Tate & Lyle Technology Ltd v Freres

Problems of obesity and diabetes are high on the international agenda in the developed world. In consequence the market for low calorie foods is one of the fastest growing sectors of the food processing industry. Thus the industry requires sugar substitutes. This case concerns a sugar substitute called maltitol. Until the 1980s maltitol was only available in commercial quantities in the form of a syrup. In the 1980s various techniques were developed for the industrial production of anhydrous maltitol crystals from the maltitol syrup. The patent in suit concerns the crystallisation of maltitol, and in particular two different shapes of maltitol crystal, described in the patent in suit as “bipyramidal” and “prismatic”. What it claims is that a by-product of the maltitol production process, called maltotriitol, can be used to modify or control the shape that the crystals take.

As granted in the EPO, the patent in suit had 10 claims. However it was opposed in 2003. As a result of the opposition the Opposition Division of the EPO decided that nine of the 10 claims were invalid on the ground that they were not new. The remaining claim (formerly claim 10) was allowed to remain, but in a modified form. Tate & Lyle (which did not participate in the opposition in the EPO) say that the remaining claim is also invalid, with the result that the patent should be revoked.

Tate & Lyle’s case was argued by Mr Colin Birss QC and Mr Tom Mitcheson. The case for the patentee, Roquette Frères, was argued by Mr Iain Purvis QC and Ms Anna Edwards-Stuart.

Sugars are saccharides. Saccharides come in many different forms. Some are monosaccharides, some disaccharides, others trisaccharides and so on. The prefix denotes the degree of polymerisation (i.e. the number of small molecules bonded together) and is abbreviated to DP in some of the prior art. Thus a trisaccharide can be abbreviated to DP3. Saccharides can also be divided into oligosaccharides and polysaccharides. The former have more than one identical molecule bonded together but not very many. If they have many they are polysaccharides.

The process of making maltitol crystals begins with a starch (typically from corn, potatoes or wheat). The starch is first liquefied by the addition of enzymes or acids. Once liquefied, any enzyme is deactivated and saccharification can take place. Saccharification is also carried out by means of enzymes which break down the long polysaccharide chains into shorter ones. Specific enzymes are added to hydrolyse the polysaccharides so as to yield the required type of sugar. On the way to making maltitol, it is necessary to yield maltose, so the additive will be a maltogenic enzyme. Maltose itself consists of two units of glucose linked together and is thus a disaccharide. But the maltogenic enzyme will also produce maltotriose, which consists of three units of glucose linked together and is thus a trisaccharide. There are other by-products of enzymatic maltose production (e.g. panose and maltotetraose). The hydrolysation results in a hydrolysate. This is filtered to remove insoluble material. It is also typically purified by column chromatography or molecular sieving to remove unwanted sugars.

The purified hydrolysate is then hydrogenated. This converts the sugars into sugar alcohols in the form of a syrup. Hydrogenation of the maltose in the hydrolysate will result in maltitol within the syrup. Hydrogenation of the maltotriose in the hydrolysate will result in maltotriitol within the syrup. The hydrogenated equivalents of other by-products of maltose production are sorbitol, panitol and maltotetraitol. Since the object of the exercise is the production of maltitol crystals, the presence of maltotriitol and other by-products is regarded as an impurity.

The syrup is then crystallised. Even if they are not polymorphic, crystals can grow in different ways, and exhibit different shapes. The different shapes are properly called “habits”. However, the patent in suit describes the different shapes as different “forms”; which is an alternative descriptor. These different habits form as a result of the symmetry of the internal crystal structure and the relative growth rates of the crystal faces. Thus if molecules of maltitol accumulate at one face more rapidly than at another, long needle-like crystals are formed. If, on the other hand, the crystals accumulate at the same rate at the competing faces, the resulting crystals are likely to be more cubic in nature. There is no accepted nomenclature for the different habits that crystals take. Different adjectives are used by different people to describe what is essentially the same habit. Thus Roquette have described the same habit as “bipyramidal”, “arrondi” (round) and “pave” (like a cobblestone). Others have described the same habit as “parallelepipedal”.

Professor Davey, called by Tate & Lyle, is a Professor of Molecular Engineering at the University of Manchester. His particular research areas are crystal nucleation and growth. He has experience of working in industry. Between 1977 and 1995 he worked for ICI dealing with crystallisation of a wide range of products in which capacity he supervised a team of chemists at PhD level or above and technicians. Together with Professor John Garside he taught an annual course on crystallisation, aimed at those working in the industry. The lectures were subsequently worked up into a book published in 2000, although the lectures themselves go back many years beforehand.

Dr Lindley, called by Roquette, is a research chemist and consultant in the area of food technology and particularly in the development and use of sweeteners. His particular area of expertise is the taste and sweetness of saccharides.

The patent in suit is addressed to a team of people interested in the commercial production of maltitol. Dr Lindley describes the team as likely to be headed by a chemical engineer with knowledge of carbohydrate chemistry and the different physical properties of sugars and sweeteners. The team would also include someone with knowledge of industrial crystallisation. They would know that different crystal habits exhibit different performance characteristics. The team would also know about the manufacturing process (including the operation of the relevant industrial equipment). Professor Davey agrees that the team will include someone with knowledge of crystallisation from solution.

The experts also agreed that the team would have access to relevant textbooks. Dr Lindley nominated Sugar Chemistry and Advanced Sugar Chemistry, while Professor Davey nominated Crystallization by John Mullin (in its third edition at the priority date), which he described as the Bible and the standard text at undergraduate and post-graduate level. Dr Lindley was not in a position to disagree.

