "in such manner, and subject to such terms ... as the court ... thinks fit". Plainly this is a wide statutory discretion. The principles on which the discretion should be exercised were summarised by Aldous J in Smith, Kline & French Laboratories Ltd v Evans Medical Ltd [1989] FSR 561 as follows:

"First, the onus to establish that amendment should be allowed is upon the patentee and full disclosure must be made of all relevant matters. If there is a failure to disclose all the relevant matters, amendment will be refused.

Secondly, amendment will be allowed provided the amendments are permitted under the Act and no circumstances arise which would lead the court to refuse the amendment.

Thirdly, it is in the public interest that amendment is sought promptly. Thus, in cases where a patentee delays for an unreasonable period before seeking amendment, it will not be allowed unless the patentee shows reasonable grounds for his delay. Such includes cases where a patentee believed that amendment was not necessary and had reasonable grounds for that belief.

Fourthly, a patentee who seeks to obtain an unfair advantage from a patent, which he knows or should have known should be amended, will not be allowed to amend. Such a case is where a patentee threatens an infringer with his unamended patent after he knows or should have known of the need to amend.

Fifthly, the court is concerned with the conduct of the patentee and not with the merit of the invention."

Although this case was decided under the 1949 Act, the court's discretion is as wide under the 1977 Act as it was under the 1949 Act; and is not (yet) constrained by the principles applicable under the European Patent Convention: Kimberly-Clark Worldwide Inc v Proctor & Gamble Ltd [2000] RPC 422. This may change when section 75 (5) of the Patents Act 1977 is brought into force. Most of the considerations relevant to the exercise of discretion to which Aldous J referred are aspects of the conduct of the patentee.

The courts have also drawn a distinction between a so-called "deleting" amendment and a so-called "rewriting" amendment. A deleting amendment is one that deletes invalid claims from a patent that contains some valid claims. A rewriting amendment is one that reformulates existing claims so as to validate them. In other words, absent the rewriting, the claim is invalid. The discretion, it is said, is more likely to be exercised in favour of the former kind of amendment than the latter.

Mr Davis submitted that there were sound reasons for this distinction. A competitor or would-be competitor of the patentee will know about the extent of the monopoly claimed by the patent as granted. He may legitimately take the view that a particular claim (say, claim 1) in the patent as granted is too wide to be valid. But he may take the view that a narrower, dependent, claim (say, claim 4) is valid; and take steps to design his own product or process around that claim. If claim 1 is ultimately deleted because it is too wide, he suffers no harm, because he has already disregarded it in his own activities. But if claim 1 is rewritten, so as to make it narrow enough to be valid, a number of evils may result. First, he is faced with a claim the precise scope of which he could not have foreseen. Second, he is exposed to the danger that the patentee will rewrite his claim with an eye to potential infringement by a known product or process, so that the rewriting may have a specific target which it would not otherwise have had. These are fair points to make, but on the other hand their force is to some extent blunted by the very existence of the statutory power to permit post-grant amendments.

In the present case I am concerned with rewriting amendments.

The underlying purpose of the power to permit post-grant amendment must be to allow the patentee to validate what would otherwise be an invalid or partially invalid patent. The grant of a patent is a trade-off between adding to the stock of public knowledge in the shape of the specification on the one hand, and the grant of a twenty year monopoly to the patentee on the other. The essential bargain between the community and the inventor, which is necessary to encourage invention and innovation, would be broken if the public received the teaching without having to cede the monopoly. It is, therefore a power conferred on the court for the benefit of the patentee. On one reading of the cases there seems to be a penitential theme running through the some judges' observations. Unless the patentee confesses all his sins, he cannot receive absolution. However, in Kimberly-Clark Worldwide Inc v Proctor & Gamble Ltd at 436 Aldous LJ firmly rejected the notion that the discretion to refuse amendment was based on moral or quasi-moral grounds. The rationale underlying the court's discretion to refuse permission to amend is "the duty of the court to protect the public from abuse of monopolies". Where a patentee has claimed a monopoly over an area that is unjustifiably wide, he may have deterred research and experiment that would have been carried out but for his unjustified claim. That is against the public interest. Where, however, there has been no abuse of a monopoly, there would seem to be no reason to refuse an amendment in order to protect a patentable invention that would otherwise be unprotected.

It is in that context that the conduct of the patentee is relevant. If he has persisted in defending or asserting claims that he knows are unjustifiably wide, he may be guilty of what in the traditional (but rather archaic and moralistic) language patent lawyers call "covetousness" or "covetous claiming". The implication is that the patentee is claiming to own something that belongs to someone else (e.g. the public). Likewise, if he has delayed too long in seeking to amend, he may have effectively persisted in maintaining claims that he knows are too wide. Nevertheless, in the context of a deleting amendment, the only effect of refusing permission to amend will be that the patent will continue to contain both valid and invalid claims, although a partially invalid patent is, under current practice, generally revoked. In the case of a rewriting amendment, the effect of refusing permission to amend will be to deprive protection to an invention which could have been protected. It seems to me that in considering whether to exercise the discretion to refuse amendment in circumstances in which that will be the result, the court ought to take into account the proportionality of the alleged culpability of the patentee as compared with the effect of loss of protection for his invention.

Aldous LJ returned to the question of disclosure in Oxford Gene Technology Ltd v Affymetrix Inc (No 2) [2001] RPC 9. He said:

The duty of disclosure is confined to what is relevant to the stated reason for the amendment. There is no duty (or need) to trawl through documents to see whether there is anything else that might bear on the exercise of the court's discretion. Where an application to amend is opposed, the grounds of opposition may, of course, give rise to matters on which disclosure should be made. And as Aldous LJ points out, grounds of abuse must be properly particularised. The particulars of the objections will normally delimit the scope of required disclosure.

