Court of Appeal · Decided 23 February 2010

Mühlbauer AG v Manufacturing Integration Technology Ltd

Feedback & support

[2010] SGCA 6Case no. Civil Appeal No 34 of 2009

What happened, in plain English

There is no plain-English summary of this judgment yet.

Ask about this judgment

Laws this judgment relies on

Patents Act 1994

Grounds of decision

Official text from eLitigation

Andrew Phang Boon Leong JA (delivering the judgment of the court):
[1] This is an appeal against the decision of the trial judge (“the Judge”) in Mühlbauer AG v Manufacturing Integration Technology Ltd [2009] SGHC 45 (“the GD”). The Judge allowed the respondent’s counterclaim that the appellant’s patent be declared invalid for want of both novelty as well as an inventive step. Accordingly, the appellant’s claim for infringement was dismissed.
[2] The appellant, Mühlbauer AG (“the Appellant”), is a company incorporated in Germany. It is the registered proprietor of Singapore patent No 117982 entitled “Device for Inspecting and Rotating Electronic Components” [emphasis added] (“the Patent”), filed in Singapore sometime in December 2005. The Patent is in force in Singapore, with its priority date being 16 February 2004.
[3] The respondent, Manufacturing Integration Technology Ltd (“the Respondent”), is a Singapore company. It is alleged to have infringed the Patent pursuant to s 66(1)(a) of the Patents Act (Cap 221, 2005 Rev Ed) (“the Act”) by making, disposing of, offering to dispose of, using or importing the patented product and/or keeping the patented product, whether for disposal or otherwise. In particular, the Appellant complained that the Respondent had manufactured and marketed a device under the trademark or name “CAERUS” which infringed the Patent.
[4] In the court below, the Respondent acknowledged that its machine infringed all 10 claims of the Patent. In its counterclaim, however, the Respondent contended that the Patent lacked novelty as well as any inventive step vis-à-vis the prior state of the art (“the state of the art”), and should therefore be found to be invalid. In any case, for the purposes of this judgment, we shall be taking the Patent and the Respondent’s “CAERUS” device to be exactly identical.
[5] Although the issues canvassed in the present appeal were straightforward, a number of significant (and broader) issues were raised, including (as we shall see) the role of characterisation in the context of deciding patent disputes as well as the (no less important) role of experts in the same context.
Facts (including a description of the Patent)
[6] The Patent describes a machine for inspecting, picking up and placing electronic components onto printed circuit boards or tape and reel packaging. The Respondent’s “CAERUS” device is also a machine that facilitates the inspection and rotation of electronic components.
[7] The 10 claims countenanced by the Patent are important, and they are as follows (see also Appendix A):
[8] Put simply, the Patent describes both a device (Claims 1 through 6, with Claims 2 to 6 being dependent claims) and a method (Claims 7 through 10, with Claims 8 to 10 being dependent claims) in which an optical inspection of a wafer chip, a pickup of a wafer chip, a turning-around and subsequent deposit of the wafer chip all occur within a single 180-degree rotation of a two-headed pivoting part of the device. At the end of the 180-degree rotation, the device is primed to pick up the next wafer chip in line. The device alternates between clockwise and anti-clockwise 180-degree rotations of the pivoting pickup heads. The apparent ingenuity of this device (and method) lies in its circumventing the need for separate stages for optical inspection and pickup. By means of a through opening located transversely between the two pickup heads, the top-mounted camera inspects the wafer chip below on the substrate – the base material on which semiconductor devices are fabricated – while the pivoting part, picking up another wafer chip simultaneously, rotates through an angle of 90 or 270 degrees. As optical inspection and pickup are concurrent, the cycle time is reduced, resulting in greater productivity.
[9] It was previously thought that since both the camera and the pickup head had to be located directly over the electronic component to carry out their respective tasks (of determining the position/orientation of the component and picking it up after correcting for any offset, respectively), this method required the pickup head to move out of the line of sight of the camera before each and every optical inspection. This was time-consuming. The device and method countenanced in the Patent, however, overcame this difficulty with apparent efficiency as well as elegance. Indeed, as we shall see, it is crucial to the present appeal that:
(a) The Patent operates with two pickup heads – and two pickup heads only.
(b) The Patent alternates between clockwise and anti-clockwise 180-degree rotations of the pivoting pickup heads, thus enabling a through opening located transversely between the two pickup heads that (in turn) facilitates visual inspection of the electronic component below.
It is also important to note that the combined result of (a) and (b) above – from the important practical perspective – is a much improved overall throughput. This has, in turn, translated to significant commercial success as well – a point that was acknowledged by the Judge (see the GD at [44]), although he correctly pointed out that the commercial success of an invention “alone is not conclusive” as to whether the invention is novel and/or one involving an inventive step (see the GD at [47]). But we are running ahead of the story as these (as well as other) characteristics relating to the Patent will need to be compared with the relevant characteristics in the other patents mentioned below (see [11], and which are relied upon by the Respondent as forming part of the state of the art) in order to ascertain whether or not the Patent is new and/or involved an inventive step.
The decision below
[10] After considering the evidence on both sides, in particular, the testimony of two expert witnesses, the Judge found (at [48] of the GD) that:
He further commented (at [45] of the GD) that:
[11] Chief amongst the Judge’s considerations was the fact that the Patent appeared to have been anticipated by four prior patents and a machine, as follows:
(a) Advanced Systems Automation Limited’s (“ASA”) Singapore patent No 104292 dated 30 September 2005, and which relates to a flip chip die bonder (“the ASA Patent”);
(b) Matsushita Electric Industrial Co’s United States patent No 5,839,187 dated 24 November 1998 (“the Matsushita Patent”);
(c) National Semiconductor Corporation’s United States patent No 6,364,089 B1 dated 2 April 2002 (“the National Semiconductor Patent”);
(d) Kabushiki Kaisha Shinkawa’s United States patent No 6,311,391 B1 dated 6 November 2001 (“the Shinkawa Patent”); and
(e) ASA’s flip chip die bonder known as the AFC 800, which incorporates the ASA Patent (“the AFC 800”).
A short note on terminology is in order at this particular juncture. A die is a small block of wafer, usually a semiconducting material, on which a given functional circuit is fabricated. “Die bonding” is the attachment of the semiconductor die either to its package or to some other substrate. The die is picked from a wafer table, aligned to a target pad on the substrate, and then permanently attached via means of a solder or epoxy bond. This process should be distinguished from that of “die sorting”, which merely involves the sorting of good and bad dies or chips on a wafer.
[12] Preferring the opinion of the Respondent’s expert, Mr John Briar (“Mr Briar”), over that of the Appellant’s expert, Associate Professor Tay Meng Leong (“A/P Tay”), with respect to the state of the art as embodied in the patents and machine referred to in the preceding paragraph, the Judge found that the Patent failed on two of the three conditions governing patentable inventions in s 13(1) of the Act (as set out below at [15]), namely, that the invention must be new and must involve an inventive step. In arriving at this conclusion, the Judge was careful to highlight the fact that while Mr Briar had “extensive experience in flip chip packaging technology and equipment development for use in the semi-conductor industry” (the GD at [18]) and possessed “personal knowledge and experience of what the state of the art was at the material time” (the GD at [28]), A/P Tay was not in the semiconductor industry and was unfamiliar with flip chip machines prior to receiving his instructions from the Appellant.