Although the experts agree that the patent in suit is addressed to a team, neither of them had the skill set of the whole team. Thus while Professor Davey was very knowledgeable about crystallisation, he was less knowledgeable about saccharides than Dr Lindley and also less knowledgeable about the practical production of maltitol. On the other hand, Dr Lindley was not an expert in crystallisation; and he, too, disclaimed expert knowledge of the production process. Both the experts gave their evidence fairly and were doing their best to assist the court.

It was common ground that the common general knowledge included methods of making maltitol. The paradigm method was to start with starch, saccharify it and then hydrogenate it. The resulting liquid would be likely to contain not only maltitol but also sorbitol and polyols. The latter would include maltotriitol. I have already described this process.

The experts also agree that it was common general knowledge that a number of parameters could influence the rate and extent of crystallisation and the habit of the crystals formed. These include:

The experts thus agree that it was common general knowledge at the priority date that the presence of impurities could affect the habit of the crystals formed. Not all impurities did this, but some did. Others slowed down the crystallisation but without altering the habit. Yet other impurities had no relevant effect. A habit altering impurity blocks the addition of molecules at one face in preference to another, thus influencing the habit of the crystal eventually formed. The best known example of this in the field of sugar chemistry at the priority date was the interference of raffinose in sucrose crystallisation. In some cases, however, an impurity was added to a solution intended for crystallisation. This is explained in Crystallization by John Mullin:

“The majority of habit modification cases reported in the literature have been concerned with laboratory investigations, but the phenomenon is of the utmost importance in industrial crystallization and by no means a mere laboratory curiosity. Certain crystal habits are disliked in commercial crystals because they give the crystalline mass a poor appearance; others make the product prone to caking (section 7.6), induce poor flow characteristics or give rise to difficulties in the handling or packaging of the material. For most commercial purposes a granular or prismatic habit is usually desired, but there are a few specific occasions when other morphologies, such as plates or needles, may be wanted. A few examples of habit modification of industrial interest are listed in Table 6.2. In nearly every industrial crystallization some form of habit modification procedure is necessary to control the type of crystal produced. This may be done by controlling the rate of crystallization, e.g. the rate of cooling or evaporation, the degree of supersaturation or the temperature, by choosing a particular solvent, adjusting the solution pH, deliberately adding an impurity that acts as a habit modifier, or even removing some impurity that already exists in the solution.”

Table 6.2 includes raffinose and sucrose. Professor Davey explains that adding raffinose to a solution of sucrose produces needles instead of the usual prisms. Dr Lindley agreed; but he said that raffinose was a very undesirable impurity which was to be avoided or eliminated if possible.

Professor Davey goes on to say that before 1997 it was well known to anyone skilled in the art of crystallisation that an impurity with a similar molecular structure to the substance to be crystallised could modify crystal growth. It was known to be critical that part of the impurity mimics the structure of the host so it can enter the growing surface, and that the remainder of the impurity must be different so as to disrupt crystal growth of the face that it binds to. Dr Lindley did not, I think, take issue with this. Both experts agreed that it was well known that the crystal habit of a molecule may well influence its properties, such as its hygroscopicity, flow properties and solubility.

Dr Lindley agreed that it was well known that the syrup to be crystallised would be likely to contain sorbitol and higher polyols including maltotriitol. He also agreed that the team would have access to an analytic department which would routinely measure the components of the syrup (including both maltitol and the impurities). In the production of crystals the other parameters would also be routinely monitored and controlled. As long as the process resulted in crystals of satisfactory quality the parameters would remain unchanged. As part of the production process the habit of the crystals would also be routinely examined and monitored; and if the habit of the crystals changed that would soon be brought to the attention of the relevant personnel. If the habit of the crystals changed that would show that one of the parameters had changed. It was thus common ground that part of the production process entailed examining the habit of the crystals that it produced.

The habit of the crystals can be important, especially if the crystals are to be subjected to further processing. For instance, if they are plate-shaped or needle-shaped, they may tend to block filters and other processing equipment; whereas if they are cuboid they may be freer-flowing and hence more desirable.

The patent in suit has a priority date of 1997. Its original language was French. The specification begins by introducing maltitol, and explains that it was first presented in syrup form, then in powder form, and that it has been known in crystalline form since 1980. These crystals are anhydrous maltitol which are subject to wide-ranging patent protection on behalf of the HAYASHIBARA company. The specification then describes two forms of processes for crystallisation: the massé technique and crystallisation in water. These are almost the only processes in industrial use. The products obtained by these methods are of very variable crystallinity and not all are particularly well suited to certain applications such as chewing gum or chocolate. It also gives other examples of situations in which crystals produced by these techniques are not satisfactory. The particular defects to which it draws attention are those of not flowing easily, being liable to cake or knot, poor solubility, being bad vehicles for compression and not meeting the criteria for identification and purity set by the different pharmacopoeias [0008].

The problem that Roquette set out to solve is described in [0009]. It sought “to perfect maltitol compositions which do not have the flow, caking, dissolving or compression defects presented by known maltitol powders.” It continues by asserting [0010] that while working to perfect these compositions it “was able to isolate, in a surprising and unexpected way, two particular forms of maltitol crystals, one bipyramidal and the other prismatic.”

