Scanning Tunneling Microscopy

The techniques that I work with are collectively called scanning probe microscopy (SPM), which encompasses scanning tunneling microscopy (STM), atomic force microscopy (AFM) and their force/current spectroscopy counterparts. These are incredibly powerful methods that allow me to look at structures down to atoms. The machines that I work with are housed in ultra-high vacuum (UHV) chamber, which we pump down to an incredible vacuum of only 0.0000000001% of atmospheric pressure: a more perfect vacuum than what you will find near the international space station. We typically conduct measurements at a chilly 4.5 degrees Celsius above absolute zero: a temperature needed to "freeze" our molecules and nanoribbons to the surfaces we are probing, and also keep electrons from "sloshing around" too much, instead keeping them in well-defined (and sometimes highly correlated) states.

Performing research with such instruments is not always straightforward, and when it breaks down, you need to fix it. I am happy to report that the numerous times of hardship in the lab have turned me into an expert at fixing and engineering STMs and other ultra-high vacuum equipment!

The scanning tunneling microscope (STM) works by scanning an extremely sharp tip over a conductive surface. Due to the quantum mechanical tunnel effect, electrons can "jump" from the last couple of atoms of the tip to the surface, through the vacuum. This happens when the vacuum gap is of the order of only one or a few atoms.

In STM, the application of a finite bias voltage between tip and sample, combined with the constant tunneling of electrons through the gap, results in a steady flow of electrons from tip to sample or vice versa: a current. Typically, we maintain a bias voltage of approximately 1 or 2 volt, and maintain a tiny tunneling current of the order of just several picoampere or nanoampere.

When we perform STM in current feedback mode, it means that we keep the tip-sample distance approximately the same by actively retracting the tip whenever the current increases beyond a certain setpoint, and moving it closer to the surface when the current drops below. When combined with horizontally moving over the surface, we can then actively trace the height profile of the surface, down to the atom.

The simulations below show the effects of feedback settings and loop gain on STM scanning. These simulations were made in Mathematica, with the aid of MathemaTB. Here, the tip is moved over the surface and the current is calculated by summing over the negative exponentials of the distances of the lowest tip atom to all atoms on the surface. The black dots represent the trace of the tip height, while the vertical position of the blue dots represents the magnitude of the tunneling current.

Since the feedback loop comprises a PID-controller, the proportional and integral gain should be set up correctly to get the tip to behave correctly without crashing into the surface or bouncing.

The effect of lowering the bias voltage is similar to that of increasing the current setpoint: it brings the tip closer to the surface. Scanning close to the surface enables features to be resolved more sharply, but it is also easier for the tip to crash into something.

Setting a larger bias voltage or lower current setpoint results in the tip tracing out the surface at a safer distance. The drawback is that the features are not as well-resolved anymore.

Electronic structure analysis

The bias voltage between tip and sample is the primary parameter that we use to perform electronic structure characterization of our materials. The energy levels or bands within a material can, hand-wavingly, be understood as electron tunneling pathways, and moving the bias voltage past them causes a more-or-less step-wise increase of the tunneling current as electrons have sufficient energy to tunnel into these levels. I customarily use a technique called lock-in amplification to efficiently measure such steps in the electronic structure of surfaces and materials in STM.

STM tunnel junctions and their behavior in terms of their I(V) transfer function can be understood in physical terms through models like the Tersoff-Hamann theory (possibly refined by correlation effects such as the Markus Ternes theory). Alternatively, one can look at the junction as an effective electronics circuit.

Recently, I have implemented multifrequency and intermodulation spectroscopy to further boost our characterization of materials, using the multifrequency lockin amplifier from Intermodulation Products. I have written the Python program Scantelligent to perform advanced spectroscopic measurements. The so-called "Dirac comb" that you see in the FFT of the tunnel current signal is exactly what we are after in multifrequency lock-in amplification.

Our hyperspectral and hyper-parameter analysis as enabled by these techniques are important because we are interested in learning how our materials behave under the influence of external parameters like laser irradiation, gate voltage, or magnetic fields. We use our advanced techniques as a tool in our exploration of this vast "parameter space" that describes the properties of our nanomaterials. While we are currently focusing on defects, excitons and trions, our specific aim is to wield optical excitations to manipulate our structures and potentially use them as qubits.

Maintenance

The life of an experimentalist is not always as shiny and glamorous as it seems. All that glitters is not gold (except for our surfaces, those are actually pure gold). Doing beautiful experiments, making great new discoveries and publishing in high-impact journals is only possible when everything is working well. But a scanning tunneling microscope is a complicated machine, as is all its peripheral machinery. So a significant portion of the work of a scientist comprises troubleshooting, repairing, maintaining and engineering. The images below give an impression of the repairing/engineering work I have been doing.

Repairing a broken sample heater

Fixing a hydrogen cracker

Repairing a molecule evaporator

Putting hair-thin wires back in place

Troubleshooting problems with the sputter gun electronics

Repairing a sample storage unit

Repairing a broken sputter gun

Soldering tiny wires in the STM stage

Repairing the piezoelectric stepper motor in the STM stage

Building a dual-wavelength laser setup

Fixing a sample transfer mechanism

Replacing a heater arm with a Peter arm

Fixing the STM

Repairing a transimpedance amplifier

Brute force removal of a galled screw

Installing a new homebuilt molecule evaporator

Life in the lab

The following pictures give a little impression of my life in the STM lab. Photo credits and credits for extra dramatic effect by switching off the lights to Christopher Mercado.

Google Sites

Report abuse