In Baker Hughes Inc v Halliburton Energy Services Inc [2002] EWHC 2524 Jacob J pointed out that even where there is a duty to disclose, that does not require disclosure of "every single document" and that it is unnecessary that "every single witness should be produced". There is, in effect, a moving evidential burden. Once the patentee has investigated the position, the evidential burden may shift to the opponent. As Jacob J put it:

"The matter must be considered in proportion."

Mr Davis submitted that the duty to disclose extended to matters within the knowledge or possession of opponents to the amendment. He relied in this respect on the decision of Mr Bridges, Divisional Director, acting for the Comptroller in Kaiser's Application (O/279/00) para 26:

"the patentee is under an obligation to make the whole story known. This includes information which is already in the possession of the other party to the proceedings, because it is not good enough for an applicant to support an amendment by mere argument, even if the other side should agree to it. It is for the comptroller to decide what is relevant and it is clearly in the patentee's interest to err on the side of caution in deciding what to disclose."

In that case Mr Bridges came to the conclusion that because the patentee had not disclosed a veiled threat to enforce the unamended patent made at a time when the patentee had no reason to believe that it was invalid, discretion to amend should be exercised against the patentee, even though, by the time that the discretion came to be exercised, the hearing officer was in full possession of all the facts. As Mr Bridges put it (para 43) the patentee had not "fulfilled the conditions necessary for discretion to be exercised to amend the patent". He therefore refused permission to amend.

I am bound to say that this strikes me as a punitive approach to the exercise of a discretion which plainly exists for the benefit of the patentee. A duty to disclose may be broken in a variety of ways, some of which may be important and some of which may be relatively trivial. The reasons for non-disclosure may likewise be various, from deliberate concealment of facts known to be relevant and damaging at one end of the spectrum to misunderstanding what the duty entails at the other. If, within material known to an opponent, there is information telling against the grant of permission to amend, one would expect the opponent to raise it in his grounds of opposition. If any failure to comply with the duty to disclose is to be treated as a failure to comply with "a condition" and hence results in the refusal of permission to amend, proportionality has gone out of the window. I do not doubt that failure by the patentee to disclose relevant matters is a factor which tells against the exercise of discretion in his favour. Depending on the gravity of the failure and the reasons for the failure, it may prove decisive. However, I do not consider that it is right to read Aldous LJ's guideline as amounting to a statutory pre-condition to be strictly fulfilled on pain of refusal of permission to amend. In this connection there is also some difficulty in applying the analogy with an application for a freezing order or for relief without notice. It is indisputable that a party has a duty of full disclosure in such circumstances. But there are three things that make the analogy difficult to apply. First, at the stage when the application is made, almost inevitably the court is in no position to decide whether there has been a failure to comply with the duty or not. In the usual case, it is only once the freezing order has been granted that the question arises, and then on an application to discharge an order that has already been made. Second, even where there has been a breach of the duty to disclose, the court still retains a discretion to continue (or, perhaps technically, to reimpose) the order. Third, the duty to disclose does not apply with the same stringency to an opposed application.

In cases where delay is alleged, there may be a difference according to whether the opponent or the public has suffered detriment as a result of the delay. If detriment is established, then the permissible period of delay may be shorter than in a case where no such detriment has been established. In Matbro Ltd v Michigan (Great Britain) Ltd [1973] RPC 823, 833 Graham J said:

"It was I think accepted by all parties that mere delay is not, of itself, necessarily sufficient to justify refusal of amendment. There must have been or be likely to be some detriment to the respondents or to the general public caused by such delay before it can be an effective bar to relief. Such detriment will normally in the case of respondents be the subject of evidence. In the case of the general public there may also well be evidence present from which the court can draw an inference of detriment, but such an inference may, it seems to me, also be drawn by the court without evidence in a proper case."

Moreover, as Aldous J pointed out in Smith Kline, a lapse of time during which the patentee knows about the prior art in question but believed, on reasonable grounds, that his patent was valid will not count as unreasonable delay, if it counts as relevant delay at all.

What is an acceptable period of delay is plainly a question of fact; and will depend on all the circumstances. In Smith Kline itself, a delay of eight years was held to be too long, even though no prejudice to the opponent or the public had been established. Mr Davis came close to submitting that a period of two months from the date on which the patentee became aware of the need to amend was, in general, a sufficient time within which to formulate an amendment. He based this primarily on the observations of Pumfrey J in Instance v CCL Label Inc [2002] FSR 27. In that case the judge said at para 33:

Counsel's advice in that case was that the patent had been anticipated; and the patentee's patent agent took the view that the case for anticipation was a strong one. Moreover, during the period that elapsed between counsel's advice and the application to amend, the unamended patent had been deployed against competitors. That, therefore, was a case in which detriment was established. I cannot regard Pumfrey J as having laid down a general rule, or anything close to one.