The key teaching of the patent is set out in [0011]. It says:

“It has indeed demonstrated that, against all expectation, the form of the maltitol crystals was a function of the maltotriitol content of a maltitol syrup intended for crystallisation. The Applicant company has noted that by controlling the maltotriitol content of a maltitol syrup, it was possible to direct the form of the maltitol crystals towards one or other of the forms or towards a mix of the two forms, when this maltitol syrup is subjected to a crystallisation stage.”

The bipyramidal crystal is described as made up of two regular tetrahedrons of base 50-500µm, to make an octahedron with edge length of 50-500µm. The experts agree that the octahedron is in fact made up of two square pyramids rather than two tetrahedrons. The prismatic crystal is described as consisting of plane faces constituting a tetrahedron and 100-400µm long by 20-100µm wide.

The patent goes on to explain the merits and demerits of the two different habits: [0014] to [0016]. A semi-crystallised mass of maltitol with bipyramidal crystals is more viscous than one with prismatic crystals. The bipyramidal crystal provides a more thickened mass before refining which is useful in the production of chocolate, and a greater consistency of flow which is useful in the production of pharmaceutical coatings etc. On the other hand the prismatic crystal is more compressible and so enables low-crystal content caking in the manufacture of chewing gum centres and the like.

The patent continues with a description of the two crystal habits. At [0027] to [0030] the patent explains that differences in habit are caused by different quantities of maltotriitol. Less than 1% maltotriitol and the crystals are bipyramidal in habit. More than 4% maltotriitol and the crystals are prismatic in habit. Between 1% and 4% and the crystals are a heterogeneous mixture of the two habits. These data are subsequently tabulated at [0077] (although one figure in the table is in the wrong place). At [0034] the patent repeats its key teaching:

“One of the essential characteristics of the invention is therefore to vary the maltotriitol contents of the maltitol syrups to be crystallised while advantageously retaining a reduced content of maltosyl 1,6 matitol.”

However, although this paragraph teaches that the maltotriitol content of the maltitol syrup can be varied, it also proposes a number of other ways in which the maltotriitol level can be controlled. The patent goes on to explain that the maltotriitol content of the syrup may be controlled either upstream or downstream. It proposes three different ways of controlling the maltotriitol content upstream (all of which result in the ultimate reduction of maltotriitol content) and two different ways of controlling the maltotriitol content downstream (one of which reduces and the other of which increases the maltotriitol content). The upstream methods of controlling the maltotriitol content of the syrup include alteration of the hydrolysate before hydrogenation (e.g. by molecular sieving). At this stage, there is no maltotriitol present in the hydrolysate, because that comes into existence only after hydrogenation. What is present in the hydrolysate at that stage is its forerunner: maltotriose. The patent explains that all these possibilities for controlling the maltotriitol content may be used singly or in combination: [0038].

The patent continues with a long (and largely irrelevant) description of the process by which the syrup is obtained. However, the patent explains ([0047]) that at the first stage of saccharification:

“it is already possible to control the maltotriose content (which after hydrogenation leads to maltotriitol) … by adjusting the amount of maltogenic α-amylase as a function of the maltotriose content and therefore of the form of the maltitol crystals that it is wished to obtain.”

The patent thus teaches that one way of controlling the crystal habit is by adjusting the maltotriose content of the hydrolysate with the result that the amount of maltotriitol content of the syrup after hydrogenation is likewise controlled. This is reflected in [0071] and [0072] of the specification, which explain:

“[0071] According to a first version of the process, a series of steps is applied to the maltitol syrup obtained in the preceding hydrogenation stage as follows

-

possibly carrying out a chromatographic fractionation, known per se, so as to obtain a maltitol rich fraction and a more or less rich maltotriitol fraction as a function of the form of crystals required

-

concentrating the maltitol rich fraction

-

crystallising and separating the formed maltitol crystals

-

recycling the crystallisation mother liquors upstream of the chromatographic fractionation stage

[0072] According to a second version of the process, a series of steps is applied to the maltitol syrup obtained in the preceding hydrogenation stage as follows

- concentrating the maltitol syrup

- crystallising and separating the formed maltitol crystals.”

The first version adjusts the maltotriitol content of the syrup after hydrogenation, but the second version does not. The maltotriitol content has already been determined by the adjustment of the maltotriose content of the hydrolysate; and thus after hydrogenation the syrup merely requires concentration and crystallisation. A similar process is described in [0074] where the maltotriose content of the hydrolysate is again adjusted before hydrogenation.

The patent gives one example to illustrate the claim [0076]. In the example the syrup to be crystallised is concentrated and placed in a dish. Crystallised maltitol is then added, and the dish is cooled. The crystals are thus formed. There is no adjustment to the maltotriitol content of the syrup. The tabular results of the example simply record that “different bases are used”.

As one might expect, the patent culminates in the claims. It is necessary to describe some of the claims that were subsequently abandoned or disallowed by the Opposition Division of the EPO. Claims 1 and 4 as granted claimed the two types of crystal: bipyramidal and prismatic respectively. These claims have been abandoned, because neither of these crystal habits was new. Each had been disclosed by the prior art. Granted claims 2 and 5 claimed compositions comprising those crystals respectively and claim 7 was to a composition comprising both kinds. These claims have also been abandoned for similar reasons. The granted patent also contained claims to processes for making the bipyramidal and prismatic crystals (and mixtures) in claims 3, 6 and 8. Those claims mentioned the maltotriitol content as appropriate (lower than 1% for bipyramidal, more than 4% for prismatic and in between for a mixture). These process claims have also been abandoned because the processes for making crystals using the specified maltotriitol content had also been anticipated by prior art.

