Australian Overturns 15 Years of Nano-Science Doctrine

Roland Piquepaille writes "Dr John Sader, from the University of Melbourne, discovered a design flaw in a key component of the Atomic Force Microscope (AFM). He 'used established mechanical principles to prove that the popular V-shaped cantilever inadvertently degrades the performance of the instrument, and delivers none of its intended benefits.' This finding may reshape the industry by proposing a single new standard and because the AFM 'has been the instrument of choice for three dimensional measurements at the atomic scale, since its invention in 1986.' Check this column for more details and an AFM diagram or read the original University of Melbourne's article. You also can visit the 'How AFM works' page."

Comments

Great.......but now what?

[Omikr0n]

Okay I just read the articles and as much as I'd like to understand it, most of it just seems way over my head. But from the minimal information I can understand, it seems that such a flaw shouldn't have been overlooked for this many years. Why did we just discover it now? Perhaps it will become more clear for me after next semester.

In the mean time, can someone possibly provide examples of any popular theories or situations that this discovery may have thrown off? I just want something more substance than "it changed a lot".

Re:Great.......but now what?

[Thurn und Taxis]

The cantilever arms, which are what differ between the V-shaped and the straight-beam cantilever arms, have characteristic dimensions on the scale of micrometers. That's six orders of magnitude larger than the atomic scale, so classical mechanical principles work just fine.

I don't have access to the paper yet, but I think the difference is fairly intuitive. To twist the tip of a V-shaped cantilever, you mostly just have to bend the center of one arm upward and the center of the other arm downward. To twist the tip of a straight-beam cantilever, though, you have to twist the whole beam. Most thin beams will bend much more easily than they'll twist (try it with a twig), so the V-shaped cantilever will twist more easily. Pretty intuitive, really, once you know the answer.

I wonder how much of a difference this really makes in the measurements, though, and whether the V-shaped cantilevers have other advantages that counteract this torsion problem. Newer AFMs use quadrature photodiodes, so it should be possible to measure the torsion of the tip and find out.

Re:Great.......but now what?

[tfoss]

AFM is basically dragging a pointer over a surface, and using a laser and fancy equipment to measure how much the pointer moves up down. This up and down motion is an indication of the height of the surface. In a way, it is very much like using your fingers to read braille. You run your fingers over a surface, and where the dots are raised, your nerves notice it, fire and you feel height.

With AFM, the finger is a little beam with a probe (often times a carbon nanotube) hanging down, running along the surface. On the top of the beam there is a mirror that reflects a laser beam onto a detector. As the surface height increases, the tip moves up, forcing the beam to flex just a little bit. This flex changes the mirror and thus the laser beam reflects to a different part of the detector. Raster scan a sample, and you get an x,y, and now z (height) value, so you have a 3d image of the sample.

If I read this correctly, the discovery is that the shape of beam that holds the tip, which is currently a V shape, works better when it is flat. The V-shape makes a beam stronger, and less likely to twist...or at least it was thought to. Intuitively, this makes sense. Fold a rectangular piece of paper into a V along the long axis. It seems stronger and more stable than if you just hold the unfolded paper out. Apparently, though, this is not the case with AFM cantilevers. Why this is the case is not mentioned, nor do I have any idea.

The reason this was not discovered is likely many reasons. First, it is obvious that a V-shape is stronger and more stable. That this is an incorrect assumption was probably not really even considered. It's as if you were building a computer, and everyone knows that a faster processor makes a faster computer. So you use the fasted one you can find. Except, in this certain circumstance a slower processor works better.

As for the effect of this, it really likely does not invalidate many experiments. It is a technical issue, not a new theory. It just means that you were not getting as much information as you could have from your machine.

-Ted

Ahh

[Timesprout]

But if an Australian overturns something does that not mean its actually the right way up ?

Godamnit :/

[dark-br]

Ill be sending my refund form right now!

Good thing?

[JohnnyKlunk]

Since the Aussie police have raided all the Universities and removed MP3/DivX collections they've had to turn their attention to work.
Hope noone at my company realises this.

Well-known

[Bowling Moses]

I don't do AFM, but my labmate has. He said that this flaw was well-known, and that most people dumped the v-shaped cantilevers in favor of nanotubes (I think) or straight cantilevers. Cool thing he said was that to get a tip, very popular was the gunk that piles up after you clean an electron microscope. One man's trash is another's treasure, I suppose.

Re:Well-known

[brarrr]

Thats not really true...

(I just tried to access the april issue of review of scientific instruments and it is not yet online, so I don't know the math behind his findings)

But no, the flaw is not well known, and no, most people haven't dumped v-shaped for nanotubes, you're confusing a few things.

One measurement technique in AFMs involves attaching a carbon nanotube to the tip of a cantilever (a v-shaped one, as thats what is available). This gives much greater resolutions (tube diameter is ~10nm) vs tip of cantilever diameter ~25nm. HOWEVER, when you do that, you can only scan very slowly, and cannot scan surfaces with steep topographies. Otherwise the nanotubes will just knock off the tip of the cantilevers.

