T 610/24: Starting from remote prior art risks ex post facto reasoning

The Board remitted an application concerning chemical pump compatibility management back to first instance for a new search. The initial refusal for lack of inventive step was based on prior art that pursued an entirely different objective and lacked the central technical feature of the invention. The Board emphasised that, when the examining division selects a starting point that is too remote from the claimed invention, the inventive step analysis risks becoming an exercise in hindsight rather than a realistic assessment of what the skilled person would have done.

Background

The application at issue (EP 3756153 A1) concerned a system for checking the chemical compatibility between a chemical composition and components of a chemical pump. The system comprised a chemical pump with at least one component exposed to the chemical composition being pumped, a chemical reservoir containing the chemical composition with an electronically readable media, a database containing chemical compatibility data including a compatibility between the chemical composition and the at least one component, and a portable computing device executing a software application. By reading the electronically readable media on both the pump and the chemical reservoir, querying the database for compatibility information, and outputting an indication of compatibility via the portable device, the system enabled an operator to verify chemical compatibility before supplying a chemical composition to a pump, thereby preventing damage to pump components and avoiding premature pump failure.

The Examining Division (ED) refused the application, considering it obvious over D5 (US 2010/0024915 A1). In their analysis, D5 taught a dispensing arrangement that checked compatibility between a chemical solution and a receiving container using radio frequency identification (RFID) tags and a database. They identified two distinguishing features: checking pump compatibility rather than container compatibility, and using a portable device. Both differences were deemed either obvious or mere design choices.

The Appeal

The Applicant (A) challenged the ED’s interpretation of D5, arguing that this document addressed an entirely different problem. According to A, D5 was concerned with preventing mislabelling errors by ensuring the correct chemical went into the correct container. This had nothing to do with analysing whether materials would be chemically degraded by contact with particular substances. 

A also questioned whether document D6, which the Board had cited regarding chemical compatibility issues with pumps, had been proven to be publicly available before the priority date.

Decision of the Board

The Board accepted A's position that D5 concerned preventing dispensing errors rather than material compatibility. D5 described a system for ensuring operators filled the right containers with the right chemicals to avoid mislabelling problems. The alleged compatibility check in D5 was simply matching a chemical to its intended destination, not analysing whether chemical reactions or material degradation would occur.

Moreover, the Board observed that D5 did not contain a pump at all. The apparatus used an eductor, which is a fluid-driven device that operates on completely different basis from a pump. While both devices move liquids, an eductor has no motor-driven mechanical parts. This meant that arriving at the claimed invention from D5 would require multiple additional steps: introducing a pump, recognising a material compatibility, and then adapting D5's identification technology to address this new problem.

The Board reinforced a vital concept for inventive step assessment. Although ED is not strictly bound to use the single closest piece of prior art, selecting a starting point with a similar purpose and technical effect is strongly advisable. When examiners begin from documents addressing different problems, the analysis often becomes distorted by hindsight. The "technical problem" formulated in such cases tends to be one the skilled person would never actually have conceived, because the circumstances prompting that problem do not exist in the chosen starting point.

The Board identified a more significant issue with the examination: the search had been inadequate. The search report listed only IPC class G06Q, which covers business and administrative data processing systems. This classification does not encompass pumping equipment or questions about chemical resistance of materials. A proper search should have covered classes like F04 (pumps for liquids) and should have investigated the technical field of chemical compatibility for pump components.

Since the relevant prior art landscape had not been properly established, the Board could not meaningfully assess inventive step. Before granting or finally refusing a patent, the deciding division must satisfy itself that the applicable prior art has been identified. The Board therefore exercised its discretion under Art. 111(1) EPC to remit the case, instructing the Examining Division to conduct a search covering the relevant technical fields, including liquid pumps and chemical compatibility considerations.

Summary written by the NLO EPO Case Law Team