Granted claim 9 was:

“A process for directing the form of the maltitol crystals, characterised in that it consists in controlling the maltotriitol content of the maltitol syrup to be crystallised.”

Roquette defended this claim before the Opposition Division. But the Opposition Division decided that this claim, too, had been anticipated. Roquette mounted and subsequently abandoned an appeal. So this claim has been revoked. That leaves granted claim 10 as the only remaining claim. It claims:

“The use of maltotriitol to modify or control the form of maltitol crystals.”

The key teaching of the patent is that is that it is the maltotriitol level which is the determinant of crystal habit. Mr Purvis did suggest that the level of maltotriitol might not be the only determinant of crystal habit and that other agreed parameters (e.g. temperature and rate of cooling) might play a part. He suggested that the patent in suit recognised this by explaining ([0033]) that it was “likely” (rather than certain) that particular crystal habits would be obtained at given levels of maltotriitol. The difficulty with this suggestion is twofold. First, neither the specification nor the claim refer to any other determinant. Second, if other determinants are relevant to crystal habit, the patent does not explain how and under what conditions they operate or what effect they have, with the result that the patent must be insufficient. It is, in my judgment, plain that the patent teaches that the maltotriitol level is the only determinant. No other parameter (such as temperature) is said to have that influence either alone or in combination with the maltotriitol level. The claim must be understood in that context.

The patent also teaches that the maltotriitol content of the syrup may be controlled directly or indirectly (in the latter case by controlling the level of maltotriose before hydrogenation).

In his opening skeleton argument Mr Purvis submitted that the better construction of the remaining claim is that it covers any process of making maltitol in which maltotriitol content is actively adjusted in the course of production so as to obtain maltitol of a particular crystal habit. In practice this requires that the maltotriitol content is monitored in the course of the manufacturing process; and is adjusted as and when necessary to change the crystal habit being created, for example from bipyramidal to prismatic or the other way around. Accordingly, on this construction the key requirements of the claim are:

He amplified this submission in closing as follows. The claim requires that maltotriitol is “used” to modify or control the form of maltitol crystals; not that the maltotriitol is modified or controlled. There are three sections to the claim that need to be construed. First what is meant by “use”; second what is meant by “modify or control”; third what is meant by “form”. In the context of the specification the word “use” denotes some active adjustment of maltotriitol content in the course of production, by adding or removing maltotriitol. This is consistent with the teaching of the essential characteristics of the invention at [0034], and the subsequent teaching of how to implement the invention in [0035] to [0039]. “Modify” and “control” mean essentially the same thing. They require active alteration of the crystal habit. “Control” is used in the sense of controlling water temperature by adjusting a mixer tap. It is common ground that “form” means habit. Thus Mr Purvis submits that requirement that the habit of the maltitol has to be changed by the adjustment of maltotriitol content means that one would test for infringement by comparing the crystal habit actually obtained from what would have been obtained in the absence of the adjustment. If the crystal habit is different, then there is infringement. If not, there is no infringement. Accordingly if one starts with a syrup already containing the required amount of maltotriitol and has no further need to adjust it the patent is not infringed.

Mr Birss, on the other hand, says that this construction is too restrictive. The claim encompasses not only a case in which the amount of maltotriitol is changed (either by increase or reduction) but also a case in which the amount of maltotriitol is checked, found to be satisfactory and left alone. It would cover a case in which the process prevented the formation of maltotriitol by reducing the amount of maltotriose in the hydrolysate before hydrogenation. It would also extend to a case in which someone used the discovery that the level of maltotriitol determined the crystal habit by buying pure maltitol crystals and recrystallising them. Such a person would know that the purity of the maltitol was such that he would always produce bipyramidal crystals. He would thus be using the maltotriitol content (or lack of it) to control the habit of the crystals. Since it was common ground that the claim did not include any subjective element, it would not matter whether the person in question knew that the reason why the crystal habit was what it was was the level of maltotriitol in the syrup.

The claim as framed simply describes the “use” of maltotriitol, without saying how it is used. This, to my mind, gives support to Mr Birss’ construction. Contrary to Mr Purvis’ submission the verb “control” seems to me to mean something other than “modify” otherwise the patentee would not have used two verbs. Modify plainly means “to change”. It seems to me, therefore, that “control” must also encompass something falling short of an actual change in the maltotriitol content or indeed in the form (i.e. habit) of the crystals. Mr Purvis’ construction does not give full weight to this verb, whereas in my judgment Mr Birss’ construction does. Moreover, the analogy of the water tap does not support Mr Purvis’ construction. In many systems it is possible to pre-set the required temperature by a thermostatic valve, so that all the user has to do is to turn the tap on or off. Likewise if I have a car with cruise control it keeps me at the same speed. The meaning of “control” in my judgment encompasses keeping things the same. No active adjustment is required. In addition the French text uses the verb “contrôler”, among whose meanings are “to monitor, inspect or check”. The contrôle de billets on the SNCF simply checks passengers’ tickets. This, too, to my mind supports Mr Birss’ construction. In my judgment, therefore, the claim encompasses a case in which the maltotriitol content is monitored, inspected or checked, without necessarily being changed.