[13] The Judge also observed that the crux of the inventive concept in the Patent, viz, the “carrying out of vision inspection concurrently during the rotation of the pick up heads and via the through opening between the pick up heads”, was already disclosed in the state of the art. Specifically, he noted that (see the GD at [36]):
[14] In the circumstances, therefore, the Patent was declared invalid and was accordingly revoked.
The law on patentability
Introduction
[15] In order for an invention to be a patentable one, it must satisfy the three conditions laid down in s 13(1) of the Act. These are that:
(a) the invention is new;
(b) it involves an inventive step; and
(c) it is capable of industrial application.
The Respondent maintains on appeal that the Patent should be invalidated as it is lacking in both novelty as well as in an inventive step (viz, that conditions (a) and (b) above have not been satisfied).
Novelty
[16] The conditions for novelty (or, utilising the statutory language, that the invention concerned is “new”) are set out in s 14 of the Act as follows:
[17] In the English Court of Appeal decision of The General Tire & Rubber Company v The Firestone Tyre and Rubber Company Limited and Others [1972] RPC 457 (“General Tire”), Sachs LJ set out the test for determining the novelty in a patent under s 32(1)(e) of the Patents Act 1949 (c 87) (UK) in the case of anticipation by prior publication as follows (at 485–486):
In other words, in seeking to establish anticipation for the purposes of discounting (or even ruling out altogether) novelty, the directions contained in a prior publication (which constitutes part of the state of the art) must be so clear that following those directions must inevitably lead to something that would, if the patentee’s patent were valid, infringe the patentee’s claim. Sachs LJ’s views were endorsed by this court in Merck & Co Inc v Pharmaforte Singapore Pte Ltd [2000] 2 SLR(R) 708 (“Merck & Co Inc”) and in Genelabs Diagnostics Pte Ltd v Institut Pasteur and another [2000] 3 SLR(R) 530 (“Genelabs Diagnostics”) as well as by the Singapore High Court in Dextra Asia Co Ltd and another v Mariwu Industrial Co (S) Pte Ltd and another suit [2006] 2 SLR(R) 154 (“Dextra Asia Co Ltd”) (for related proceedings, see Mariwu Industrial Co (S) Pte Ltd v Dextra Asia Co Ltd and another [2006] 4 SLR(R) 807 and Dextra Asia Co Ltd and another v Mariwu Industrial Co (S) Pte Ltd [2007] 3 SLR(R) 29). It has further been held in Genelabs Diagnostics that the prior publication must “not only identify the subject matter of the claim in the later patent” [emphasis added] but must also be an “enabling disclosure” (at [24]). This means that an invention “would be anticipated by a piece of prior art if the teachings disclosed in this prior art are sufficiently clear and complete to allow the skilled addressee to make the invention” [emphasis added] (see Ng-Loy Wee Loon in Law of Intellectual Property of Singapore (Sweet & Maxwell Asia, Rev Ed, 2009) at para 30.1.38 as well as SmithKline Beecham Plc’s (Paroxetine Methanesulfonate) Patent [2006] RPC 10, where the House of Lords clarified that the concepts of disclosure and enablement are distinct).
[18] And, in the Singapore High Court decision of Trek Technology (Singapore) Pte Ltd v FE Global Electronics Pte Ltd and others and other suits [2005] 3 SLR(R) 389 (“Trek Technology”) (affirmed by this court in FE Global Electronics Pte Ltd and others v Trek Technology (Singapore) Pte Ltd and another appeal [2006] 1 SLR(R) 874 (“FE Global Electronics”)), Lai Kew Chai J laid down the following principles for determining anticipation by the prior art (at [87]):
In conjunction with the observations of Sachs LJ in General Tire, the principles set out by Lai J above are instructive in resolving the question of fact as to whether or not an invention may be considered “new” for the purposes of s 14 of the Act.
Inventive step
[19] Section 15 of the Act defines an “inventive step” by stating that “[a]n invention shall be taken to involve an inventive step if it is not obvious to a person skilled in the art”. In the Singapore High Court decision of Ng Kok Cheng v Chua Say Tiong [2001] 2 SLR(R) 326 (“Ng Kok Cheng”), Judith Prakash J observed as follows (at [16]):
[20] The four-step test for inventiveness espoused in the English Court of Appeal decision of Windsurfing International Inc v Tabur Marine (Great Britain) Ltd [1985] RPC 59 (“the Windsurfing test”) provides a structured approach to resolving what would be an otherwise vague inquiry, and has been followed and applied in numerous cases, both foreign as well as domestic. As set out in Trek Technology (at [94]), the Windsurfing test may be formulated as follows:
(a) Identify the inventive concept embodied in the patent in suit.
(b) The court then assumes the mantle of the normally skilled but unimaginative addressee in the art at the priority date, imputing to him what was, at that date, common general knowledge in the art in question.
(c) Identify what, if any, differences exist between the matter cited as being “known or used” and the alleged invention.
(d) The court then asks itself the question 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.
It should be noted that the reference to “common general knowledge” in (b) above does itself raise difficulties which were noted by this court in First Currency Choice Pte Ltd v Main-Line Corporate Holdings Ltd and another appeal [2008] 1 SLR(R) 335 (“First Currency Choice Pte Ltd”) at [38]–[41] (for related proceedings, see the Singapore High Court decisions of Main-Line Corporate Holdings Ltd v Overseas Bank Ltd and another [2008] SGHC 55, Main-Line Corporate Holdings Ltd v United Overseas Bank Ltd and another (First Currency Choice Pte Ltd, third party) [2010] 1 SLR 189 and Main-Line Corporate Holdings Ltd v United Overseas Bank Ltd and another [2009] SGHC 212), but which do not arise in the context of the present appeal (cf also the proposed solution proffered by Ng-Loy Wee Loon in Law of Intellectual Property of Singapore at para 30.1.47). It should also be noted that whilst the various criticisms of the Windsurfing test itself were acknowledged by this court in First Currency Choice Pte Ltd, V K Rajah JA, who delivered the judgment of the court observed (at [44]) that “[n]onetheless, the Windsurfing test … appears to be here to stay”. Indeed, Rajah JA was of the view that the Windsurfing test could also prevent erroneous ex post facto reasoning (see also [101] below). He also proceeded to add thus (at [44]):
However, Rajah JA also sought (bearing in mind the various criticisms of the Windsurfing test) to bring some practical balance to this test by observing as follows (at [45]):
[21] In so far as the concept of the “notional skilled person” is concerned, Rajah JA made the following observations in First Currency Choice Pte Ltd (at [28]):
For the purposes of steps (b) and (d) of the Windsurfing test for obviousness (see [20] above), the court should therefore assume the mantle of only a skilled but unimaginative addressee. Care must be taken not to either go beyond or fall short of this rather nuanced standard as set out above.
Patent construction
[22] The courts have consistently endorsed adopting a “purposive construction” of patent claims so as to determine the essential features of any particular invention. This approach received judicial affirmation in the seminal decision of the House of Lords in Catnic Components Limited and Another v Hill & Smith Limited [1982] RPC 183, where Lord Diplock observed as follows (at 242–243):
[23] This approach was reiterated more recently in the (also) House of Lords decision of Kirin-Amgen Inc v Hoechst Marion Roussel Ltd [2005] RPC 9 (“Kirin-Amgen”) as constituting the bedrock of all patent construction. Lord Hoffmann elaborated on this principle (at [32]) as follows:
[24] This sensible approach has also been adopted by this court in FE Global Electronics, where it was held (at [14]) that the purposive construction of patent claims was preferred as “it balance[d] the rights of the patentee and those of third parties” (see also the decision of this court in First Currency Choice Pte Ltd (at [25]–[26]) as well as the Singapore High Court decision of Merck & Co Inc v Pharmaforte Singapore Pte Ltd [1999] 3 SLR(R) 1072 (“Merck & Co Inc (1)”) at [54] (affirmed by this court in Merck & Co Inc)). A purposive construction of the claims would give the patentee the full extent, but no more than the full extent, of the monopoly which a reasonable person skilled in the art, reading the claims in context, would think that he (the patentee) was intending to claim.
[25] Clearly then, the starting point in patent construction is to ask the following threshold question: What would the notional skilled person have understood the patentee to mean by the use of the language of the claims? This question, however, is difficult to answer without first coming into possession of the knowledge that a notional skilled person would have. Accordingly, a brief excursion through the state of the art, as referred to above (at [11]), is necessary. However, a preliminary objection concerning the novelty of the ASA Patent must be dealt with first.
A preliminary objection: Did the ASA Patent form part of the state of the art having regard to ss 14(2) and (3) of the Act?
[26] Before proceeding to set out what the state of the art was in the context of the present appeal vis-à-vis the requirements in relation to both novelty as well as an inventive step, it would be useful to set out the relevant parts of s 14 once more. Sections 14(2) and (3) of the Act provide as follows:
[27] Indeed, the provisions just set out in the preceding paragraph are especially relevant in the context of a preliminary objection raised by counsel for the Appellant, Dr Stanley Lai (“Dr Lai”), to the effect that the ASA Patent did not, on a proper construction of ss 14(2) and (3) of the Act, even form part of the state of the art to begin with. Let us now turn to the detailed argument itself, commencing with the relevant factual background.
[28] The ASA Patent was published on 21 June 2004 and was granted in September 2005. All this took place after the Patent’s priority date of 16 February 2004. However, the priority date for the ASA Patent was on 7 January 2002. In other words, while the ASA Patent was published after the priority date of the Patent, it had been filed before that same priority date.
[29] Section 14(3) of the Act makes it clear that the state of the art in relation to the Patent comprises matter contained in an application for another patent (ie, the ASA Patent) which was published on or after the priority date of the Patent, if both conditions in ss 14(3)(a) and (b) of the Act are met. The condition in s 14(3)(b) of the Act is easily met in the context of the present appeal: the priority date for the ASA Patent was earlier than the priority date for the Patent. It is, however, less clear whether the requirements in s 14(3)(a) of the Act have been met. It is to this particular issue that our attention now turns.
[30] The Appellant argued that the Respondent had failed to satisfy the requirement that the subject matter impugning the novelty of the Patent was contained in the application for the ASA Patent as filed and as published, pursuant to s 14(3)(a). It contended that the burden lay on the Respondent to demonstrate that all the requirements of s 14(3)(a) had been met, since it was the Respondent who was asserting that the ASA Patent was part of the state of the art.
[31] In particular, the Appellant argued that in order to qualify to be considered as prior art under s 14(3)(a), the Respondent had to demonstrate that what was published under the ASA Patent corresponded exactly to what was filed for that same patent. This, however, the Respondent had apparently failed to do.
[32] With respect, this particular argument is extremely technical in nature. More importantly, it appears to discount the fact that s 14(2) of the Act already provides (by way of an alternative, and no less efficacious, “route”, so to speak) that the state of the art includes “all matter … which has at any time before the priority date of [the Patent] been made available to the public (whether in Singapore or elsewhere) by written or oral description, by use or in any other way” [emphasis added]. Under s 14(2), therefore, the ASA Patent would have already been considered as well as classified as prior art, given the international publication date of its Patent Cooperation Treaty (“PCT”) application on 17 July 2003. We pause here to note the Appellant’s attempt to curtail reliance on s 14(2) on the basis that it had not been raised in the court below. Even assuming that this was the case, the nature of the argument is such that it ought to be considered by this court and would not, in any event, entail any adduction of further evidence of any kind (see also generally the decision of this court in Panwah Steel Pte Ltd v Koh Brothers Building & Civil Engineering Contractor (Pte) Ltd [2006] 4 SLR(R) 571).
[33] We do not, with respect, consider the Appellant’s objections to the effect that “[t]he Respondent has not satisfied proving that the PCT Application meets the publication requirement under the Act” to be persuasive. While the PCT system does not grant full patent protection in the same way that a national patent would, it is well-known that PCT applications, once lodged, are subject to search, examination and challenge. The PCT system permits the suspension - for up to 18 months - of national or regional filing procedures while the inventor refines and assesses the prospects of his patent. Such suspension, and the corresponding delay in national registration, would only make sense if the inventor is assured of the priority of his invention.
[34] In any event, during and after publication of a PCT application, written descriptions are made available to the public – whether in Singapore or elsewhere – in line with the stipulations in s 14(2). It is perhaps useful to note at this juncture that in FE Global Electronics, a published PCT application, filed by the Lexar Corporation, was considered sufficient to constitute prior art for the purposes of assessing novelty. Another PCT application in the same case, subsequently converted to a domestic Singapore patent, was also considered by this court to form part of the state of the art.
[35] It is also noteworthy that the Appellant filed its own PCT application for the Patent on 31 January 2005, almost one and a half years later than ASA’s PCT application.
[36] Returning to the Appellant’s concerns centring on s 14(3)(a) and relating to the possible non-correspondence of matters filed against matters published (see [31]-[32] above), we are of the view that this is something that can be easily verified, even by the court, by retrieving the relevant records and effecting the necessary comparisons. However, there is no necessity to adopt this particular course of action in the context of the present appeal (or, indeed, to even contemplate utilising s 14(3) in the first place), given the applicability of s 14(2) as noted above. That having been said, and although there is no necessity for us to express a concluded view on this particular argument, it would appear that, adopting an approach that looks more to the substance rather than the form, there is some doubt as to whether such an argument would be persuasive in the light of the views we expressed at the commencement of this particular paragraph.
[37] With respect to the state of the art for an inventive step, s 15 states as follows:
[38] Accordingly, given the operation of s 14(2), regardless of the applicability of s 14(3), the Appellant cannot avail itself of the arguments contained in the preliminary objection for the purposes of assessing the state of the art under s 15.
[39] Before we turn to consider what the state of the art in the context of the present appeal in fact was (which will set the appropriate backdrop required for considering the two legal issues in this appeal, viz, whether or not the Patent is new (or novel) and whether or not it involved an inventive step – both of which the Judge held were absent, thus resulting in (as we have already recounted above) judgment in favour of the Respondent in the court below), we would like to address two important preliminary points which, albeit more general in nature, are not only germane to the present appeal but also have wider implications for other cases as well. The first preliminary point relates to the role of logic and legal argument in the context of patent disputes or, to be more precise, what ought to be avoided by counsel in relation to such disputes and (as we shall see) in relation to any other dispute in any other area of the law for that matter. The second preliminary point relates to the role of expert evidence, which evidence is almost invariably (and understandably) adduced in the context of patent disputes. Let us turn, then, to each of these preliminary points seriatim before proceeding to consider the particular issues and facts that are raised in the context of the present appeal.