The proposed amendment is an amendment to claim 1 (by the replacement of "and/or" with "and") and the insertion of a series of dependent claims, each one narrower than the last. Mr Davis submits that this form of cascade of claims is an impermissible form of amendment. He relied on observations of Aldous LJ in Lubrizol Corp v Esso Petroleum Ltd [1998] RPC 727. That was a case in which the Court of Appeal held that the patent in suit was invalid. The question then arose whether the patentee would be entitled to relief under section 63 of the Patents Act 1977, which applies only when a patent has been found to be "partially valid". If relief was not available under this section then the patentee did not wish to amend. So the issue before the court was whether the patent was partially valid. Aldous LJ said (p. 790):

There are a number of reasons why these observations do not support Mr Davis' proposition. First, the issue before the court was whether the patent in suit was partially valid within the meaning of section 63. It was not an application to amend, but a preliminary to a potential application. The patentee did not want to amend unless the patent was partially valid. Second, Mr Thorley's submission did not relate to an actual application to amend. The alternative forms of claim to which he referred were potential alternatives depending on the outcome of the court's decision on the issues (hence Aldous LJ's reference to "drafts of claims", which he appears to contrast with "a series of claims"). So he was not addressing a formulated cascade of claims at all. Third, the principle underlying Aldous LJ's conclusion is that of abuse of process, in the sense of having multiple suits where one would have been enough. If the patentee applies to amend before trial, then he does what Aldous LJ says he should do, and brings forward his whole case. This is what Terrell on Patents recommends at para 9-66.

I can see no objection in principle to an amendment formulated as a cascade of claims. If, as a matter of case management, the court decides not to consider the whole cascade, that is a different matter; and, moreover, a matter of discretion.

The sole witness on the question of amendment was Dr Armin Prasch, who takes responsibility within the Glatt group of companies for the prosecution of European patents. He also has responsibility for the prosecution of patent litigation; and in particular for parallel proceedings in Germany in which Glatt has sued Vector, relying on the patent as granted.

He dealt first with the date on which Glatt first became aware of the pleaded prior art. Mr Naunapper (the inventor of "Naunapper") was an employee of Glatt; and Naunapper was cited in the patent as granted. So was Larson. So Glatt knew of these pieces of prior art right from the start. But the draftsman of the patent in suit referred to both of them in the specification itself; and plainly thought that he had adequately distinguished both of them. So, presumably, did the EPO because the patent was granted. Dr Prasch said that Glatt did not become aware of Kurita until it was introduced in evidence by Vector in the German proceedings in September 2005. Likewise Dr Prasch said that Glatt did not become aware of Suzukawa until it was introduced by Vector into the German proceedings in February 2006. I accept his evidence. Dr Prasch continued:

"Glatt has always considered the Patent valid in the light of Naunapper and Larson, particularly as they were both cited during prosecution in the EPO. Although the two Japanese patents have only recently become known to Glatt, we have been advised that the Patent is valid in the light of those patents as well. Glatt is seeking to amend the patent so as further to distinguish the claimed invention over the pleaded prior art. The amendment was originally sought so as to further distinguish Kurita in particular although I am advised that the same amendment also further distinguishes Suzukawa (and indeed Naunapper and Larson)."

As is well known in this field, in Germany infringement and validity are considered by two separate courts. In the present case, Glatt has sued Vector in Germany for infringement. Dr Prasch explained (without challenge) that in those proceedings Glatt has complied with all its obligations about disclosure; and that under German law amendments to a patent may be made at any stage during validity proceedings. He also says (again without challenge) that, apart from the proceedings in Germany against Vector, Glatt has not sued or threatened to sue anyone else in relation to the patent in suit or related patents anywhere in the world.

Mr Davis strongly criticised Glatt's failure to put the whole story before the court. He said that the burden lay on Glatt to persuade the court to exercise its discretion in its favour. If the whole story was not laid before the court, Glatt had not discharged the burden of proof. That, he submitted, is what happened in this case.

He relied in particular on the following:

i) Dr Prasch did not review the "file wrapper" either for the European patent or its American equivalents. However, the file wrappers (or at least parts of them) were put in evidence and did not, so far as I could see, contain anything which contradicted what Dr Prasch had said about the date or dates on which Glatt became aware of the cited prior art. Moreover, Dr Prasch said that although he personally had not reviewed the US files, they had been reviewed by Glatt's patent attorneys. I do not attribute any significant weight to this complaint which, moreover, was not pleaded.

ii) Dr Prasch had not investigated the prosecution of the Japanese patent. This might have been of relevance, because both Suzukawa and Kurita were Japanese pieces of prior art, and might have been cited in Japan. As I understood Dr Prasch's evidence, the Japanese patent was prosecuted by Japanese patent attorneys who reviewed the file. I do not attribute any significant weight to this complaint which, again, was not pleaded.

iii) Glatt did not conduct a search for relevant documents at the offices of its patent agents, or at the offices of its US premises. Nor has it disclosed any e-mails. It is not at all clear what it is that Glatt is supposed to have been looking for in these searches that it did not conduct. As Dr Prasch explained the various patent attorneys reviewed their files. The US premises was not the place from which European patents were prosecuted: that was the responsibility of Dr Prasch's office. So far as e-mails are concerned, it is true that Dr Prasch said in cross-examination that he communicated with the USA by e-mail, but that does not indicate that he communicated anything of relevance to this application.

iv) Dr Prasch did not produce any information about the grant of licences. Mr Davis submitted that if there were any (as to which there is no evidence) they might have taken advantage of the unamended patent across its full width (as to which there is again no evidence) and thus resulted in Glatt abusing a wider monopoly than that to which it was properly entitled (as to which there is again no evidence). Nor was this complaint pleaded. It is pure speculation.

v) Dr Prasch did not explain what Glatt was "up to" in Germany. However, since Vector were the only other party to the German proceedings, and if there had been anything relevant to the application to amend, they could have produced it, this does not seem to me to be a weighty complaint.

vi) Although Dr Prasch refers to advice that Glatt has received, he did not indicate the name or nature of the adviser nor the substance of the advice. However, revelation of these aspects of the advice would have involved a waiver of legal privilege. It is clear that a patentee who applies to amend is not bound to waive prejudice, and that no adverse inference should be drawn against him if he declines to do so.

