T 1009/96 (Alcatel N.V.) 31-05-2001

Reasons for the Decision

1. The appeal is admissible.

2. Novelty

No objection has been raised under Article 100(a) EPC on the ground that the subject-matter of granted claim 1 is not patentable within the terms of Articles 52(1) and 54 EPC. The Board takes the same view that the claimed subject-matter is new, having regard to all the documents cited.

3. Inventive step

3.1. Closest prior art

3.1.1. In selecting the closest prior art, a document should be considered, which is directed to the same purpose or effect as the invention. This means that the document should relate to the same or a similar technical problem or, at least, to the same or a closely related technical field as that of the patent in suit (see T 989/93, not published in OJ EPO, point 12 of the reasons).

In the present case, the invention relates to an optical power amplifier with an active fibre (see the patent specification, column 1, lines 3 to 5). According to column 2, lines 43 to 47, the object of the invention is to provide such an amplifier "that has a high amplification efficiency in correspondence with relatively high pumping wavelengths, substantially more than 520 nm", the term "amplification efficiency" being defined in column 2, lines 2 to 4, as "the ratio between the power of the transmission signal in output with respect to the supplied pumping power". The Board notes that prior art documents (see, for example, D6 or D11), while describing the properties of an optical amplifier, often mention, besides other parameters like the gain or the pumping wavelength, the "pumping efficiency", which is commonly used for estimating the output signal power that can be obtained with the available pump power (for a given input signal). In the Board's view, the amplification efficiency K=Pout/Ppump, as defined in the patent in suit, and the pumping efficiency mentioned in the prior art documents refer to the same physical aspect concerning power conversion, although different terminologies are used. Therefore, the closest prior art should be chosen among the cited documents which refer to an optical power amplifier with high pumping efficiency at relatively long pumping wavelengths.

Document D6 relates to Erbium-doped fibre amplifiers operating at 1.5 µm, which are suitable for telecommunication applications, in particular as in-line repeaters, power and pre-amplifiers. D6 teaches that a pumping wavelength of 980 nm is particularly advantageous because it ensures greater pumping efficiencies (see page 207, Conclusions). Thus, document D6 is regarded as an appropriate starting point for assessing inventive step since it discloses an optical amplifier of the kind claimed and addresses the problem of having a high pumping efficiency.

3.1.2. The respondent cites T 686/91 and T 59/90 (supra) in support of his statement that none of the cited documents could represent the closest prior art.

In T 686/91 (see point 4 of the reasons), the board concluded that "a document not mentioning a technical problem that is at least related to that derivable from the patent specification, does not normally qualify as a description of the closest state of the art on the basis of which the inventive step is to be assessed, regardless of the number of technical features it may have in common with the subject-matter of the patent concerned." Thus, rather than supporting the respondent's statement, this decision leads to the opposite conclusion that D6 can be regarded as the closest prior art document, because, as already mentioned, it discloses an Erbium-doped fibre amplifier with a high pumping efficiency.

As to T 59/90 (see points 4-8 of the reasons), the board, after having identified the closest prior art, formulated the technical problem underlying the invention. The addressed problem was then considered not to be known from the closest prior art, and the solution to the problem was regarded as involving an inventive step. Thus, the closest prior art document was identified independently of whether or not the case dealt with a "problem invention".

3.2. Problem to be solved and solution

3.2.1. Document D6 discloses an optical power amplifier suitable for use in an optical fibre telecommunication line (see page 204, end of second paragraph, end of third paragraph, page 205, last paragraph, page 206, last paragraph, page 207, Conclusions, last sentence). The amplifier comprises an active doped fibre containing GeO2 as refractive index modifying dopant and Erbium as fluorescent dopant in the core (see page 208, Table 1, in particular line 2). It is supplied with pump light from one pumping laser through a dichroic coupler, the pumping wavelength being 980 nm (see page 204, last paragraph, Figure 1, Table 1, line 2). Moreover, since D6 specifies that the large signal gain saturates at a certain value, it is implicit that the power amplifier may operate in saturation conditions (see page 205, end of third paragraph).

Therefore, the subject-matter of granted claim 1 differs from the amplifier disclosed in D6 only in that Al2O3 is used as refractive index modifying dopant.