In addition Mr Purvis’ construction imports a requirement of the active adjustment of maltotriitol itself. Maltotriitol is only formed after hydrogenation. But as the patent explains, the composition of the syrup may be adjusted before hydrogenation, by controlling the level of maltotriose. Mr Purvis’ response to this point was that two of the four versions of the process described respectively in [0072] and [0074], in which the contents of the syrup was adjusted before hydrogenation, were simply not within the claim at all. I do not consider that the skilled person reading the patent would treat two of the four versions of the process as being simply irrelevant. He would understand them as teaching him something about the intended compass of the claims. In addition, if Mr Purvis’ construction is right then the single example given in the patent (and its tabulated results) would not infringe the claim, because the example does not disclose any active adjustment of the maltitol content of the syrup. No skilled reader would understand that the sole example given in the specification is irrelevant to the claim. The thrust of the example is that it is simply the content of the maltotriitol in the syrup that determines the habit of the resulting crystals. This, too, supports Mr Birss’ construction.

I reject Mr Purvis’ construction of the claim, and accept Mr Birss’ submission that the patent will be infringed by any process which, when reproduced, will consistently result in bipyramidal crystals, or prismatic crystals, or a mixture of the two.

Tate & Lyle attack the validity of the last remaining claim on three bases:

Tate & Lyle cited a plethora of prior art: nine citations in all. It is not necessary to refer to all of them in detail.

This patent dates from 1983. It was the first to show the world how to crystallise maltitol, which until then had been produced in syrup form. Hirao describes the creation of a seed crystal from which anhydrous crystals of maltitol can be obtained. It is common ground that, according to the photographs forming part of the specification, the crystals produced by the method described in Hirao have two habits. The syrup used to make the crystals (after hydrogenation) consisted of 0.8% sorbitol, 92.2% maltitol, 4.6% maltotriitol and 2.4% higher sugar alcohols. According to the teaching of the patent in suit syrup with a maltotriitol level exceeding 4% ought to have resulted in prismatic crystals only.

Boursier dates from 1989. It describes a method of covering a confectionery or pharmaceutical product with a hard (i.e. crystalline) sugarless coating. The coating is made from a maltitol syrup. An example that has proved satisfying contains 97.1% maltitol by weight; 1.1% sorbitol and 1.8% maltotriitol. However, Boursier also gives data about comparative experiments. The experiments begin with a maltitol syrup containing 95% maltitol, 1.6% sorbitol and 4.3% maltotriitol. These were crystallised at different temperatures, but neither was satisfactory. The next series of comparative experiments used different concentrations of maltitol in the syrup. A table in the specification (Table III) gives comparative results for the syrups with different maltitol contents: 92%, 95%, 96% and 97% respectively. The syrups with 96% or 97% maltitol gave good results (97% being better than 96%) whereas those at lower concentrations gave poorer results.

Devos also dates from 1989. It describes a continuous process for the preparation of crystalline maltitol. The patent is owned by Roquette and it is the process that they use to manufacture their commercial product Maltisorb. Devos points out that it is important that the syrup should contain very little impurities of the polyalcohol type of a glucidic nature; and explains that the reason for this is that they interfere with or even prevent the crystallisation of maltitol. The process described by Devos has a number of steps. The starting point is starch which is then saccharified. The saccharified syrup is then hydrogenated. This results in a hydrogenated syrup which is then subjected to chromatography. The chromatography fractionates the syrup into a number of fractions. One of the fractions is rich in maltitol and another is rich in maltotriitol. There are other fractions rich in sorbitol and higher oligosaccharides, but these do not matter for present purposes. What matters is that the maltitol rich fraction is taken forward for crystallisation and the maltotriitol rich fraction is discarded. The maltitol rich fraction consists of at least 87% maltitol, small proportions of sorbitol and other polyols and a content of maltotriitol of between 2.5% and 13% by weight. When this syrup is crystallised it results in anhydrous crystals which Devos describes as “flowing quite freely”. The mother liquor used in the crystallisation is then recycled for further use in crystallisation in the next batch of syrup. Thus at the stage of chromatographic fractionation (after hydrogenation) maltotriitol is removed from the syrup; and the recycling of the mother liquor after crystallisation will add maltotriitol to the next batch of syrup.

Niimi dates from 1995. It describes a method for manufacturing high purity maltitol. It points out that a contamination of sugar alcohol of DP 3 or more (i.e. by trisaccharides or higher) checks the crystallisation of maltitol leading to a longer period of crystallisation and a lower yield of maltitol. The invention consists of simpler methods of decreasing the amount of oligosaccharides of DP 3 or more (which would include maltotriitol). The way in which it does this is primarily by selecting enzymes in the saccharification process which will reduce the amount of maltotriose in the hydrolysate before hydrogenation.

Magara dates from November 1996. It describes a process for manufacturing crystalline maltitol. The starting point is a syrup containing 81% to 90% maltitol. The syrup is then subjected to chromatographic separation which results in an aqueous solution of maltitol with a maltitol purity of between 94% to 99.9%. This solution is then crystallised. Magara does not specify the precise amount of maltotriitol in any of the syrups: it simply describes DP 3 saccharides (of which maltotriitol is one). However, within the given ranges Dr Lindley accepted that the content of maltotriitol would range from less than 1% to more than 4%. Magara gives a number of examples of the process at work. In example 3 a hydrogenated aqueous solution of maltitol contains 84.4% maltitol, 8.0% sorbitol and 7.6% sugar alcohol of DP 3 or more (which would include maltotriitol). That solution is then passed through a chromatography column in order to recover a fraction containing a higher proportion of maltitol: in this case 98.2% maltitol, 1.2% sorbitol and 0.6% sugar alcohol (which would include maltotriitol). This syrup is then crystallised.