Mr Davis summarised his contention thus:

"As is often the case when faced with instances of non-disclosure, it is very difficult to know what has not been disclosed for the very reason that it has not been disclosed. However, it is entirely possible that, in the relatively long life of this patent, there were meetings to review IP assets, meetings with licensees, correspondence with various third parties (including those suspected of infringing the patent) and other instances of the patentee considering its patent. It is up to the patentee to put the whole story of the patent before the court and it is completely insufficient simply to state "the Defendant is unaware of any reason why the Court's discretion should be exercised adversely to it"."

The converse of this submission is that Glatt would be required to prove a negative. Without undertaking a wholly disproportionate search of all its records world-wide, and those of its agents, and stretching back over fifteen years, I do not see how it would have been possible for Glatt to have met Mr Davis' stringent criteria. Although it must be acknowledged that the evidence adduced could have been fuller, having heard Dr Prasch cross-examined, I do not consider that I have been told anything other than the full relevant story.

Mr Davis relies on delay in three different ways. First, he says that to the extent that the amendments are designed to distinguish the patent in suit from Larson, the amendment is too late. Glatt have admittedly known about Larson for years, and cannot be permitted at this late stage to amend round it. Second, to the extent that the teaching of Kurita and Suzukawa do no more than reflect common general knowledge at the priority date, it is equally far too late to amend to distinguish the patent in suit from common general knowledge. Third, he says, taking at face value Dr Prasch's evidence that it was not until September 2005 that Glatt became aware of Kurita, and that it was not until December 2005 that it received advice that the patent should be amended, the gap between that advice and the application made was still too long.

I think that Mr Davis is right in saying that the additional limitation introduced into claim 4 is designed to distinguish the patent in suit from Larson. I think that Dr Prasch accepted this, although he rightly pointed out that claim 4 was one of a series of dependent claims whose primary purpose was to distinguish the patent in suit from Kurita. However, as I have said, the draftsman of the patent in suit must have considered that he had sufficiently distinguished Larson in the patent as granted; and so, for that matter, must the EPO. Moreover, Glatt has maintained (and continues to maintain) that the patent as granted is valid. Bearing in mind that the application to amend is unconditional, I cannot see why the patentee should not be allowed to amend what it says is a valid claim in order to narrow the scope of the monopoly, especially as the interpretation of Larson on which the amendment depends is one that is still contentious. So far as common general knowledge is concerned, this submission is predicated on the assumption that what Professor Hounslow said in cross-examination about common general knowledge of three fluid nozzles is correct. But Mr Davis does not accept that what Professor Hounslow said on this subject is correct. Moreover, the views expressed by Professor Hounslow are those of an independent expert and cannot be attributed to Glatt itself. But even if they could, I do not see that they can be attributed to Glatt earlier than Professor Hounslow expressed those views, which was not until his cross-examination. So in my judgment there was no relevant period of delay. That leaves the third aspect of delay. The position here is that Glatt was advised in December 2005 that the patent should be amended. A conditional application to amend was made in January 2006; and that turned into an unconditional application in February with differently formulated amendments. I do not regard that as delay sufficient to result in a refusal to exercise discretion in Glatt's favour.

Mr Davis submits that the fact that Glatt has sued Vector in Germany relying on the patent as granted means that Glatt has taken unfair advantage of too extensive a monopoly. It has been guilty of covetous claiming. As Mr Davis stressed the law in Germany and the United Kingdom is the same. Both member states are applying the European Patent Convention. It is common ground that Glatt have done nothing that German law does not permit them to do. So Mr Davis' proposition comes to this: that to conduct patent litigation according to German procedure is an abuse of a monopoly. I do not accept this submission. To hold otherwise would be an affront to comity. Moreover, I do not know whether the alleged infringement in Germany would still have fallen within the narrower ambit of the claims as proposed to be amended. I do not consider that covetous claiming has been established.

I conclude that, if I can, I should exercise my discretion in Glatt's favour and allow the amendments.

Mr Davis did, however, have a fall-back position. Even if, in principle, I came to the conclusion (as I have) that I should exercise my discretion in Glatt's favour, nevertheless I cannot because the patent as amended discloses matter that was not disclosed in the application as filed.

This is a statutory bar on amendments, to which Aldous J referred in the second of his guidelines in Smith Kline. An amendment is not permitted under section 75 of the Patents Act 1977 if, in the words of section 76 (3), it results in the specification disclosing additional matter. Section 76 (3) does not itself define "additional matter"; but it is common ground that it has the same meaning as in section 76 (1) which prohibits an amendment if it "discloses additional matter, that is, matter extending beyond that disclosed in ... the application for the patent as filed." The specification is normally the place for disclosure, whereas the claim normally defines the scope of the claimed monopoly, which need not necessarily be of co-extensive width with the disclosure: Texas Iron Works Inc's Patent [2000] RPC 207, 245. However, this is not a hard and fast rule. The amendment of a claim, including the deletion of a feature or features of a claim, may amount to the disclosure of additional matter: Southco Inc v Dzus Fastener Europe Ltd [1992] RPC 299, 324.

It is common ground that the test to be applied is that stated by Aldous J in Bonzel v Intervention Ltd [1991] RPC 553, 554 (which I have modified slightly to take account of the fact that this is an application to amend a granted patent):

"The decision as to whether there was an extension of disclosure must be made on a comparison of the two documents read through the eyes of a skilled addressee. The task of the Court is threefold:

(a) To ascertain through the eyes of the skilled addressee what is disclosed, both explicitly and implicitly in the application.