Also, getting the tube on the tip is a hit or miss process, and rarely repeatable with the same length/angle/etc - and usually held on using electrostatic forces.

I haven't read anything about AFMs in a year or so, but this is what I remember from when I was involved with them.

Now I'm on to bigger things (ducks)

Most nano-science won't work anyway...

[Krapangor]

If modern string theory is true then most nano science applications will fail to work.
Recall that standard supersymmetry work with strings in 11 space dimensions on Yang-Calbai manifolds. At sizes below 15 angstroem you'll effects from these 11 dimensions. Especially has the wave equation non-trivial, non-analytic solutions and Hygens' principle fails (due to the topology of the Yang-Calbai manifolds, recall that the 5th deRham cohomology group is non-trivial).
So you'll get the effect of string resonance - strings are coupled together the 3rd order Laplace equation which overrules strong and weak interaction. This means that control of dynamical systems below the 15 angstroem barrier is impossible - you'll always get 5th order resonance which collapses the control Lie-algebra.
So all these nifty little nano-machines won't work, they'll be just little protein blob wiggling around and doing nothing useful.
As an example see this example.

woosh

[Cyno01]

*passes hand over head*

Now it needs to be proven empirically

[cerulean]

It's very intriguing that a mathematician has been able to mathematically prove that V-shaped cantilevers are worse for Atomic Force Microscopy.If the proof is so conclusive, however, it would have been nice for them to wait until they'd fabricated some straight-beam cantilever AFM tips, so that they could do a nice thorough study proving that they get better performance using them for actual data.

(It shouldn't be any more difficult, and it might be a little easier, even, to make straight beam cantilever tips than to make V-shaped ones. This is because the cantilever part of the tip is typically made by some sort of photochemical etching, and a straight beam is certainly a simpler shape to etch.)

Anyway, even with recent advancements in tip design technology atomic force microscopy is still rather inexact when it comes to getting good results consistently. As much as they try to design good tips, you'll never really know if you'll get good images from it until you mount it in the AFM and actually use it. I've certainly heard of grad students who will find a good tip (through trial and error) and become very protective of it (which is hard to do because they're extremely delicate), just because getting good results from Atomic Force Microscopy can often be tricky business, and a tip that you know is good is a great advantage.

Re:Making an AFM microscope shouldn't be that hard

[cerulean]

The tips are very delicate, and so far, they only seem to be made by photochemical etching. This is in contrast to Scanning Tunneling Electron micrscopes, which you can (and people do) make out of reasonably cheap parts.

This is because an STM tip can just be a pointy piece of wire, snipped off with pliers, and still give decent results some of the time. Also, there are easy techniques for making sharper STM tips yourself, such as electrochemical etching, which in this case is a very simple, easy-to-do-at-home process.

What's the problem with twist?

[photonic]

I happened to play with an AFM for an introductory lab course some years ago. What i remember is that by bouncing laser-light of the tip onto a 4-quadrant detector you could detect both the deflection and the twist of the tip. By scanning the tip sideways you get a twist depending on the local 'sticky-ness' of the sample, which could give some extra information about the sample.

Does somebody know why twist is a problem? I tried to look up the RevSciInstr article, but couldn't find it.

Yeah... but they work

[helix_r]

V-shaped cantilevers work fine. People can obtain atomic resolution with them. What more could you want?

I have used both straight and V-shaped. If there is a difference in performance, the difference is mostly likely very small and over-shadowed by other factors.

Re:Making an AFM microscope shouldn't be that hard

[mOdQuArK!]

I heard that they when they started making ATMs, the first tips were "made" by smashing a diamond between two plates of steel, then trying as many of the resultant crystal fragments as possible to see which one gave the best resolution. They "estimated" that the ones which gave the best resolution had a tip with the sharpness of a single atom.

Not Just Incorrect Measurements

[agg123456789]

AFMs are being used to do alot more than measure nowadays. This summer I worked on Dip-Pen Nanolithography which uses an AFM like a fountian pen of sorts. It's pretty cool stuff, and if that cantilever is off (the piece which holds the "nub" of the pen) then all of the work done could be rendered incorrect... DPN Information

Re:Making an AFM microscope shouldn't be that hard

[today]

I've been working on the software for these types of instruments since 1991. Making something that resolves atoms at room temperature is quite a daunting task. In electronics, just the basic Johnson Noise of resistors becomes significant when trying to resolve such tiny measurements. On top of that, the thermal drift of the metal in your instrument which moves your measuring device relative to what your measuring is enough to prevent you from seeing atoms. Then you also have to worry about digital noise generated by your processors radiating into the sensor electronics over ground and power leads.

To make a commerically viable AFM, you need a lot of smart people from several different fields. But even then, these people have to have a few years of building this sort of instrumentation under their belt. It is not easy at all. And the machining costs alone will always dictate a high price for these instruments.

-todd-

PS - Although atoms get a lot of press, I think the most interesting uses of AFM are in biology and hard drive research. These certainly produce the more spectacular looking images.