FridayAFM - Bismuth

Héctor here, your AFM expert at Nanosurf calling out for people to share their Friday afternoon experiments. Today I remove the oxide layer on Bismuth and perform electrical measurements on the exposed surface.

This fridayAFM closes a chapter on electrical measurements, hence I would like to start by recalling what has been done so far and what we learnt with it.

We started with the:

In that occasion, we performed current vs distance measurements to determine the oxide layer thickness. To disentangle the tunneling from the probe piercing through the oxide, we varied the voltage and fitted the trend to obtain the oxide layer thickness. It came about 307 nm for aluminum kitchen foil. We also highlighted how it is possible to see the formation of the water meniscus in the deflection data, but that is a separate topic.

Then, using a logarithmic amplifier, we repeated the experiment with aluminum kitchen foil, gold on mica, and ITO:

In this repetition of the experiment we spent more time discussing the limiting current and the origin of the different resistances in the circuit. We obtained a similar value for the oxide thickness in aluminum, and what is more interesting, we also provided values for the contact resistance between the probe and the surface. For the diamond probes used (here material and geometry matters), and the indentation considered, the contact resistance came about 20 kΩ.

With the accumulated knowledge, we then studied a galena crystal, which is an early semiconductor, and by playing with the voltage applied during the imaging, we managed to grow the oxide layer and create a diode:

We observed the diode effect by looking at the I-V curves on the "clean" surface and on the part that was oxidized. They were short-lived diodes because the voltage used to perform the I-V was degrading them, but if we had limited the voltage to a smaller range, they could function as diodes (maybe one day I will come back to that and see how many cycles they last using low voltages).

Doing an experiment of a different kind (focusing on microwaves this time), we test how good are these probes (i.e. the diamond ones from Adama) for digging onto samples:

We learnt that the 40 N/m probe and the 450 N/m probes can apply more than 10 μN, dig onto the sample surface, and still image with good resolution.

So, today, we will use a bit of everything. Our test sample is bismuth, which, when cooled down slowly tends to form geometrical patterns with nice colors. The thing is... bismuth real color is grey, it shows rainbow colors because it has an oxide layer on top, and the oxide thickness changes across its surface. As usual, the oxide layer is insulating, but you will not know if you test with a multimeter, because the force you apply when putting the probes down is enough to break the oxide barrier (as we discussed with the aluminum).

So... using bismuth and diamond probes and conductive measurements... can we dig the oxide layer and reveal the metal?

1 Bismuth sample mounted on the conductive sample holder. Electrical diagram for the conductive + force modulation AFM.

As you can see on the figure above, I obtained a piece of bismuth which was prepared to show the geometrical patterns and has a nice oxide layer all over it (thanks Etsy). I used the metallic clip on the C-AFM sample holder to contact it, and the clamp (covered in Kapton tape to prevent conduction) as additional support.

On the circuit diagram you can see that the bismuth is grounded and the tip applies a voltage and measures the current. On this occasion, while using static deflection for topography feedback (and while recording the current between the tip and the sample), I will be oscillating the piezo shaker. Why doing this? As long as we stay away from the cantilever harmonics (and subharmonics), the probe will not break the electrical connection with the sample, and thus the shaker will not affect the current measurements. So...? So we gain information about the subsurface features and without affecting the main measurement. (And in addition the lateral resolution improves, but this is a claim for which I don't have enough evidence).

1 C-AFM + force modulation images taken with a AD-450-AS diamond probe from Adama. Representative I-V curves from the area with oxide and from the exposed metal.

As you can see in the images of the figure above, the phase and amplitude were not a bad idea, they reveal more details about the oxide layer than the topography (or the deflection, as you will see later on). What about the current? Well, I was very lucky and landed next to a hole on the oxide layer. How I know it is a hole? Because the current saturated to the maximum measurable by the current meter inside of the AFM. And how do you know the rest is oxide? By doing I-V curves we can see that on the hole the contact is ohmic, with the I-V curve being a straight line over zero. On the oxide, on the other hand, the contact is Schottky type, it requires certain voltage to start conducting.

Can we corroborate this by... for instance removing more of the oxide? See the figure below.

1 C-AFM + force modulation images taken before/after digging a hole in the oxide layer.

Now, while the hole we just dig shows the exposed metal, it is too deep (sorry, I used too much force). For this reason, I zoomed out and repeated the digging process. You can see consecutive scans in the figure below (where for a change, I show you the deflection channel).

1 Digging using the diamond probe with a large setpoint. Consecutive deflection images.

This below is how the area looks like after the second digging.

1 C-AFM + force modulation images taken after the second digging. Note the details hidden to the topography but revealed by force modulation, some of which will be visible on the current image.

Note that while deflection highlights the topographic details (e.g. scratches), there are things that are only revealed in the force modulation channels. Looking at the phase and amplitude it is possible to see for instance contrast in the oxide layers. This contrast, as you can see in the figure below, matches some of the features of the current channel, hence either they are related with the oxide thickness the degree of oxidation, or contamination.

1 The second digging was more gentle, and is just barely enough to reveal the metal, giving us a better estimation of the oxide layer thickness.

Now, looking at the current channel, we can see which areas conduct and which parts are still covered with oxide. Using this information, it is possible to look at the topography and determine how thick is the oxide (and also notice that I used too much force for the first dig). We see that at about 149 nm there is conduction, but we also see that an exposed area (top of the image), with 50 nm less of oxide is not conducting, hence the oxide thickness must be between these two values. Not the most precise estimation, but at least it gives us some reference and it is all I had time to measure this week.

Let's recap. Using C-AFM (combined with force modulation) we explored a bismuth sample with a thin oxide layer on top. We used a diamond probe for the electrical measurements and also for digging onto the oxide layer.  We estimated the oxide layer thickness for this particular area to be between 50 and 149 nm. 

I hope you find this useful, entertaining, and try it yourselves. Please let me know if you use some of this, and as usual, if you have suggestions or requests, don't hesitate to contact me.