A first preliminary point: the importance of logic and rational argument as well as the danger of over-generalisation and over-abstraction
[41] It is important to emphasise, at this juncture, that the use of the general principles of logic and rational argument do not, in any way whatsoever, undermine the legal principles that are germane to a particular area of the law (in this case, that relating to patents). Indeed, on one view, the legal principles themselves are the product – in part at least – of the process of logic as well as rational argument. It is no surprise, therefore, that the application and/or development of general principles of logic and rational argument must (as we shall see) be integrated with the particular legal principles as well. In other words, the general principles of logic and rational argument and the particular legal principles (here, in relation to the patent law) are complementary in nature.
[42] Turning, then, to the perspective of general principle and logic, there is an irreducible tension between universality and particularity (or, to use equally clear terminology, between generality and specificity; cf also the decision of this court in Spandeck Engineering (S) Pte Ltd v Defence Science & Technology Agency [2007] 4 SLR(R) 100 at [28] (in the context of liability in negligence for pure economic loss)). However, as we shall see, this is not a tension that necessarily leads to paralysis or inaction; on the contrary, it (as we shall attempt to demonstrate in a moment) is, in fact, mediated via an interactive process that defines what the law and decision-making are all about in the practical sphere. Indeed, the unique combination (or, perhaps more appropriately, integration) of the universal and particular, of the general and the specific, and of the normative and the descriptive, produce not only decisions for the case at hand but also precedents for the future.
[43] To elaborate, when an argument is pitched at too high a level of abstraction, it will probably attract no controversy (and may even achieve consensus) but, because it is at such a “rarefied” level, it is actually unhelpful in most cases (for example, virtually all persons will agree that it is desirable to “do good”, but, of course, once we descend into the specifics (ie, what “doing good” means to each person), the opposite extreme will be reached, with little (if any) agreement amongst all concerned). Indeed, such an approach is especially unhelpful in cases involving patents, simply because, in many cases (such as this), a necessary step involved in the court’s decision entails a comparison of the competing patents. At this point, specific details must necessarily come into play. However, it is also acknowledged that the court cannot go to the other extreme in focusing only on the trees and missing the wood. This is, however, unlikely to happen in practice because the court must perforce come to an overall conclusion, and that would entail gathering together all the relevant (and specific) details and analysing them in the context of the applicable legal rules and principles before arriving at an overall conclusion. Here, we see the interaction between the universal and the particular, between generality and specificity, as well as between the descriptive and the normative. It is not something which can be reduced to a mechanistic formula simply because life in general and the decisions of courts (as a consequence) in particular are too complex to admit of a rigid formulaic approach. However, that does not mean that there are no general principles that ought to guide the court. As already mentioned, there are, of course, the general principles of logic and argumentation. More importantly, each area (as well as sub-area) of the law will also embody general legal rules and principles which are to be applied to the facts at hand. In the specific sphere of patent law, we would think that, in addition to the general legal rules and principles (embodied in ss 13, 14 and 15 of the Act; see also generally [15]-[25] above), the court would also need to apply those rules and principles (as well as those relating to general logic and argumentation) to the specific claims which are contained within the patent(s) concerned. Indeed, these claims constitute the specific factual matrix as well as context which the court takes into account in arriving at its decision. Again, however, it is of the first importance to emphasise that this is an interactional process that is not susceptible of mechanical application. Put crudely, it entails an exercise of judgment by the court concerned in applying the general principles of law as well as logic to the facts concerned, and as understood in the context in which they occur.
A second preliminary point relating to the role of expert evidence
[44] The difficulties engendered by the issue of bias with regard to experts for the respective parties are, unfortunately, perennial in nature. Indeed, in the Singapore High Court decision of Khoo Bee Keong v Ang Chun Hong and another [2005] SGHC 128, this difficulty, amongst others, was referred to, as follows (at [68] and [82]-[87]):
[45] The issue of the need for review in this area of the law referred to in the last paragraph of the passage quoted in the preceding paragraph was also alluded to by counsel for the Appellant, Dr Lai. Dr Lai emphasised that the issue in relation to the possible (or even probable) bias of experts is an especially significant one in patent cases and suggested that a new system involving impartial assessors might be a possible solution. The special difficulties just mentioned might well be the case, given the very nature of the inquiry in such cases, in which the court also often does not possess the requisite technical expertise. However, it is obviously also significant with respect to all other areas of the law as well. In the meantime, however, it may well be wise and prudent for the parties concerned in future cases (especially of this nature) to apply to the court to appoint an impartial and objective expert (whose views they would agree to abide by) pursuant to O 40 r 1 of the Rules of Court (Cap 322, R 5, 2006 Rev Ed) (which permits the court to appoint one or more independent experts on the application of any party which, under O 40 r 1(2) “shall, if possible, be a person agreed between the parties and, failing agreement, shall be nominated by the Court”). It should, however, also be noted that this same Rule also permits the court to appoint one or more independent experts “on its own motion”, although, in the nature of things (particularly in patent disputes), this particular avenue will probably prove to be less than practical.
[46] Returning to this particular appeal, we note that one of the Appellant’s main arguments centred on the alleged bias on the part of Mr Briar. In particular, Dr Lai noted that he was one of the actual inventors of the machine which constituted the subject matter of the Respondent’s (and, more importantly, opposing) patent. Indeed, he argued, the Judge appeared to accept wholly this particular expert’s views without more (cf also [12] above).