3.2.2. The use of Al2O3 as refractive index modifying dopant in combination with a high pumping wavelength is related to the problem of improving the pumping efficiency. This follows from the fact that the pumping efficiency is negatively affected by ESA (see D11, page 1084, left-hand column, first paragraph, and D25, page 403, right-hand column, lines 1, 2), whereby ESA is influenced by factors like the pumping wavelength and the fibre glass composition, as is well-known to the skilled person (see D25, page 403, right-hand column, second paragraph). Indeed, considering the energy level diagram of the Er3+ ion and the associated transitions (see D11, Figure 1), non-radiative decays to intermediate levels causing undue pumping energy losses may take place depending on the pumping wavelength, whereas the fibre glass composition determines the location of the energy levels and their broadening.

Document D6 addresses the problem of ESA, although it does not clearly indicate to which extent its negative effects can be avoided merely by using the 980 nm pumping band (see page 205, line 6, "relatively clear of ESA", and line 10, "entirely free of ESA"). In any case, D6, Table 1, second line, discloses the combination of 980 nm as pumping wavelength with SiO2/GeO2 as fibre type. According to D11, ESA can be completely avoided by using fibres containing Al2O3, instead of GeO2, for a pumping wavelength of 980 nm (see page 1085, Table 2, last line, page 1086, right-hand column, first paragraph). Therefore, following the teaching of D11, the pumping efficiency or, using the terminology of the patent in suit, the amplification efficiency of the power amplifier according to D6 can be further improved.

3.2.3. Summarising, starting from the power amplifier disclosed in D6, which is pumped at a wavelength of 980 nm, the skilled person would, in the Board's view, consider to use Al2O3 instead of GeO2 as refractive index modifying dopant because, according to document D11, no ESA at all would be present and a higher pumping efficiency could be achieved. Therefore, in combining D6 with D11, the skilled person would arrive at the subject-matter of granted claim 1 without any inventive activity.

3.2.4. The Board notes that high values of the amplification efficiency, which the respondent presents as an essential feature of the present invention, have already been achieved in the prior art. Document D22 (see page 455, Experiment and results, Figure 1) discloses an experimental configuration including an Er3+-doped fibre optical amplifier operating at 1.535 µm. The gain characteristics of the amplifier pumped with 15 mW at 980 nm is shown in Figure 2. For large input signals, when the amplifier is operated in saturation, a maximum saturated output power of 8.5 mW is achieved. As the appellant I points out in the letter of 1 August 1996, this gives an amplification efficiency K = Pout/Ppump = 8.5 mW/15 mW = 0.57, whereas the quantum efficiency is Eq = pump/ out = 980 nm/1535 nm = 0.64. Thus, K is 89% of Eq. This result is not far from that achieved by the patent in suit (see column 7, lines 37-39), which cannot, therefore, be considered as surprising.

As regards this calculation, the Board does not agree with the respondent's objection, that it reflects an ex post facto analysis. Indeed, it is legitimate to rely on the definition of amplification efficiency given in the patent in suit in order to compare the amplifier known from D22 with that of the invention.

3.3. The Board does not share the respondent's view that the idea itself of optimising the amplification efficiency K to approach the quantum efficiency Eq, while designing an optical fibre power amplifier, had not been known in the state of the art at the priority date of the invention. It is admitted that the mathematical definitions concerning K and Eq are not per se disclosed in the cited prior art documents. However, the physical concepts underlying these definitions can indeed be found in the cited documents, although under a different terminology (see above). As regards the theoretical maximum value of K, i.e. Eq, the skilled person can easily infer from the energy level diagram of the Er3+ ion, that, for the signal wavelength of 1.5 µm, the pumping band at 980 nm provides the highest value of Eq that can be achieved, when compared with the other possible pumping wavelengths of 532 nm, 670 nm and 807 nm mentioned in D6, page 204, last paragraph.

3.4. The respondent's argument that, at the priority date of the invention, operation of fibre-based optical power amplifiers had not yet been investigated in detail in real telecommunication systems, is not convincing. Indeed, the Board notes that both D6 and D11 make numerous references to application of Erbium-doped fibre amplifiers in telecommunications systems. The respondent also underlines that "permanent" operation of the claimed amplifier should be regarded as a distinguishing feature over the prior art. Such a feature, however, is not mentioned in the granted claim 1, and, therefore, does not need further consideration.

4. In conclusion, having regard to the patent documents according to the respondent's request, the ground of lack of inventive step set out in Article 100(a) EPC prejudices the maintenance of the European patent.