Ribadeau-Dumas dates from December 1996. It describes a grainy confectionery product and a process for manufacturing it. Ribadeau-Dumas explains that there are two habits of maltitol crystal. It describes one as having “a needle shape”. When the crystals are this shape, with a length of 120 microns and a cross section of 35 microns, the confectionery product tends to change rapidly and harden excessively. The other is described as having a “parallelepipedal shape” with a size of less than 60 microns. This shape results in confectionery that is stable for several months. The needle shape corresponds to the prismatic crystals described in the patent in suit; and the parallelepipedal shape corresponds to the bipyramidal crystals described in the patent in suit. In the case of the crystals of parallelepipedal shape Ribadeau-Dumas points out that the crystals have a high purity with respect to Maltitol, exceeding 95% and most often 98% or even 99%. The specification goes on to say:

“In order to control the crystallization of maltitol … and to obtain sufficiently fine crystals, it will be preferable to adjust the oligosaccharide and polysaccharide contents in the grainy confectionery product in accordance with the invention.”

Example 1 describes the preparation of maltitol fondants. The syrup that is crystallised contains approximately 92% maltitol, 1% sorbitol and 4.6% maltotriitol. Maltitol powder is then added to the syrup. The maltitol powder has been made according to the teaching of Hirao. The syrup is then crystallised and on examination the crystals are “essentially … large parallelepipedal crystals, rather cubic, with edges of about 100 microns for the smallest crystals”. Further trials were made using different seed crystals in crystallising the syrup. Two were needle shaped and two were parallelepipedal. Following crystallisation and cooking the fondants made with the needle-shaped seed crystals were “particularly hard, brittle and sandy in the mouth”. By contrast the fondants made with the parallelepipedal seed crystals were firm but had an ideal texture.

Caboche dates from 1997. It describes a novel crystalline maltitol composition which is porous and honeycombed; and a process for manufacturing it. Caboche teaches that the composition should contain a low level of polyols (including maltotriitol). The polyol content should be preferably less than 5% and, better still, less than 2% because “their presence significantly impairs the crystallinity of the composition according to the invention”. Example 1 describes the crystallisation of a syrup containing 99.8% maltitol. The process used is atomisation.

The claim will have been anticipated if someone following the instructions given by a piece of prior art would inevitably infringe the patent. Lord Hoffmann summarised it thus in Synthon BV v Smith Kline Beecham plc [2006] R.P.C. 10 (§§ 22, 23):

“…the matter relied upon as prior art must disclose subject-matter which, if performed, would necessarily result in an infringement of the patent. That may be because the prior art discloses the same invention. In that case there will be no question that performance of the earlier invention would infringe and usually it will be apparent to someone who is aware of both the prior art and the patent that it will do so. But patent infringement does not require that one should be aware that one is infringing: “whether or not a person is working [an] … invention is an objective fact independent of what he knows or thinks about what he is doing”: Merrell Dow Pharmaceuticals Inc v H N Norton & Co Ltd [1996] R.P.C. 76 , 90. It follows that, whether or not it would be apparent to anyone at the time, whenever subject-matter described in the prior disclosure is capable of being performed and is such that, if performed, it must result in the patent being infringed, the disclosure condition is satisfied. The flag has been planted, even though the author or maker of the prior art was not aware that he was doing so. …

…But the infringement must be not merely a possible or even likely consequence of performing the invention disclosed by the prior disclosure. It must be necessarily entailed. If there is more than one possible consequence, one cannot say that performing the disclosed invention will infringe. The flag has not been planted on the patented invention, although a person performing the invention disclosed by the prior art may carry it there by accident or (if he is aware of the patented invention) by design. Indeed, it may be obvious to do so. But the prior disclosure must be construed as it would have been understood by the skilled person at the date of the disclosure and not in the light of the subsequent patent.”

Mr Purvis pointed out that in Synthon Lord Hoffmann quoted with approval the following passage from T-396/89 UNION CARBIDE/high tear strength polymers [1992] E.P.O.R. 312:

“It may be easy, given a knowledge of a later invention, to select from the general teachings of a prior art document certain conditions, and apply them to an example in that document, so as to produce an end result having all the features of the later claim. However, success in so doing does not prove that the result was inevitable. All that it demonstrates is that, given knowledge of the later invention, the earlier teaching is capable of being adapted to give the same result. Such an adaptation cannot be used to attack the novelty of a later patent.”

He thus submitted that it was not legitimate to select particular conditions that happen to fall within the range of what the prior art discloses, in order to try to meet the requirements of the claim. This submission applies, in particular, to Devos. But it seems to me that the force of this submission must depend on how, in the real world, the skilled addressee would put the teaching of the prior art into practice. The policy underlying anticipation is that it is wrong to allow a patentee to prevent a person from continuing to do that which he was lawfully doing before the grant of the patent: Windsurfing International Inc. v Tabur Marine (Great Britain) Ltd [1985] R.P.C. 59, 77. If, therefore, the practical implementation of the prior art would necessarily involve selecting from the range, then it seems to me that Mr Purvis’ objection fails.