(b) To do the same in respect of the patent as [proposed to be amended].

(c) To compare the two disclosures and decide whether any subject matter relevant to the invention has been added whether by deletion or addition.

The comparison is strict in the sense that subject matter will be added unless such matter is clearly and unambiguously disclosed in the application either explicitly or implicitly."

The specification in the application and the patent as granted is, for practical purposes the same. So is the specification as proposed to be amended. I did not understand Mr Davis to argue otherwise. He concentrated on the claims. So far as claim 1 is concerned, the effect of the amendment is that the claimed invention requires both the "means" and the "gas stream" to act together to shield the initial spray pattern. The application in my judgment also requires both. It points out that in the cited prior art, some patents do not disclose protection of the spray pattern "by an upwardly moving column of air", and others do not disclose a structure having this effect. The essence of the invention, as described in the specification in the application, is:

"Accordingly, the present invention introduces a shielding or barrier means around the lower portion of the nozzle, within the upbed, for shielding the nozzle and allowing up flow of air within the shielding means around the nozzle."

It refers both to the shield and to the "columnar shield of upwardly moving air". The description of the embodiment also includes:

"The partition 40, as well as the tubular column of upwardly rising air about the spray nozzle assembly 32, shields the lower, beginning portion, of the spray pattern 56."

This passage also points to both the physical partition and the air shielding the spray pattern in combination. Claim 1 of the application also appears to me to require both the means and the air flow.

It is true that some of the features that are described in the embodiment illustrated by Figure 1 do not feature in amended claim 1. Nor do they feature in claim 1 as originally filed. But they do appear in the embodiment described in the application. However, I accept Mr Campbell's submission that this is not a case of deleting features; it is simply a question of not claiming them. The features are progressively added into the new dependent claims.

I do not regard this as a case of added matter, even if the case is put on the basis of an intermediate generalization.

Mr Davis had a separate point on claim 13 relating to added matter. The point is this. Claim 1 of the patent application as filed required the combination of both the physical shield (the means) and also the upwardly moving column of gas (the gas stream). I have held that this construction of the application is correct. Although claim 13 requires both a means and a gas stream, it is only the gas stream that shields the initial spray pattern. Thus, says Mr Davis, there is added matter, because the application does not reveal that the gas stream alone can shield the initial spray pattern. The place to look for disclosure is the specification. One passage in the specification reads:

"In addition, although the shielding of the spray nozzle is preferably provided by an inner cylindrical partition 40, other shielding arrangements may be utilized. For example, shielding of the spray nozzle may be accomplished by an air wall or stream that surrounds the nozzle and prevents particles from prematurely entering the spray pattern."

In my judgment this is sufficient disclosure to support amended claim 13. This objection fails.

Mr Davis' second point on claim 13 was that it was prohibited by section 76 (3)(b) of the Patents Act 1977. This prohibits an amendment which "extends the scope of the patent". The reference here to "the patent" is a reference to the patent as granted; not to the claims as they appeared in the application as filed.

Claim 1 of the patent as granted contained the phrase "and/or", which is now to be deleted in amended claim 1. If there are two features, X and Y, "X and/or Y" means, X alone, Y alone or both X and Y together. Claim 13 as proposed to be amended limits the claim to a shield consisting of the gas stream, without the means performing a shielding function. However, since claim 1 of the patent as granted allowed for either alternative independently, I do not consider that amended claim 13 widens the scope of protection. This objection also fails.

A patent may only be granted for an invention if it is "new": Patents Act 1977 s. 1 (1) (a). If an invention is not "new" it is said to have been anticipated. The classic statement of principle is that of the Court of Appeal in General Tire & Rubber Company v. Firestone Tyre & Rubber Company Ltd [1972] RPC 457, 485 as follows:

Lord Hoffmann summarised the requirement of disclosure as follows:

He approved the following statement of principle by the Technical Board of Appeal in Union Carbide [1999] EPOR 312:

Lord Hoffmann went on to summarise the requirement of enablement as follows:

Lord Hoffmann pointed out that in considering whether the requirement of enablement had been satisfied:

The experts agreed that Larson does not disclose a gas distribution plate or a screen. Mr Davis argued that his own expert, Professor Geldart, was wrong, because he had misunderstood the meaning of "gas distribution plate or screen" in the patent in suit. This was a bold submission. If both experts understood the significance of this integer in the same way, why would the skilled addressee understand it any differently? It is all the bolder since Larson itself emphasises that it does not use a gas distribution plate, but a venturi instead. In any event, I have rejected Mr Davis' argument about the meaning of the quoted phrase in the patent in suit. Since this is an integer of the claim, it follows that Larson does not anticipate. There was also some debate whether Larson disclosed a plenum chamber. Professor Hounslow explained that the rate at which air entered that part of the Larson apparatus (which is called a "duct" rather than a plenum chamber in Larson itself) that was said to constitute the plenum chamber would have entered at such velocity as to cause turbulence incompatible with the function of a plenum chamber. I found his explanation cogent, and I accept it, despite Professor Geldart's disagreement. In my judgment Larson does not disclose a plenum chamber either.

Larson does not therefore anticipate.

