Speeding Up STM Imaging

Roland Piquepaille writes "Probably not many of you have used a scanning tunneling microscope (STM), the essential tool of nanoscience. And you might think that it's as easy to take a picture of an atom with an STM as it is to take a shot with your digital camera. In fact, the imaging of individual atoms with an STM is quite slow. Now researchers at Cornell University have shown how to accelerate this process — by adding a radio transmitter, they are able to speed up atomic-level microscopy by a factor of at least 100. A typical STM currently has a sampling rate of about one KHz. This new radio-frequency STM can operate a thousand times faster."

Sampling rate is limiting factor?

[Anonymous Coward]

I thought the limiting factor of SPMs (including STMs) is the feedback loop: one has to keep the probe tip from crashing into the surface as it's dragged back and forth, which means that the scan has to be slow enough that the piezo stack that's moving the probe tip up and down can do its job (limited by speed of sound through the material), as well as the electronics that have to decide how to move the thing in the first place. This might help with the electronics, but 1000x speedup in sampling rate doesn't mean 1000x speedup in imaging speed.

Re:Sampling rate is limiting factor?

[ndg123]

On atomically flat surfaces with small scan areas, you can scan in constant height mode (rather than constant current, where the tunneling current is the input to the feedback loop to adjust the probe height ). Still, a 400x400 point image of a 20 x 20 nm area still used to take a couple of minutes. Not 1/1000 second.

Re:Sampling rate is limiting factor?

[jibster]

The speed a piezo stack responds is related to the speed of sound but not in the way you think.

Each active element of the piezo receives the electrical signal to expand\contract at the speed of electricity through the material. This is usually very close to the speed of light. So the entire stack basically gets the signal move in parallel.

At that point we require a mechanical movement but since we are typically asking it to change by about 1nm/s this doesn't take a long time to do.

One day the response time of the material will become the limiting factor but right now its collecting the electrical signal.

On a separate note, we built STMs all the time in uni. We had a Russian genius who could do amazing things with no budget. He had a technique for making STM tips just by cutting a wire. We got better results from those tips than any of the commercial tips or the techniques published at the time (KOH, drawing etc.).

Re:I must be missing something here..

[IWannaBeAnAC]

In an STM machine, there is a single tip that moves over the sample. The sampling is done one pixel at a time, in much the same fashion as the beam of a CRT for example. So 1kHz is rather slow; for your 3 megapixel digital camera it works out at 3000 seconds (almost 1 hour) per frame. So a 1000x increase in speed is really significant!

Re:Huh?

[Anonymous Coward]

AFMs are SPMs, as are STMs. But AFMs work on the principle of van der Waals forces, not electrical current, so this wouldn't apply to them. Most AFMs on the market today are actually general-purpose SPMs that can work in AFM mode, with a plethora of optional modes (including STM... you can find an incomplete list of modes here http://en.wikipedia.org/wiki/Scanning_probe_microscopy [wikipedia.org].