Mr Purvis also submitted that there was a false assumption underlying the entirety of the novelty attack. He said that it might be the case in the examples given in the patent in suit, where particular crystallisation parameters have been adopted (e.g. 50 degrees C, very slow cooling to 20 degrees over 4 days), that alteration of the content of maltotriitol would change the crystal habit. But unless you are using precisely the same crystallisation parameters, there is no reason to assume that this would be the case. And here, the prior art uses different parameters. The way to seek to prove it would be to carry out an experiment to replicate what was being taught in the prior art and see what resulted. But Tate & Lyle had not done that. The difficulty with this submission is that the patent in suit teaches quite clearly that maltotriitol is the sole determinant of habit, or at least is the sole determinant of habit at the levels stated in the specification. The novelty of the patent in suit must, in my judgment, be determined on its own terms.

Hirao instructs the reader to crystallise a syrup with a maltotriitol content of 4.6%. According to the teaching of the patent that will consistently result in prismatic crystals. Mr Birss submitted that if you were to make anhydrous maltitol crystals following the instructions in Hirao, you would infringe the claim. The reason for this is that you are told what level of maltotriitol to use. According to the teaching of the patent in suit that would determine the crystal habit. By selecting the maltotriitol content of the syrup you are using it to control the crystal habit, even if you did not know that it had that effect. I accept this submission.

I note also that the Opposition Division of the EPO decided that claim 9 of the patent as granted was anticipated by Hirao. Granted claim 9 was:

“A process for directing the form of the maltitol crystals, characterised in that it consists in controlling the maltotriitol content of the maltitol syrup to be crystallised.”

I was unable to understand (and Mr Purvis could not explain) what the difference was between claim 9 in the patent as granted and the remaining claim in the patent in suit. If the one is anticipated by Hirao, so must the other be.

I hold, therefore that the patent in suit is anticipated by Hirao.

Boursier teaches the reader to crystallise syrups with different levels of maltitol. The syrup consists of maltitol with “a certain content of sorbitol and maltotriitol”. Thus different levels of maltitol inevitably entail different levels of maltotriitol. Table III shows that the higher the level of maltitol (and hence the lower the level of maltotriitol) the better the results. The best results are obtained with a syrup containing 97% maltitol and thus, by necessary implication, less than 4% maltotriitol. Mr Birss submits that by selecting a syrup containing less than 4% maltotriitol in preference to a syrup containing more, you are using it to control the crystal habit, even if you did not know that it had that effect. I accept this submission.

Boursier is concerned with the manufacture of a hard sugarless coating for confectionery or pharmaceutical products. Are products of this kind within the claim of the patent in suit? The specification of the patent in suit is unspecific about the kinds of products to which its teaching relates; and on its face the claim applies to any process in which maltitol crystals are used. In my judgment the kind of products contemplated by Boursier are within the claims.

I hold, therefore that the patent in suit is anticipated by Boursier.

Devos manipulates the level of DP3 (which would include maltotriitol) in two ways. First, it does so at the chromatography stage when the maltitol rich fraction is separated from the maltotriitol rich fraction. Second, it does so after crystallisation when the mother liquor (also containing maltotriitol) is recycled. According to Devos the range of content of maltotriitol in the syrup runs from 2.5% to 13% by weight. Dr Lindley agreed that anyone putting Devos into practice would choose a level within that range and stick to it. The chosen level would be that which would maximise yield. Dr Lindley’s evidence continued as follows:

“Q. Therefore what I suggest, Dr. Lindley, is that if you were putting this into practice, what you would do is where it indicates here the range of maltotriitol levels in the syrup, which we looked at at column 5, lines 6 and 7, staying within the four corners of Devos, you would find out whether 13% maltitol let you run your crystalliser all right but gave you bad/good yield and whether 2.5% and few in between. That is what you would do.

A. Probably.

Q. It is not probably. It is what you would do, is it not?

A. Yes.

Q. The point then is, on the hypothesis that the patent is right, what would happen is that the habit of your crystals would change from prisms at 13% to a mixture of bipyramids and prisms at 2.5%.

A.

All things being equal, yes.”

This evidence shows, to my mind, that someone working the invention disclosed by Devos would actively manipulate the maltotriitol content of the syrup and in so doing would change the crystal habit.

It follows, in my judgment, that the patent in suit is anticipated by Devos.

Nimi teaches the control of maltotriitol at the processing stage upstream of hydrogenation, by removing maltotriose. This indirect control of maltotriitol in my judgment falls within the claim. The ranges of maltotriitol content given in the examples fall on each side of the parameters taught by the patent as changing crystal habit. Thus Example 14 describes a solution containing 4.5% trisaccharides or higher. Dr Lindley agreed that this would result in a maltotriitol content exceeding 4%. On the other hand, Example 8 described a syrup with a trisaccharide content of 0.5%. Using these two syrups would result in different crystal habits, whose habit had been determined by the amount of maltotriitol in the syrup.

In my judgment the patent in suit is anticipated by Nimi.

Example 3 in Magara starts with a hydrogenated aqueous solution of maltitol containing 84.4% maltitol, 8.0% sorbitol and 7.6% sugar alcohol of DP 3 or more (which would include maltotriitol). Dr Lindley agreed that the maltotriitol content of the solution would probably exceed 1%. That solution is then passed through a chromatography column in order to recover a fraction containing a higher proportion of maltitol: in this case 98.2% maltitol, 1.2% sorbitol and 0.6% sugar alcohol (which would include maltotriitol). The syrup necessarily contains less than 1% maltotriitol. This syrup is then crystallised.

Thus the maltotriitol content of the syrup has been deliberately adjusted. According to the teaching of the patent if the maltotriitol content had not been adjusted in this way the crystals would have been a mixture of bipyramidal and prismatic. But as a result of the adjustment they will be bipyramidal. It follows that the deliberate adjustment of the maltotriitol content has altered the crystal habit. This falls squarely within the claim.