[47] However, as was observed in the Singapore High Court decision of Asia Hotel Investments Ltd v Starwood Asia Pacific Management Pte Ltd [2007] SGHC 50 at [206]–[207], for the purposes of determining whether the evidence of an expert should be discounted, the relevant test is one of actual partiality, rather than merely the appearance of partiality. In any event, however, there is insufficient evidence on the record to demonstrate that the Judge did not discount the views of Mr Briar accordingly. The mere acceptance (even in its entirety) of such views does not necessarily entail the absence of such a discount. As it turns out, however, it is clear (for reasons that will be elaborated upon below) that the views by the experts appointed by both parties were not, in our view, critical, except in so far as they enabled this court to understand what were the precise claims in the respective patents. To this end, both the experts’ views were helpful, bearing in mind that it is, in the final analysis, the court which decides whether or not the requisite legal criteria have (or have not) been satisfied. It should also be borne in mind that the expert evidence is - in so far as the issues as to whether the Patent is novel and/or whether an “inventive step” exists vis-à-vis the state of the art - itself subject to the legal criteria contained in ss 14 and 15 of the Act, respectively (as well as the accompanying legal rules and principles).
[48] In particular, the court must bear in mind the fact that the legal criteria just set out in the preceding paragraph does not (in the context of whether there is an inventive step within the meaning of s 15) assume knowledge and expertise that goes beyond what a reasonable person “skilled in the art” would possess (see also [21] above as well as FE Global Electronics (at [22])). This is not an unimportant practical point, simply because many experts would not themselves fall within this particular category of persons “skilled in the art” as they would possess extraordinary knowledge as well as expertise. It is therefore important for such experts to bear this particular point in mind when they tender their evidence before the court in future. It would also be wise for counsel (and, if necessary, the court) to remind the experts concerned of this point if they appear to have forgotten it.
The state of the art described
[49] In light of the Appellant’s (not unwarranted) contention that “significant portions of the [Judge’s] grounds of decision originate from the evidence of the Respondent’s expert [ie, Mr Briar]” (see also [12] and [46] above), care will be taken here to describe each and every piece of prior art as objectively and neutrally as possible.
The ASA Patent and the AFC 800
[50] The ASA Patent (see Appendix B) is described in its certificate as a “flip chip bonder and method therefor”. Essentially, it is a wafer chip bonding apparatus comprising a rotating pickup turret “having a plurality of pick-up heads, wherein each of the plurality of pick-up heads is adapted for picking at least one component having a picking orientation” [emphasis added]. This is different from the Patent, which specifically lays out its requirement for two and only two rotating pickup heads (see also [9] above).
[51] In the ASA Patent, it is mentioned that there is an optically assisted pick-and-place system for inspecting the wafer chips, but few details are provided. In particular, specific instruction as to concurrent pickup and inspection is not alluded to. At p 7 of that patent, it is only stated as follows:
This is repeated at pp 11–12 of the patent.
[52] The idea of using two heads for the pick-up turret assembly is mentioned in the ASA Patent, but in a negative, “teaching away”, sense:
[53] It is also significant to note that the ASA Patent also envisages an increased throughput (this is stated right at the outset of the patent itself), much like the Patent does. In the case of the ASA Patent, however, the increased throughput is:
Indeed, the overall process of the ASA Patent relies on unidirectional circular rotations, as opposed to the clockwise and anticlockwise 180-degree rotations of the Patent (see [9] above). It is significant that the “concurrency of operations”does not include the simultaneous inspection of the next wafer chip (or die) to be picked up.
[54] The AFC 800, a die bonder machine with four pickup heads, incorporates the ASA Patent (see Appendix C). As noted pointedly in the Appellant’s Case (“the AC”), while the machine implements an optical system for the inspection of the next wafer chip to be picked up, the camera is positioned parallel to the substrate, at a nine o’clock position to the rotating pickup turret. In order to establish an optical line to the wafer chip to be picked up, the optical path is accordingly bent by 90 degrees via a prism/mirror located in the centre of the rotating pickup turret. This is in contrast to the optical system countenanced in the Patent, where the optical path between the wafer chip and the camera is vertical, with the camera mounted directly above the two-headed pickup and the wafer chip below.
[55] It is not disputed by the Appellant, however, that the arrangement of the optical system in the AFC 800 does allow for the concurrent inspection of wafer chip(s) as each of the four pickup heads is rotated around the rotating turret.
The Matsushita Patent
[56] The Matsushita Patent (see Appendix D) describes an apparatus and a method for bonding a chip via three pickup heads, and in a manner similar to that found in the ASA Patent. Importantly, however, the Matsushita Patent does make express mention (although the word “preferably” is also utilised as well in so far as the use of camera is concerned) of the ostensibly concurrent inspection and pickup of two wafer chips as follows:
[57] Simply put, Claim 2 (as set out in the preceding paragraph) defines that the camera observes the next wafer chip to be picked up whilst the “plurality of nozzles” (in the incarnation laid out in the Matsushita Patent, being three in number) are displaced from their regular pickup positions, thereby leaving a gap for an optical path to be established between the camera and the wafer chip below. As the Appellant’s own expert, A/P Tay, notes in his own report:
Further corroboration of the concurrent pickup and inspection afforded by the device may be found at column 6 of the Matsushita Patent:
[58] However, an important difference between the process envisioned by the Matsushita Patent and that embodied in the Patent is that the former does not countenance true concurrence in inspection and pickup. Under what is envisaged by the Matsushita Patent, the inspection of the next wafer chip in line takes place when the three-headed pickup head has rotated away from the “position for picking up”. The inspection does not take place as the pickup head is in the midst of rotating. This is necessarily a function of the unidirectional incremental rotation mechanism employed by the Matsushita Patent, which is not dissimilar to the mechanism mentioned in the ASA Patent and implemented in the AFC 800. As can be seen from the extract in the preceding paragraph, the inspection for defects only takes place when “all three heads have been in positions displaced” [emphasis added]; not (instead) when “all three heads are being displaced” or are “in the process of rotation”. Similar language is employed in the relevant claims (see [56] above), where it is stated that observation of the next chip to be picked up only takes places “under a condition wherein one nozzle of said plurality of nozzles has been horizontally rotated to a position where said one nozzle hands one of said chips over to said transferring head” [emphasis added].
[59] It should be noted, in passing, that the rotation of the pickup head turret in the Matsushita Patent is executed about a vertical axis, whereas the rotation of the two-headed pivoting pickup in the Patent is executed about a horizontal axis. In the court below, the Respondent’s expert, Mr Briar, opined that the horizontal orientation disclosed in the Matsushita Patent was not a material difference for the purposes of comparison vis-à-vis the Patent. He was not challenged on this view.
The National Semiconductor Patent
[60] Like the ASA Patent and the Matsushita Patent, the die-handling device countenanced in the National Semiconductor Patent (see Appendix E) comprises a “plurality of stations” [emphasis added] mounted on a rotary flipper. The device is responsible for transferring a wafer chip (or die) from a stretched wafer to a tape cavity. Briefly, it functions as follows. A wafer chip is placed within a first station in a first orientation. While the chip is held, the wheel portion of the rotary flipper rotates and the next station receives another chip. When the first station is in an unloading position, the chip is released. At this point, the chip is oriented in a second orientation. While the chip is released into a cavity of a tape and reel, vacuum pressure is applied to hold on to the second chip.