The essence of the invention claimed by Suzukawa is that it facilitates a greater liquid dispersion rate; that is to say, more liquid can be injected into the fluidizing bed. It does this by means of a three fluid nozzle. Professor Geldart said that what was shown by Suzukawa would not be described as a nozzle with something (i.e. a means) attached. I agree. Suzukawa does not, in my judgment, disclose a means adjacent to the nozzle. Moreover, the nozzle is not intended to shield the spray pattern, although in itself this lack of intention might not matter, if it in fact shielded the spray pattern. But Professor Hounslow said that there was in fact no shielding. The real effect of Suzukawa was to increase bulk motion of the fluidizing bed itself. The induction of bulk motion does not shield the particles; it merely makes them move faster. Professor Geldart agreed that the shielding effect of Suzukawa was "somewhere between negligible and non-existent". He considered that in order to obtain any significant shielding effect, the outer tube would have to be significantly extended. Although he said that extending the outer tube in this way is the "sort of adjustable parameter which would be varied during experimentation", that kind of experimentation is not the kind that is contemplated in a novelty case. Although the skilled addressee may be assumed to carry out trial and error experiments to make the prior art work, he is not to be assumed to carry out trial and error experiments to turn the prior art into something else.

In my judgment Suzukawa does not anticipate.

Professor Hounslow (with whom Professor Geldart agreed) considered that the intention of Kurita is two fold: first, to improve movement of particles in the area around the spray nozzle, and second to ensure that the nozzle does not foul itself. The mechanism by which it does this is by what it calls the "activation" of the particles. Professor Geldart explained that this meant causing the movement of particles, to stop them settling in any one region. The reason for this was that whenever particles in a fluidized bed land on an upwardly facing surface where there is no gas, they tend to become defluidized and settle down to become a packed mass. In essence, Professor Hounslow agreed.

What, therefore, Kurita is concerned with is keeping particles away from the nozzle; not keeping particles away from the spray.

What Kurita describes is a nozzle. It does not describe a nozzle plus a means. Although it may be said to describe a gas stream, the gas stream does not "surround" the nozzle. At best it surrounds the spray port. In addition, Kurita does not disclose any shielding effect of a magnitude that would be of any significant effect.

In my judgment Kurita does not anticipate.

Naunapper refers to an "additional air casing". Mr Davis argued that the additional air casing shielded the nozzle because "put simply, it is a bit of metal and if a particle strikes it first then it will not move in to the region just above the nozzle where the initial spray pattern is. The concept of an -�air casing' is plainly addressed to something which is supposed to encase the nozzle tip and, hence, the initial spray pattern."

Professor Geldart said in his report that the pipe providing the air casing appeared to terminate at a level below the end of the two-phase nozzle; hence any shielding would be "very limited". He resiled to some extent from this in cross-examination, saying that the shielding would be "limited" rather than "very limited". But Professor Hounslow was firmly of the view that there was no shielding. He said that he could not identify a means which might prevent the entrance of particles, nor is there an air stream to do so. He did not modify his view in cross-examination. I prefer Professor Hounslow's view.

In my judgment Naunapper does not anticipate.

In my judgment the arguments in favour of anticipation suffer from the defects identified by the Technical Board of Appeal in Union Carbide. They selected individual features from the prior art, and applied them for different purposes to the patent in suit. The arguments were in my judgment strongly coloured by hindsight. In the absence of the patent in suit, I do not consider that the pleaded prior art contained "clear and unmistakable directions" to do that which, if done, would have infringed the claims in the patent in suit. I conclude therefore, that the patent in suit was not anticipated.

A patent may only be granted for an invention if it involves an inventive step: Patents Act 1977 s 1 (1) (b). An invention does not involve an inventive step if it is "obvious to a person skilled in the art": Patents Act 1977 s. 3. An attack on the ground of obviousness must be distinguished from one based on lack of novelty. The classic description of the difference between the two is again that of Sachs LJ in General Tire:

In Synthon Lord Hoffmann added:

"If performance of an invention disclosed by the prior art would not infringe the patent but the prior art would make it obvious to a skilled person how he might make adaptations which resulted in an infringing invention, then the patent may be invalid for lack of an inventive step but not for lack of novelty."

The question whether a step was obvious has been described as a jury question. In Windsurfing International Inc v. Tabur Marine (Great Britain) Ltd [1985] RPC 59, 73 Oliver LJ said:

"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. Thereafter, 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 degree of invention."

In commenting on the last (and crucial) Windsurfing step, in Molnlycke AB v. Procter & Gamble Ltd [1994] RPC 49 Sir Donald Nicholls V-C said that:

"obviousness connotes something which would at once occur to a person skilled in the art who was desirous of accomplishing the end."

It was common ground between the experts that the behaviour of fluidized beds is not well understood in scientific terms; and that the behaviour of particles within such beds is almost impossible to predict, even with today's computer modelling. Although the science of fluidization has ceased to be a "black art", it is still very much empirically based, and reliant on trial and error. Partly for this reason the community of those involved in fluidization is conservative in outlook. As it is put in Kunii and Levenspiel on Fluidization Engineering (p. 19):

"On looking back, it seems that the path to commercial success with fluidized processes has been unusually painful and complex, with many stages of scale-up and more than its expected share of embarrassing failures. These difficulties stem largely from the lack of satisfactory answers to the many questions on which design decisions should be based, and this in turn stems from a lack of reliable predictive knowledge about what goes on in these beds. Such design uncertainties, coupled with the large investment involved, have led to general conservatism and caution in many of these developments."

This mindset means, in my judgment, that the skilled addressee would be very wary of modifying designs on paper in the prior art.

The inventive concept claimed by the patent is that particles should be kept away from the region near the spray by a combination of a barrier (a "means") and a gas flow. This both protects the particles (thereby preventing them from agglomerating) and also protects the nozzle (thereby preventing it from becoming abraded). What, in my judgment, was particularly inventive about it was that Mr Jones had identified the entry of particles into the wet zone at the inception of the spray pattern as being a rate limiting factor for the introduction of spray; and had realised that the way to overcome that rate limitation was to stop particles from entering the spray pattern before it had become more fully dispersed. As Professor Hounslow put it, what the patent has taught is to think about where the particles are. Mr Jones had discovered both the cause of the problem and the way to overcome it.

