Scanning Tunneling Microscope design

Reply

replied on Tue, Mar 6 2012 3:14 PM

rated by 0 users

First off you may want to read a bit more before you start to fire up the soldering iron but here we go:

Q: First, I know that the tunneling current will be very, very small; however, I'm not exactly sure what to expect. What values are typical of the tunneling current (prior to amplification)? What minimum and maximum current should I expect?

A: A few nA down to maybe tens of picoAmps with maybe 50mV to 1V bias would be a "typical" scenario. So 10E7 to 10E8 amplification should get you there.

Q: Second, I was wondering what's the best method of making Pt-Ir tips that are sharp enough to resolve images with atomic resolution? I've read that Pt-Ir tips can be electrochemically etched or mechanically cut at very steep angle from Pt-Ir wire. How effective/reliable at making sufficiently sharp tip is cutting the wire? How does that compare to the ease/difficulty and cost ot electrochemically etching Pt-Ir tips?

A: I would cut them with sharp cutters at an angle. The reason why you get high resolution on (flat) samples is that the tunneling current depends exponentially on the distance, so even a seemingly dull tip will have som atom up front. Etching help if you have samples that have some pronounced topograpphic features as you get less convolution effects.

Q: Also, I haven't yet decided on a design for the scanning system; however, I was thinking of useing a piezo tube scanner. One thing that I noticed about tube scanners is that, because they "deflect" or bend to achieve the x and y displacements, the tip would be progressivly tilted more and more as x and y displacements increase. Would that, no doubt, slight tilt be problematic? That is, would it effect the tunneling current at all?

A: Tube scanner are a good, straightforward to build choice. Yes, a tube has mechanical x-talk but that is actually quite small and for the scan ranges you intend to adress <100nm not an issue at all. You can quite easily estimate that x-talk using geometry. If your tip would change the tip-sample distance your z-feedback would compensate by chaning the "length" of the piezo.

Q: Additionally, I've found plenty of information on piezo electric actuators and was wondering if closed-loop operation of the positioning system is recommendable? From what I've found online, the feedback electronics which correct positioning errors useing data from some type of position sensor (capacitive and piezo resistive strain gauge sensors being the only ones which can obtain the nessacary precision, although I've been told that piezo resistive sensors are prone to thermal drift) introduce additional signal noise into the piezo's drive signals, which effectivly reduces the scanning resolution. Is closed-loop operation a good idea or would I be better off just using an open-loop?

A: Open loop is a great start and for the small ranges you intend sufficient. Once you have everything together you can always build version 2.0 using closed-loop.

Q: Regarding position sensors, I also had an idea which I'd apreciate some one else's opinion on. My idea is that I could place capacitive position sensors on the microscope and use thier positioning data to record the actual position of the tip/scanning head in space, without the feedback loop. That way I'd have the actual position of the tip when each data point (z hieght/tunneling current) is taken, while retaining the resolution of open-loop operation. While non-linearity errors wouldn't be corrected during the scan, I'd have the actual position of the tip for each data point : instead of simply assuming that the tip's position is always following the correct, linear voltage/displacment relationship. For example, if a small area of the sample missed due to some non-repeatable positioning error, the actual locations of the data points which would other wise have been assumed to correspond to that small area which was skipped over when the tip was breifly misaligned could be related to thier actual locaions (likely very near by). A subsequent scan, in which the previously missed area is then correctly scanned, the data for that area would be correctly shown (for the first time) instead of showing an aparent "change" in the local surface topography when the two scanns are compared. This would allow the higher scanning resolution to be maintained (since the actuators are being operated in an open-loop) while aquireing the acuracy/repeatability of closed-loop operation.

A: You seem to have intersting ideas but I would advise you to do two things: i) start simple. Build an open loop STM and get some atomic resolution. II) Do some calculations on how accurate you have to be in order to achieve what you are proposing taking also into account that your STM will show some drift, that is relative movement of sample and tip due to differences in thermal expansion.

Q: The book which I've been reading describes the feedback circuit for constant-current mode. In the book's explanation, it mentions a "proportional amplifier." I'm not familiar with this type of amplifier, can anyone tell me what it is? I've searched online for an explanation of what it is but found nothing - the closest matching results of a google search are sites which are selling circuit boards which they're reffering to as proportional amplifiers. They don't give any real description of what they are but they all seem to indicate that thier products are some how related to solenoids and servo motors. The book shows a functional block diagram of an STM's control circuits. In that diagram, the tunneling current is first amplified, then the amplified signal goes to another amplifier before going to the controling computer. That second amplifier is in parallel with an logarithmic amplifier which has a switch in parallel with it, connecting the input to the out put (the feedback loop maybe?) from the logarithmic amplifier, the book shows the current going to a comparator (it refers to it as a "Comp amp") where its presumably compared to a referance for the desired constant tunneling current (post-amplification and linearization). I find the use of a comparator odd however, since the only comparators I've seen have two output levels - high and low - which wouldn't be enough information to determine how much higher or lower the current is and correct the height of the tip to compensate, would it? After the comparator, the diagram shows a block labeld "sample/hold." It shows two outputs from there, one to the computer (the connection is labled "DAC5" - likely refering to a digital to analog converter generating the feedback/offset/correction signal), the other goes to two amplifiers in parallel with eachother. One is and integrator, the other is the "proportional amplifier." the outputs of those amplifiers are connected, then sent to two places. One is the cmputer ("Tip position ADC" - measurement of the correction/offset/feedback signal for height determination) the other is a voltage amplifier which is driving the z piezo. The part that I'm confused by is the integrator and proportional amplifier. Can anyone explain this? The book's explanation seems to assume that the reader knows more about the circuits (specifically the proportional amplifier, comparator and sample-and-hold circuit) than I do.

A: The first thing you want to do is to amplify the tunneling current. An inverting amplifier can be used here with a high enough feedback gain. After that you can use a logarithmic amp to account for the exponential distance depende of the tunneling current. You will then enter your "control" circuit wich adjust the voltage to the z-piezo to keep a user selected "setpoint" value. Here a simple PI-circuit will suffice. The output of that PI circuit is then fed to a (High Voltage) amplifier to drive the z-piezo. The voltage you need here depends on the type and the dimensions of the ceramic you choose.

I hope that helps,

Stefan