[61] As noted by Mr Briar himself, the National Semiconductor Patent is “very limited in text description”, and only mentions that a camera may be used to determine if the chip or die is correctly released into the cavity. The only graphical illustration of any vision inspection system is in an example of prior art found in FIGS 1A and 1B of the National Semiconductor Patent.None of these allusions to an optical inspection system or a camera dictates the concurrent inspection of any wafer chips during operation of the rotary flipper. Significantly, in a graphical representation of the work steps of the National Semiconductor Patent, prepared by Mr Briar and the Respondent, there are no indications of any optical inspection system present. Indeed, Mr Briar is at pains to qualify himself in asserting the anticipation of the Patent by the National Semiconductor Patent, as follows:
The Shinkawa Patent
[62] The Shinkawa Patent (see Appendix F) describes a “flip-chip bonding apparatus” in which a “die inverting device is installed on an optical recognition device” so that a single “vacuum suction chucking nozzle that is provided on the die inverting device” can rotate between a “pick-up position” and a “die transfer position”.
[63] In other words, the Shinkawa Patent envisages a one-armed pickup head swivelling about an optical inspection system located directly above the wafer sheet (the surface on which the chips or dies are placed prior to pickup). When the pickup head is swivelling to transfer an already-picked-up wafer chip to the “die transfer position”, a direct optical line may be established between the optical inspection system and the wafer sheet, since the one-armed pickup head is out of the way. Accordingly, the optical inspection system immediately inspects the next wafer chip for pick-up at the “pick-up position”. The pickup head then swivels back towards the “pick-up position”, having deposited the chip it was holding, and the cycle of pickup and inspection begins again.
[64] While reference to an “optical recognition device” is explicit in the Shinkawa Patent, inspection is via a bent optical path courtesy of a prism/mirror, much like that seen in the AFC 800. Visual information about the next wafer chip to be picked up enters the optical recognition device via an open window (21a) and is refracted into an image pick-up element (25). Again, this is in contrast to the situation of the camera in the Patent, which is directly (ie, vertically) above the wafer chip to be picked up. It should also be noted that the optical recognition device in the Shinkawa Patent is integrated with the swivelling die inverting device, whereas the camera in the Patent is mounted independently of the rotating two-headed pickups.
[65] Another significant difference between the Patent and the Shinkawa Patent is the fact that the device envisaged by the Shinkawa Patent possesses only one arm. This is true of whichever of the three embodiments one chooses to opt for under the Shinkawa patent. Therefore, although the die inverting device in the Shinkawa Patent (not unlike the Patent, but only in this particular respect) also alternates between rotating 180 degrees in one direction and then 180 degrees in the other direction, it operates with only one pickup head throughout.
[66] Mr Briar stated in his report that “[w]hilst the [Appellant’s] Patent has a vision inspection system that is not an integrated part of the pivoting unit and that Shinkawa’s Patent utilizes a vision inspection system that is an integrated part of the pivoting unit, nonetheless, the concept of vision inspection through an opening in the pivoting unit is similar”. With respect, this is factually incorrect. The Patent performs its vision inspection through an opening in the pivoting dual pickup head (see [9] above). The Shinkawa Patent, on the other hand, performs its vision inspection only when the single pickup head has moved out of the way. The difference is a subtle one, and should be borne in mind when applying the law on novelty as well as that on obviousness to the Patent later in this judgment.
[67] We turn now to consider the respective patents themselves in the context of the two main legal issues in the present appeal, viz, whether the Patent was novel and/or whether the Patent involved an inventive step. It bears reiterating that, in the court below, the Judge held in favour of the Respondent in respect of both these issues.
Is the Patent novel vis-à-vis the state of the art?
[68] Turning, first, to the issue as to whether the Patent is novel vis-à-vis the state of the art (which has been described at [50]-[66] above), generally speaking, it is not permissible to assemble all the pieces of prior art together into a “mosaic”, and then to compare the invention under question against this “mosaic” (see [18] above as well as Dextra Asia Co Ltd (at [53])). The invention must be compared against each individual piece of prior art, and the question is repeated each time: has this particular piece of prior art anticipated the invention? The limited exception to this rule against “mosaicing” only states that where an earlier document is referred to in a later document, it is possible to refer to the earlier document; but, even then, only to the extent of the specific portions referred to by the later document. Reference may also be made, for example, to the decision of this court in Genelabs Diagnostics (at [26]–[28]). This limited exception, however, does not apply in the context of the present appeal simply because the examples of prior art under consideration do not cite one another as references.
The ASA Patent and the AFC 800
[69] The ASA Patent may be dismissed as a potential anticipator of the Patent right from the outset. Even taking the Judge’s interpretation of the inventive concept at face value – “inspection on the fly”, as noted at [12] of the GD – the ASA Patent fails to pass muster. As has already been noted at [51] above, the ASA Patent does not provide any enabling disclosure with respect to a concurrent pickup and inspection system; it only alludes to “[o]ptically assisted pick and place systems, as will be known to one skilled in the art”. Even so, it sanctions these optically assisted pick and place systems in a non-imperative sense, stating that they “may be employed to determine the location and orientation of the target location 122A [ie, the die]” [emphasis added].
[71] A similar approach was, in fact, adopted by the expert for the defendant in Dextra Asia Co Ltd, where the expert “assumed that cold forging was not new or inventive because others who had patented processes before had used the term “forging”, which could mean either hot or cold forging” (at [54]). Not surprisingly, Tan Lee Meng J rejected such an approach which was too general and, indeed, led to confusion instead (at [54]).
[72] Reference may also be made to the House of Lords decision of Biogen Inc v Medeva Plc [1997] RPC 1, where Lord Hoffmann observed as follows (at 34):
[73] The learned law lord also observed (in the same decision) thus (id at 52):
However, much would, of course, depend on the precise facts of the case concerned itself (see, for example, the recent House of Lords decision of Generics (UK) Ltd v H Lundbeck A/S [2009] RPC 13).
[74] Consider, also (to take an everyday example), the ubiquitous television, which inventive concept might be summarised as “a viewable screen for the projection of moving images”. Liquid crystal display (“LCD”) and plasma technology (quite apart from CRT (ie, cathode ray tube) and projection televisions) have separately attained this end through starkly contrasting processes. LCD screens are, in layperson’s terms, sandwiches made up of liquid crystal pushed into the space between two glass plates. Images are created by varying the amount of electrical charge applied to the crystals. Plasma screens, on the other hand, utilise a matrix of tiny gas plasma cells charged by precise electrical voltages to create a picture. Both technologies have their own separate patents, similar end results notwithstanding.