An important feature of Larson was the venturi; and one of its most important claimed advantages was that it did away with the need for a gas distributor plate. At one stage in his cross-examination Professor Geldart agreed with the reasons set out in the specification itself in Larson which pointed out the advantages of the venturi over the conventional gas distribution plate. At another point, however, although he recognised that Larson taught away from the incorporation of a gas distribution plate, he said that he disagreed with the underlying reasoning. These two positions seemed to me to be self-contradictory. But even if Professor Geldart's second thoughts were better, he nevertheless agreed that the skilled reader of Larson would have taken its reasoning at face value. One of the embodiments of Larson (Figure 3) and the text of the specification refers to the continuous gas wall. Another of the figures (Figure 5) does not include it. There is nothing in the text of the specification to indicate what are the advantages of the continuous gas wall.

Mr Davis submits that the difference between the patent in suit and Suzukawa is that the former does not use a Wurster insert. Thus the only modification called for is the modification of Suzukawa to make it a Wurster fluidized bed. Alternatively, he says, the only difference is that the Suzukawa nozzle is not used in a Wurster bed, so that all that is required is to insert the Suzukawa nozzle into a Wurster bed. This formulation does not identify all the differences. On the basis of the construction of Suzukawa that I have adopted it lacks a "means" other than the nozzle itself; and there is no significant shielding. Moreover, the idea underlying Suzukawa is to make the spray pattern more dispersed. The patent in suit does not do that. What it does is to protect the particles against entry into the spray pattern.

Professor Hounslow said that Kurita was quite unlike the invention claimed in the patent in suit for the following reasons:

i) It describes a nozzle designed to be self-cleaning, whereas the patent has nothing to do with nozzles.

ii) The additional air flow from the second air channel is unlikely to have sufficient momentum to prevent particles moving into the wet zone

iii) The additional air flow will, on the contrary, attract particles into the wet zone.

Professor Geldart did not identify the differences between the patent in suit and Kurita in his report.

Apart from the differences identified by Professor Hounslow, which I agree with, there is also the difference that Kurita does not disclose or discuss shielding.

Naunapper did not disclose or discuss shielding either. Moreover, it was, as I have said, directed towards monitoring "sticking"; not to preventing it from happening in the first place.

Professor Geldart agreed that the problem of over-wetting of particles was a real problem. He also thought that it was one which had been appreciated by the industry for many years before 1991. In his view the attitude of the industry was that that "something needed to be done and it should be [the] subject of some serious research". However, he agreed that no solution had been found before the patent in suit. In the hackneyed phrase, there was, in my judgment, a "long felt want". Moreover Professor Geldart agreed that the idea of incorporating a spray nozzle surround which kept the substrate away from the spray pattern until it was more fully developed was "a clever idea". To say that the solution of a problem needed to be the subject of "some serious research" does not easily lead to the conclusion that the solution is obvious. Equally, a "clever idea" is unlikely to be obvious.

Professor Geldart's report did not really address the question of obviousness in the light of Larson. However, in cross-examination he said that if he had seen Larson and deduced from it the invention claimed by the patent he would have been pleased with himself for thinking of it. I do not regard this as supportive of a case of obviousness. Moreover, the case in favour of obviousness requires the reintroduction of the gas distributor plate, which Larson teaches against; and which Professor Geldart himself said was the most difficult part of the fluidized bed processor to design. Mr Davis recognised that the (re)introduction of a gas distributor plate would be a "retrogressive" step; but in my judgment that understates the position. Professor Hounslow was firmly of the view that the bridge between Larson and the patent in suit was not obvious. In his opinion in order to adapt Larson it would be necessary to abandon the role of the venturi (which was Larson's main claimed advantage), introduce a plenum and distributor plate, and very substantially reduce the gas velocity. Professor Hounslow did not consider that any of these steps were obvious. I accept this evidence. I conclude that the invention claimed by the patent in suit was not obvious in the light of Larson.

Suzukawa is a three fluid nozzle. Professor Hounslow explained that three fluid nozzles are not generally used in Wurster processors, because they are more expensive, slightly harder to clean and harder to set up. He did, however, agree that it would be possible to redesign the baffle plates depicted by Suzukawa in order to substitute a conventional gas distributor plate; although he questioned why anyone would want to do that. Professor Hounslow also expressed the view that if Suzukawa did not disclose a "means" (and I have held that it did not), then it was not obvious to add one. Professor Geldart's view was that if one needed to increase the physical shielding, one would ensure, by trial and error during a development stage, that the outermost pipe extended beyond the two middle pipes; and that such a modification would be routine during the development stage. This evidence was in my judgment highly coloured by hindsight. I do not accept that these would have been the thought processes of an unimaginative person skilled in the art.

I accept Professor Hounslow's evidence; and conclude that the patent in suit is not obvious in the light of Suzukawa.

Professor Hounslow said that the differences between Kurita and the patent in suit were not obvious. First, there is no teaching in Kurita to introduce a means separate to the nozzle so as to form a gas stream that surrounds the spray nozzle. It would be natural to carry on using a three-fluid nozzle if following the teaching of Kurita. Second, the idea of protecting the particles is not suggested either. It involves a conceptual change from Kurita's idea of only using the secondary fluid channel to enhance movement of the particles and prevent nozzle fouling. Professor Geldart agreed that there was a conceptual change between Kurita and the patent in suit.

