T 1430/15 - Perpetual motion devices and the GL

Key points

• This examination appeal is directed to an 'energy transducer' which is specified in claim 1 to be able to ‘generate useful work’. According to the Examiner (para. 1.1 of this Communication), it is a perpetual motion device. 
• “Claim 1 is directed to an energy transducer [...], which is able to generate useful work by lowering the internal energy of the material. Since the claimed transducer is limited to exhibiting this effect, the disclosure has to enable a skilled person to achieve it in order to meet the requirements of Article 83 EPC.”
• The Board then analyses the application as filed and the physics involved.
• “To summarise, the Board is not convinced that the model predictions [submitted by the applicant] are correct, there is no disclosure in the application as filed concerning the transduction of internal energy to useful work, merely a hypothesis to this effect, and there is no experimental evidence that the transduction from internal energy, as hypothesised in the application as filed, will necessarily occur when the claim prescriptions regarding the application of forces and choice of material are followed. For these reasons, the application in the version of the main request does not meet the requirements of Article 83 EPC.”
• This analysis is perhaps not very surprising. However, it's interesting to compare this case with the  - in my view rather cryptic - remark in the GL G-III.1 that “An objection could arise under Art. 57 only in so far as the claim specifies the intended function or purpose of the invention, but if, say, a perpetual motion machine is claimed merely as an article having a particular specified construction, then an objection is made under Art. 83”.
• I find this remark in the GL rather confusing. Say I have a claim directed to 'a stack of hexagonal copper plates and pentagonal gold plates' (i.e. merely as a device having this configuration) and the description says that it will generate net energy when exposed to some magnetic field. The skilled person can easily manufacture the stack of plates, still the rejection is under Art. 83 according to the cited GL passage. If on the other hand, the claim has additionally the functional feature ‘the stack is capable of generating net energy’, which is something the skilled person can not reduce to practice according to thermodynamics, the rejection is under Art.57, still according to the GL. The other way around would make much more sense to me (i.e. the impossible functional feature causes a problem under Article 83; merely specifying a structure of a device without plausible function give a problem under Art.57).
• T0541/96 explains that "An invention or an application for a patent for an alleged invention which would not comply with the generally accepted laws of physics would be incompatible with the requirements of Articles 57 and 83 because it cannot be used and therefore lacks industrial application. Also the description would be insufficient to the extent that the applicant would not be able to describe how it could be made to work." 
• GL F-III,3 state that: If the claims for such a machine are directed to its function, and not merely to its structure, an objection arises not only under Art. 83 but also under Art. 52(1) in that the invention is not "susceptible of industrial application". Implicitly, if the claims merely specify a structure, the objection is supposed to be under Art.83.
• In the auxiliary request, the functional feature was deleted. This is not accepted under Art.123(2), because “the application as filed consistently discloses that the transducer produces useful work by lowering the internal energy of a material”.

T 1430/15 -  link

Reasons for the Decision

1. The appeal is admissible.

2. Main request

2.1 The patent application does not meet the requirements of Article 83 EPC because it does not disclose the invention in a manner sufficiently clear and complete for it to be carried out by a skilled person.

2.2 Claim 1 is directed to an energy transducer containing a material with unequal cross coupling coefficients between first and second forces and corresponding energy conjugate physical properties, such as for example strain and magnetisation, which is able to generate useful work by lowering the internal energy of the material. Since the claimed transducer is limited to exhibiting this effect, the disclosure has to enable a skilled person to achieve it in order to meet the requirements of Article 83 EPC.

2.3 The application discloses model calculations which predict that during a full cycle of two forces applied to the sample in sine and cosine form, respectively, net work should be performed by the transducer if the cross coupling coefficients are assumed to be different from each other. The amount of work generated only depends on the difference in cross coupling coefficients. Stated differently, more energy per cycle could be produced by such a system than was put into the system per cycle.

2.4 Correctness of the model predictions

2.4.1 The Board is not convinced that the model predictions concerning generation of work by lowering the internal energy are correct. The Board had informed the appellant in a communication pursuant to Article 15(1) RPBA in point 1.3 that they considered unequal cross coupling coefficients to be rather exceptional and that "it appears from the description and the underlying assumptions made for deriving the constitutive equations that the coupling constants would be equal by definition". The Board stated that the cross coupling coefficients were derived by twice partially differentiating the Gibbs free energy.

2.4.2 In response to this, the appellant has adduced documents D1 and D4 to D9 to present evidence for the existence of materials which have different cross coupling coefficients and for the assertion that a skilled person was able to verify whether a given material had unequal cross coupling coefficients. The appellant also directed the Board to documents D10 to D12 in order to demonstrate the fact that in scientific literature the constitutive equations used by him were accepted to be valid under certain assumptions about the material and experimental conditions.

In view of this evidence, the Board accepts that materials with unequal cross coupling coefficients exist in nature and that a skilled person was able to measure whether these coefficients were different for a given material. The Board also accepts the correctness of the constitutive equations of the model used under the conditions under which they are derived in D10 to D12. For these reasons the aforementioned documents need not be identified in detail and their contents require no further discussion in this decision.