[75] Applying the above principles to the facts of the present case, the Patent (the key characteristics of which have been set out above (at [6]-[9])) is, in our view, a genuinely different (and specific) method of achieving the desired result of “inspection on the fly”. In particular, the true inventive concept which the Patent embodies (and which is required under the first stage of the Windsurfing test (at [20] above)) is the genuinely concurrent inspection (as opposed to stop-start “concurrence” (see [58] above)) of a wafer die through a pivoting part located between two pickup heads.Neither the ASA Patent nor the AFC 800 achieves the said result in the same manner.
[76] It should also be noted that the Patent also includes claims for a method and process (see [8] above), which at first glance operates according to a procedure that is very markedly different from the existing art (two pickup heads only, and alternating 180-degree motion of the pickup heads, with a through opening for establishing an optical path within the pivoting portion of the device). We have already noted that Claims 1 through 6 of the Patent (see [7] above) envisage a device that is not, in the final analysis, similar to the state of the art. This finding of novelty is further confirmed (as just noted) via the processes enshrined in Claims 7 through 10 of the Patent (see [7] above).
[77] Turning now to compare both the device and process countenanced in the Patent with those in the AFC 800, we see that the primary material differences between the two lie in the following aspects:
(a) The optical inspection system described by the Patent does away with the need for a prism/mirror to bend the optical path (which is the case with the AFC 800), since the optical path is a vertical one from the camera down to the wafer table.
(b) The two pickup heads in the Patent move in a reciprocal clockwise/counter-clockwise 180-degree pivoting motion, in contrast to the rotating pickup turret in the AFC 800, which performs unidirectional incremental rotations.
[78] The differences referred to in the preceding paragraph relate to neutral and objective facts that are not - and, indeed, cannot be - disputed by either party and which are, more importantly, claimed in the Patent itself (see [7] above). What is disputed are the alleged advantages of the Patent over the AFC 800 as a result of these two differences. With respect to the difference in para 77(a) above, the Appellant argued (at para 325 of the AC) that the Patent is novel vis-à-vis the state of the art as it implements:
In so far as the difference in para 77(b) above is concerned, the Appellant argued that the Patent is novel vis-à-vis the state of the art, as follows (see the AC at paras 328-329):
[79] The Respondent chose to respond to the above with the usual procedural arguments of “not raised before the Court below” and also that “A/P Tay did not give evidence on this before the court”. This was not, however, for want of trying to meet the Appellant’s substantive reasoning; indeed, before resorting to these technical procedural arguments, the Respondent had attempted to field its own substantive rebuttals. In contrast to the Appellant’s contentions, however, these rebuttals appeared to be extremely general and lacking the convincing specificity possessed by the former. For example, the Respondent argued thus:
And, in a similar vein, the Respondent also argued as follows:
[80] The disclosure in a patent claim must enable a “skilled person in the art” to make the invention (see Merck & Co Inc (1), applying the House of Lords decision in Merrell Dow Pharmaceuticals Inc and Another v HN Norton & Co Ltd and Others [1996] RPC 76). This standard is arguably similar to that of the “normally skilled but unimaginative addressee”, as required by the second stage of the Windsurfing test for obviousness (see [20] above). Here, however, even without assuming the mantle of a person skilled in the art, the logical arguments put forth by the Appellant (at [78] above) are, in fact, extremely clear and certainly do not require any expert clarification. A mirror/prism (that is an integral part of the AFC 800) is an additional, unnecessary step that results in a less efficient process. In contrast, the 180-degree alternating rotation (that is an integral part of the Patent) is not only different but also results in a different (and more efficient) process. Indeed, the Respondent’s own witness, Mr Boh Teck Keong (“Mr Boh”), admitted that “[i]f you do a continuous turning [as opposed to a 180-degree alternating rotation], you will have a cumulative error that eventually affect[s] the stopping accuracy”.
[81] In any event, regardless of the precise advantages (or disadvantages) engendered by the changes wrought in the Patent vis-à-vis the AFC 800, it is plain to see that both the device and the method envisaged in the Patent facilitate “inspection on the fly” in a manner different from that countenanced by the AFC 800. When we recall the analogy furnished above (at [74]) with respect to separate patents being granted to plasma and LCD televisions, this finding becomes almost self-evident. Accordingly, we find that that there is novelty in the Patent relative to the AFC 800.
The Matsushita Patent
[82] As has already been pointed out (see [56] above), the Matsushita Patent is (unlike the Patent) capable of only a limited form of concurrent inspection and pickup. As a result of the three rotating pickup heads having to come to a halt before inspection can take place, precious time is consumed. The Patent, on the other hand, envisages only two pickup heads at opposite ends. When one pickup head picks up a wafer die, the other pickup head is simultaneously depositing the wafer die it had already been holding onto a placing facility. When both pickup heads are in motion (ie, the pivoting part is being rotated), transferring a wafer die on one hand and returning empty to collect another wafer die on the other, concurrent inspection via the camera mounted above takes place through the transverse opening on the pivoting part. Therefore, instead of this relentless (1) transfer-inspect and (2) deposit-pickup cycle facilitated by the rotating two-headed pickup in the Patent, the three-headed pickup head in the Matsushita Patent requires an intermediate stage in which two pickup heads are stationary and doing nothing while a wafer die is picked up. Indeed, in this aspect, the Matsushita Patent resembles the sort of prior art that the Patent sought precisely to avoid. This is, in fact, underscored by reference to the following part of the description of the Patent:
From the above, it can be seen that the description of the prior art in the Patent corresponds closely to the operation of device exemplified by the Matsushita Patent. Accordingly, the Patent does not appear to have been properly anticipated by the Matsushita Patent.
[83] Another difference between the Matsushita Patent and the Patent lies in the fact that the former utilises three pickup heads (or a “plurality of nozzles”) rotating unidirectionally in increments (not unlike a ferris wheel), whereas the latter utilises two pickup heads alternating between 180-degree rotations in clockwise and anti-clockwise directions (not unlike a see-saw). The unidirectional movement of the pickup assembly is similar to that countenanced in the rotary pickup turret described in the ASA Patent (with its concomitant problems) (see [50]-[52] and [78] above).
[84] The Respondent’s expert, Mr Briar, however, cited another Matsushita patent to demonstrate that “the clockwise/anticlockwise rotation [was] nothing novel”. This, however, does not the change the fact that, for the purposes of the Matsushita Patent, the description of such alternating rotation (as found in the Patent) is absent. Indeed, this additional citation only underscores the fact that the Matsushita Patent lacks something which the Patent possesses. Further, as has already been noted, there can be no mosaicing of prior art with respect to an inquiry into novelty (see [18] and [68] above).
[85] For the above reasons, therefore, it is our view that the Patent is novel vis-à-vis the Matsushita Patent as well. Even if it is the case – which has not (significantly, in our view) been argued – that the Matsushita Patent is more efficient than the Patent, the fact remains that the Patent has accomplished the same task in a markedly different manner.
The National Semiconductor Patent
