This article reads like it was written by GPT. Bendable X-ray mirrors are an extremely common thing (using a conventional linear actuator to bend silicon, for example). The novelty here seems to be:
"Single-crystal lithium niobate is useful in X-ray technology because it can be expanded and contracted by an electric field and polished to make a highly reflective surface. This allows it to serve as both the actuator and the reflective surface, simplifying the device."
Pretty interesting kernel of info there, but it would be nice to know how fast they're able to make corrections, and how they're measuring the aberration. From the sparse info provided, they take an image, look at this '2' character, and tweak 'something' until it looks a little better. This suggests it's a slow process, so the advantage over conventional methods is how fine a correction you can make? And how much control over the surface do you have, is it just 1 dimensional?
Most articles today are written with LLMs. I found even entire web sites devoted to certain niche hobbies are LLM generated and ranked high on duck/google.
For the curious people, there are no practical refractive materials for x-rays, and mirror shapes have to be 4 times more accurate than equivalent refractive systems for the same diffraction limit.
Do you know what the objective limits are with current mirrors? They just say that nanometer objectives are not possible with current mirrors, but no measure of SOTA
"Single-crystal lithium niobate is useful in X-ray technology because it can be expanded and contracted by an electric field and polished to make a highly reflective surface. This allows it to serve as both the actuator and the reflective surface, simplifying the device."
Pretty interesting kernel of info there, but it would be nice to know how fast they're able to make corrections, and how they're measuring the aberration. From the sparse info provided, they take an image, look at this '2' character, and tweak 'something' until it looks a little better. This suggests it's a slow process, so the advantage over conventional methods is how fine a correction you can make? And how much control over the surface do you have, is it just 1 dimensional?