1: Does NEG work at very low temperatures only or does it work at ambient temperatures as well?

For the NEG effect to work, we need the following conditions being fulfilled: (a) All geometric features of our NEG-samples need to be large as compared to the de Broglie wavelength of the conduction electrons (which is around 1-10 nm. So that the electrons can be regarded as quasi-ballistic particles); (b) the mean free path of the conduction electrons needs to be comparable to the feature sizes of our samples (i.e. the mean free path length needs to be as long as or not much shorter than the feature sizes of our samples. So that our samples need to be roughly as small as the mean free path length of the electrons. If these conditions are fulfilled, then it can even be shown mathematically (solving the Liouville-equation) that the NEG-effect will inevitably occur. Our chosen material system AlGaAs/GaAs shows mean free electron path ways around 0,5 to 1 µm at low temperatures (< 50 K) only. At higher temperatures the mean free path length is reduced significantly due to electron collisions with phonons etc. That is why our AlGaAs/GaAs NEG-samples only operate reliably at temperatures < 50 K. At those temperatures we could enjoy minimum feature sizes of our NEG-elements to the tune of 500 nm, which makes fabrication of the geometries which we need much easier than smaller sizes. Other materials such as InGaAs/InP have been reported to exhibit mean free path lengths to the tune of 150-200 nm at ambient temperatures. That is why we have manufactured our second generation of NEG-elements using InGaAS/InP with minimum geometric feature sizes to the tune of 100 nm. For those NEG-elements we demonstrated the NEG-effect as well, however, although the effect is less pronounced because electron scattering increases significantly at room temperature. Therefore: Yes, if the right materials are chosen and the above mentioned conditions are met, the NEG-effect will occur also at ambient room temperature . A scientific paper with all our measurement data on this topic is expected to be published within the next few weeks.

2: How come that AI tells me, that it is not possible to draw electricity and cool the environment at the same time?

Large language models (AI) repeat and recombine information found in the published literature on which their responses are based. Consequently, when a scientific topic challenges widely accepted assumptions or is only sparsely represented in the literature, AI systems tend to reproduce the prevailing scientific consensus rather than critically evaluating new experimental evidence. As "countless publications have been written about the second law of thermodynamics, the impossibility of a perpetuum mobile of second order or the limitations of the Carnot cycle, an AI system that is not being advised to read and recombine the right kinds of literature, will mislead you. Before the invention of the wheel, an AI system trained exclusively on the knowledge available at that time would most likely have concluded that such a concept was impossible because no evidence or prior examples existed from which to infer it. AI is an excellent tool for summarizing established knowledge, but it is not a substitute for experimental validation when evaluating genuinely novel scientific concepts.

Concerning NEG and Artificial Intelligence

Just for curiosity:

What is wrong in the answers of AI to the first and second round of questions? AI does not realize, that NEG - due to its interaction with the boundaries that are not all exclusively specular - does not constitute a closed system in phase space in equilibrium but it constitutes an open system in stationary or dynamic conditions. That is why invoking the second law of thermodynamics is not appropriate.

3: Will it be possible to scale up to W, KW or MW by multiplication of NEG-Elements tiny output, which we see from one single NEG-Element which is to the tune of fW?

Yes, it has been demonstrated and is now being demonstrated on large scale. Comment: It is true that several effects in the nano-world have not been suitable for upscaling to the macroscopic world because the aggregation of many of those nano-devices by itself influenced the effect to all participating nano-devices. In the case of NEG, however, we are not talking about nano-devices (smaller than the deBroglie-wave-length of electrons) but of mesoscopic devices well in the region of 1/10 µm. Seen from the outside a NEG-element acts like a small battery of 1/10 µm size. If you couple two sich "batteries" in series, of course you double the voltage at the gates of this couple.We have tested that with 2, with 3 and with 5 NEG-elements and the voltage at the end-gates multiplid accordingly and in case of an outside resistor closing the gates the current and thus the electric powere behaved like batteries would have behaved as well. We are presently working on a chip that carries around 100 NEG-elements and we expect to show in September/October 2026 that the electric power multiplies accordingly. By then the questions - whether the power can be scaled up, what ergodicity means to NEG the second law of thermodynamics will be no questions any more.