[86] The National Semiconductor Patent, like the ASA Patent and the Matsushita Patent, dictates a “plurality of stations” [emphasis added] on a wheel moving in only one direction. Additionally, the only illustration of any vision inspection system – and not even an explicitly concurrent vision inspection system at that – occurs in diagrams referring to the state of the art. In other words, any vision inspection system, concurrent or otherwise (this is not, in fact, taught by the patent), is alluded to only in diagrams of a “conventional semiconductor device transfer machine” that the National Semiconductor Patent purports to improve upon.
[87] In view of the above, we are of the view that the Patent is novel with respect to the National Semiconductor Patent as well.
The Shinkawa Patent
[88] As has already been noted earlier, the device envisaged by the Shinkawa Patent possesses only one pickup head. Further, its optical inspection system utilises the same prism/mirror concept as the AFC 800, thereby being subject to the same vulnerabilities found therein (as explained at [80] above).
[89] More importantly, the ingenuity of optical inspection taking place through a pivoting part in the Patent is absent in the Shinkawa Patent (as noted at [64] above). Optical inspection in the Shinkawa Patent simply takes place when the single pickup arm has moved out of the way. This is decidedly different from the concept in relation to optical inspection enshrined in the Patent. Indeed, it would appear that the concept utilised by the Shinkawa Patent is no different from the prior art outlined and critiqued by the Patent (quoted at [82] above). Without a second pickup head and without a pivoting part through which vision inspection may take place, the Shinkawa Patent is condemned to the same “pickup head moves away – inspection – pickup head moves back to pick up – pickup head moves away” cycle as the prior art that the Patent has sought to distinguish - and, in our view, has in fact succeeded in distinguishing - itself from.
[90] In this respect, the diagrams (ie, FIGS 9-11) provided by the Respondent’s own expert, Mr Briar, in his report are instructive. Inspection of the next flip chip die to be picked up can only occur (as illustrated in FIG 9), where the open window is facing the wafer table. At this juncture, the pickup head is facing the nine o’clock direction. The moment the pickup head moves upwards or downwards, the angle of the half-mirror for bending the optical path is changed and the optical connection line to the next flip chip die is immediately lost. This is, a fortiori, the case when the pickup head is completely at rest pointing downwards to the wafer table (for pickup (see FIG 10)), or when the pickup head is completely at rest pointing upwards to the bonding nozzle (for deposit and bonding (see FIG 11)). In both these situations, the open window is precluded from sighting the wafer table entirely (as it points towards three o’clock in the former and nine o’clock in the latter). In short, there is no real contemporaneous inspection of the wafer die in the Shinkawa Patent. In contrast, such contemporaneous inspection, which would occur without requiring the pickup head to get out of the way completely, is in fact the hallmark of the Patent. Even so, whatever concurrent inspection that can realistically take place in the Shinkawa Patent is limited to only a scintilla of time when the pickup head is pointed exactly in a nine o’clock direction (see FIG 9). This is quite different from the process envisaged in the Patent, where, by virtue of the fixed camera position, optical inspection may take place during a broader range of the rotating movement of the pivoting two-headed pickup.
[91] In addition, the fact of the fragile optical apparatus in the Shinkawa Patent having to rotate along with the pickup head surely brings with it potential complications, particularly since parts in the half-mirror and camera involve the use of mirror and glass. In any case, it is difficult to see how this rotating optical inspection system anticipates the fixed camera system implemented in the Patent. Further, it is also difficult to see how the Shinkawa Patent’s one-armed pickup head anticipates the Appellant’s elegant two-headed pickup solution, which solution permits the establishment of an optical connection line via a transverse opening in the middle of its pivoting part.
[92] In the light of the reasons outlined above, the Patent is novel with respect to the Shinkawa Patent as well.
Does the Patent involve an inventive step over the state of the art?
[93] Turning to the second main issue in this appeal (viz, whether the Patent involved an inventive step), unlike the inquiry for novelty considered in the preceding part of this judgment, it is permissible to construct a “mosaic” out of the various pieces of prior art in the inquiry for obviousness (ie, an inventive step) – unless the act of “mosaicing” itself is not obvious to the notional skilled person. This was made clear in the decision of this court in Peng Lian Trading Co v Contour Optik Inc and others [2003] 2 SLR(R) 560 (“Peng Lian Trading Co”) at [18], where the court cited with approval Lord Reid’s observations in the House of Lords decision of Technograph Printed Circuits Ltd v Mills & Rockley (Electronics) Ltd [1972] RPC 346 at 355 to the effect that the “mosaic” must be one which is “put together by an unimaginative man with no inventive capacity” (reference may also be made to the Singapore High Court decision of ASM Assembly Automation Ltd v Aurigin Technology Pte Ltd and others [2010] 1 SLR 1 at [41]). In other words, as pertinently observed in Law of Intellectual Property of Singapore (at para 30.1.50):
[94] In the present case, it would be fair to mosaic the various pieces of prior art outlined earlier for the purposes of identifying any inventive step contained in the Patent. The ASA Patent, the Matsushita Patent, the Shinkawa Patent and the AFC 800 are all obviously concerned with the business of handling wafer dies, either for the purposes of sorting or bonding. The notional skilled addressee would not find it too far-fetched to mosaic these pieces of prior art to form a considered view of the state of the art.
[95] If, however, in order to arrive at the subject-matter of a claim, it is necessary to make a mosaic of extracts from documents published over a period of time, taking suggestions from one with suggestions from others independent of it, there can be little doubt that the claim has inventive subject-matter.
[96] The point in the preceding paragraph, notwithstanding, even allowing for “mosaicing”, there is still the need for the state of the art to be considered as a whole for the purposes of locating whether any inventive step exists in the patent being considered (in this case, the Patent). To this end, a more global approach is appropriate and the pieces of prior art will therefore not be examined individually vis-à-vis the Patent. In any event, we pause to observe that even if the Patent was compared individually against each of the pieces of prior art, the result would have been (particularly having regard to our analysis in the context of the issue of novelty) an a fortiori one simply because (as we have seen) none of the individual patents relied upon by the Respondent was as comprehensive as a mosaic of extracts constructed from these very same patents. Indeed, as we have seen, some of these patents were, viewed from an individual perspective, clearly lacking what the Patent possessed (for example, in the context of optical inspection). Put simply, the “mosaicing” of the various extracts from the individual patents relied upon by the Respondent would, in fact, present the case for the Respondent in its strongest possible light.

The judgment continues past this page. Read the rest at eLitigation.

Mühlbauer AG v Manufacturing Integration Technology Ltd

Laws this judgment relies on

Grounds of decision

Facts (including a description of the Patent)

The decision below

The law on patentability

Introduction

Novelty

Inventive step

Patent construction

A second preliminary point relating to the role of expert evidence

The state of the art described

The ASA Patent and the AFC 800

The Matsushita Patent

The National Semiconductor Patent

The Shinkawa Patent

Is the Patent novel vis-à-vis the state of the art?

The ASA Patent and the AFC 800

The Matsushita Patent

The National Semiconductor Patent

The Shinkawa Patent

Does the Patent involve an inventive step over the state of the art?