2.4.3 However, in attempting to address the Board's argument, the appellant fails to properly distinguish between whether materials with unequal cross coupling coefficients occur in nature and whether within the assumptions of the model calculations unequal cross coupling coefficients can be assumed without creating inherent contradictions. The appellant has not questioned the fact that the cross coupling coefficients were derived by twice partially differentiating a function (a thermodynamic potential, such as the Gibbs free energy) with respect to the same variables but in a different order. If the second partial derivatives of a function exist and are continuous, then the order of the differentiation can be shown to be irrelevant. Under these conditions the coefficients are equal by definition. The conditions apply to the appellant's model calculations because the appellant clearly treats the coefficients as constants in his calculation of the magnetic and mechanical energy because he takes them out from the respective integrals, see for example the equations on page 5, line 21 and page 6, line 7. If they are constant over the integration path, it follows that the second partial derivatives exist and that they are continuous on any point over which the integration runs. Therefore, the appellant's assumption of unequal coupling coefficients is in contradiction with the constitutive equations on which the model calculation is based. Within the framework of the model the coupling coefficients are equal by definition. It is noted that when accepting this, the model predicts that exactly the same amounts of conjugate forms of energy are transduced, which is entirely in line with expectations based on classical thermodynamics.

The appellant has not further addressed this point in his letter of reply dated 23 January 2020 or in the oral proceedings. Hence, the Board has no reason to deviate from its preliminary view that cross coupling coefficients are equal by definition within the model. Hence, the model appears to contain an inherent contradiction and its predictions can thus not be assumed to correctly describe a real physical system. This is not to say, that no materials with unequal cross coupling coefficients exist, but merely that the equations used in the appellant's model cannot correctly describe such materials.

2.5 No disclosure concerning conversion of internal energy

2.5.1 The disclosure of the application is silent on any details of the conversion of internal energy to useful work. It merely stipulates this effect. The appellant's model takes into account only two energetic contributions, namely on the one hand the mechanical and and on the other hand the electric or magnetic energy of the sample in an external electric or magnetic field. However, it does not take into account any details of the internal energy of the sample or any other mechanism allowing its conversion to work.

In these circumstances, the application could only be seen to contain an enabling disclosure if the generation of work by lowering the sample's internal energy was an inevitable consequence of applying forces to a material as prescribed by the claim. The application itself does not contain any experimental evidence in this respect.

2.5.2 The appellant has adduced document D2, which shows an experimental study of the transduction cycles on a Galfenol rod as support for the occurrence of this effect.

D2 uses a system which allows external stresses by a load frame and external magnetic fields by a magnetic field coil to be applied to a Galfenol sample such that the resulting magnetic induction B and strain epsilon can be controlled. Galfenol is an alloy of gallium and iron which exhibits magnetostriction and which, in previous research, had apparently been shown to have unequal cross coupling coefficients. Two closed cycles were measured around each of three different operating points, see page 5, right column middle paragraph. The mechanical and magnetic energy densities were calculated from the second of these cycles around operating points of (22.5 MPa, 0.32 T) and (22.5 MPa, 0.39 T). D2 reports that in two cases more mechanical energy is extracted from the sample than magnetic energy exerted on it, see page 7, left column, second paragraph.

The Board does not consider D2 sufficient evidence for the inevitable occurrence of the claimed effect because D2 does, in fact, not report any changes in the samples internal energy. The experiment is not designed to measure changes in internal energy. The internal energy is not normally directly observable, instead the experiment would have to be designed so as to measure all exchange of energy of the sample with the surroundings to conclusively detect a change in the internal energy. However, no experimental precautions are disclosed to have been taken which would allow it to be verified whether a change of the internal energy of the Galfenol rod corresponding to the observed anomalous excess energy could be observed. In particular, the authors of D2 hypothesise in the last paragraph of D2 that thermal fluctuations in the sample might account for the findings. If this is experimentally not ruled out, then the conclusion that a reduction of internal energy of the sample is the source of the anomalous energy gain has not been experimentally demonstrated in D2.

It is important to note that the subject of discussion is whether the application as filed discloses the invention in a manner sufficiently clear and complete for a skilled person to carry it out, not whether the reported findings of D2 are correct or incorrect. The correctness of the observation of excess energy in D2 would still not support the conclusion that the internal energy of the Galfenol was the origin of this energy for the above reasons.

The appellant has argued that the sentence concerning temperature fluctuations was to be understood in context to mean that due to the application of stress and magnetic field it may be possible that the sample cools down and that the additional energy flows to the sample in the form of heat from the surroundings.

This argument is not suitable to demonstrate that the invention is sufficiently disclosed in the application as filed. Given the details of the disclosure of the present application, it would have had to be demonstrated that the claimed effect of lowering the internal energy was an inevitable consequence of applying forces to a sample in the manner prescribed by the claim. The fact that tentative explanations as to the origin of the observed anomalous excess energy are needed merely demonstrates that D2 does not contain any experimental proof of the internal energy being the origin of the excess energy.

2.6 To summarise, the Board is not convinced that the model predictions are correct, there is no disclosure in the application as filed concerning the transduction of internal energy to useful work, merely a hypothesis to this effect, and there is no experimental evidence that the transduction from internal energy, as hypothesised in the application as filed, will necessarily occur when the claim prescriptions regarding the application of forces and choice of material are followed.

2.7 For these reasons, the application in the version of the main request does not meet the requirements of Article 83 EPC.