Accordingly I hold that the patent in suit is anticipated by Magara.

Ribadeau-Dumas replicates the process of making crystals described by Hirao. I do not consider that it adds anything of significance in the present context. In so far as it gives other examples or processes, the starting point for all of them is pure maltitol crystals. Apart from the initial crystallisation using Hirao’s method, maltotriitol does not feature at all.

Caboche contrasts the crystals made by this method with crystals made by methods described in the prior art. However, the properties of the crystals described in the table on which Mr Birss relied are properties of the crystals themselves rather than the syrup from which they were crystallised. Apart from the general teaching that polyols are undesirable because they impede crystallisation, I do not consider that the method described by Caboche falls within the claim. The general teaching is not, I think, put into practice by the process described in the patent, with the result that someone carrying out the teaching of Caboche would not necessarily infringe the patent in suit.

I hold therefore that the patent in suit is not anticipated by Caboche.

I have held that the patent in suit is anticipated by many pieces of prior art. On the face of it this is surprising. But this conclusion feeds in to Mr Birss’ next attack; namely that the patent in suit claims what is essentially a discovery. Section 1 (2) of the Patents Act 1977 provides that anything which consists of “a discovery … as such” is not an invention. As Lord Hoffmann put it in Kirin-Amgen Inc v Hoechst Marion Roussel Ltd [2005] R.P.C. 9 (§ 77):

“An invention is a practical product or process, not information about the natural world.”

Mr Birss submits that claim 10 simply informs the reader that maltotriitol, a known impurity in the manufacture of maltitol, can change the habit of maltitol crystals. This is simply a property of maltotriitol which Roquette have discovered. That is why so many known processes, if carried out, would infringe the patent. Roquette have explained why maltitol crystals take the habit that they do, but have not added anything else to the sum of human knowledge. The claim is not saved from unpatentability simply by the addition of the phrase “the use of”. What matters is the substance of the claim rather than its form. It would have been possible to claim particular processes or products that took advantage of the discovery, for instance by claiming certain levels of concentration of maltotriitol within the syrup, or crystals produced with the aid of the discovery. Indeed that is what Roquette did, but all their claimed products and processes have been declared invalid (or have been abandoned).

I accept this submission. In Molière's play Le Bourgeois Gentilhomme, Monsieur Jourdain asks something to be written in neither verse nor prose. A philosophy master says to him, “Sir, there is no other way to express oneself than with prose or verse”. Jourdain replies, “By my faith! For more than forty years I have been speaking prose without knowing anything about it, and I am much obliged to you for having taught me that.” That is this case. The industry has been using maltotriitol to control or determine crystal habit without knowing it. What is left of the patent as granted is no more than a discovery as such.

For good measure Mr Birss submits that the patent in suit is invalid for obviousness over the cited prior art. In view of my conclusions so far I can express my view on the question of obviousness briefly. It is common ground that the court should approach the question whether something was obvious by following the structured approach commended in Pozzoli Spa v BDMO SA [2007] FSR 37:

I have already identified the notional person skilled in the art and the common general knowledge.

I have already construed the claim. I pause only to note that Mr Purvis characterised the inventive concept as “a process which modifies the crystal habit of maltitol by adjusting maltotriitol content” (emphasis added). I have rejected this construction of the claim; and consequently I also reject this description of the inventive concept. The inventive concept is to use maltotriitol to determine the habit of the maltitol crystals. The use may take the form of adding or removing maltotriitol to or from the syrup or selecting a syrup with a particular maltotriitol content, or removing maltotriose from the hydrolysate before hydrogenation.

None of the cited prior art focuses on the significance of maltotriitol as a determinant of crystal habit, and that knowledge was not part of common general knowledge. This is the difference between the inventive concept and the prior art.

I remind myself that, as Mr Purvis submitted:

The House of Lords has recently considered the notion of something being “obvious to try” in Conor Medsystems Inc v Angiotech Pharmaceuticals Inc [2008] RPC 28. Lord Hoffmann said (§ 42):

“In the Court of Appeal, Jacob LJ dealt comprehensively with the question of when an invention could be considered obvious on the ground that it was obvious to try. He correctly summarised the authorities, starting with the judgment of Diplock LJ in Johns-Manville Corporation's Patent [1967] RPC 479, by saying that the notion of something being obvious to try was useful only in a case in which there was a fair expectation of success. How much of an expectation would be needed depended upon the particular facts of the case. As Kitchin J said in Generics (UK) Ltd v H Lundbeck A/S [2007] RPC 32, para 72:

“The question of obviousness must be considered on the facts of each case. The court must consider the weight to be attached to any particular factor in the light of all the relevant circumstances. These may include such matters as the motive to find a solution to the problem the patent addresses, the number and extent of the possible avenues of research, the effort involved in pursuing them and the expectation of success.””

My conclusions on the question of obviousness, having regard to the factors identified by Kitchin J, are:

So far as Professor Davey’s approach to the cited prior art was concerned:

The case put to Dr Lindley was a classic step-by-step approach, often embellished by hypothetical problems that had not arisen.

In short I was not persuaded that the patent in suit, if otherwise valid, was invalid for obviousness either over common general knowledge or over any of the individual citations.

However, for the reasons I have given, I conclude that the patent is invalid because it is not new; and also because it claims an unpatentable discovery. The patent must therefore be revoked.