The main way of driving the noise down is to use ever higher field gradients. A normal medical MRI uses a field gradient much less than 1 tesla/meter, but by using teeny magnets, they can get about 5 megateslas/meter, or about 50 Gauss/nanometer. It's still not quite good enough, though, and apparently they are not sure why they're not getting better field gradients. The gradients are off from their calculations by about an order of magnitude, for unknown reasons.
He did show some nice images of single tobacco mosaic virus particles:
But the 3D scan shown here contains about 8000 points, at 1 minute per point, so it required about 135 hours to collect! His comment about the time, though, was interesting: if you're talking about getting full scale 3D images of every single atom in an object with atomic resolution, and the ability to resolve chemical species, then the current technology you're competing against, X-ray crystallography, has a time scale of several months to infinity. So, 135 hours isn't so long to wait.
Speaking of Gauss: Steve Quake is now apparently blogging for the New York Times as a guest blogger. Check it out. His first column described the life of an academic researcher for those outside of the field, and his next column explored the divide between pure and applied sciences, in which he compares himself to Archmides, Gauss, and Lord Kelvin.
On the home front, I've been recently thinking about giving it all up and opening a sandwich stand. Or, perhaps, when I finish here, getting a degree in ECE from Stanford. Honestly, I think I'll do just about anything to stay out of the workforce until at least 2012 at this point.
0 comments:
Post a Comment