However, in my judgment it was that conceptual change which was an important part of the invention claimed by the patent. In my judgment this means that arriving at the same conceptual change (without knowledge of the patent) would have been inventive. Accordingly, although Professor Geldart may be right in saying that it would have been technically simple to introduce an insert into the Kurita nozzle so as to provide a shield, there would have been no reason for the uninventive skilled person to have thought of doing so. If (as I think) what was inventive in the patent in suit was identifying the cause of the problem and what to do to overcome it; the fact that once these had been identified the solution was technically simple does not make it obvious.

I conclude that the patent is not obvious in the light of Kurita.

Professor Geldart accepted that Naunapper would have required modification to introduce shielding. But the introduction of shielding would have required the same conceptual change that I have discussed in connection with Kurita. Professor Hounslow's evidence (which I accept) was:

"The simple point is that there is no teaching in Naunapper about shielding the spray pattern from entry of particles at all. Consequently Naunapper misses the whole point of the Patent. It is to me so thoroughly obvious that the Patent is inventive beyond Naunapper, that there is little more that I can add."

In my judgment the patent is not obvious in the light of Naunapper.

Since I have decided that claim 1 is not anticipated, and does not fail for want of an inventive step, I can take the subsidiary claims shortly. I need deal only with claims 2, 3 and 13; because if Glatt succeeds on these, the other claims follow.

Claim 2 adds "a substantially cylindrical container with an upbed" within the produce container (i.e. a Wurster insert) and a "means" provided within the upbed.

There is no separate argument in relation to Larson. However, neither Suzukawa nor Kurita contain a Wurster insert. So far as Naunapper is concerned, the "means" is not provided within the upbed. These three pieces of prior art do not anticipate claim 2. Larson only anticipates if I am wrong about claim 1.

Both Suzukawa and Kurita disclose three fluid nozzles. As I have said three fluid nozzles are not generally used with Wurster inserts. If a redesign of Suzukawa went so far as to incorporate a Wurster insert, Professor Hounslow's opinion was that, although this was a possible intellectual exercise, "at this point we are effectively throwing Suzukawa away". He said in his report that the side entry nozzle used in Suzukawa would make the introduction of a Wurster insert difficult; and would be difficult to assemble and dis-assemble, thereby making cleaning difficult. Professor Geldart agreed with this when it was put to him in cross-examination. I conclude that claim 2 is not obvious in the light of Suzuku.

Kurita would suffer from an additional problem compared with Suzukawa. If it were used with a Wurster insert, its nozzle would soon block up, as both experts agreed. Professor Geldart said in cross-examination that Kurita was "less useful as a model to improve upon to get to claim 2". He agreed with paragraphs 105 and 106 of Professor Hounslow's report, when put to him:

I conclude that claim 2 is not obvious in the light of Kurita.

Professor Hounslow's evidence was that Naunapper is even further away from the dependent claims than from claim 1. I accept this evidence in relation to all the dependent claims; and conclude that claim 2 is not obvious in the light of Naunapper.

Claim 3 adds the requirement that the "shielding means" comprises "a cylindrical inner partition within the upbed" positioned "adjacent [the] spray nozzle" which allows the gas stream to "flow up within the cylindrical inner partition around the spray nozzle".

Mr Davis argued that wall (69) in Figure 3 of Larson disclosed this integer. I disagree. Even if wall (69) disclosed a cylindrical partition, the gas stream flows up outside it, and not within it. In addition, wall (69) is not adjacent the spray nozzle: it is part of the nozzle itself.

None of the other prior art discloses an upbed, so that there is no cylindrical partition "within the upbed".

I conclude that claim 3 is not anticipated.

Professor Geldart agreed with the following part of Professor Hounslow's report, when put to him:

"Larson does not disclose Integer 2 "a shielding means which comprises a cylindrical inner partition (40) within the upbed (30)". The ends of the gas foil guidance element 48 tapers inwards forming a truncated cone, and this shape deliberately permits particles with inward trajectories to get into the Wet Zone. In addition the upper wall of the air foil and the nozzle protector sleeve 69 constitute a convergent nozzle and I would expect that the air issuing from this annular space would be at very high velocity and the ensuing venturi effect ... would suck particles in to the Wet Zone."

Professor Hounslow's view was that this feature of Larson went against the purpose of the inner cylindrical partition; and that it was not obvious to modify it so as to fall within claim 3. I accept this evidence. I conclude that claim 3 is not obvious in the light of Larson.

Both experts agreed that claim 3 was not obvious in the light of either Suzukawa or Kurita. Although Professor Geldart said in cross-examination that the introduction of an inner cylindrical partition into Naunapper would have been obvious, I did not find his reasons cogent, and they seemed to me to have been tainted with hindsight. For the reasons I have given, I also conclude that claim 3 is not obvious in the light of Naunapper.

This claim requires the presence of a Wurster insert. It is, I think, common ground that my conclusion on claim 2 effectively decides the fate of claim 13 on the questions of novelty and obviousness. Accordingly, for the reasons I have given in relation to claim 2, claim 13 is neither anticipated nor obvious.

I permit Glatt to amend the claims as sought, and hold that the patent is valid. Consequently, the action fails.

Finally, may I reiterate and amplify the request in paragraph 9(e) of the Patents Court Guide? Important documents should be supplied to the trial judge in electronic format. These will include skeleton arguments, witness statements, and experts' reports (preferably in Word format). The patent in suit should also be supplied electronically, preferably in pdf text-readable form. Any relevant drawings or figures should be supplied electronically also in pdf form. This should become a routine part of preparing trial bundles.