Hello this space is for my work related to a physics lab, nothing serious here, it's just for record.

2025/2/27

Thought about substrate holder, if still the substrate can be directly mounted on the holder plate, but no current flow in the plate if the plate is cut in a certain way. Cut out some slits or a central hole. Not very enlightening. Then checked several companies, Ferrovac and Scientia Omicron, search for "direct heating sample holder." Except for knowing we can cut a groove for better contact, not very interesting. Went back to earlier thoughts about sliding in a metal block, but I always dislike the idea of needing mechanical operation. Went back to the idea of a bimetallic strip, but still worried about the scale of deformation and the sudden pressure causing sample break. In summary this is still quite unhopeful.

I mean how do you even approach this kind of problem? The constraints are kind of there but we are too far from a defined optimization problem. At this point just keep thinking about random designs.

2025/3/23

This is entry covers the operation procedures of the AFM multimode 8 machine.

1. Login to the system, activate service, also fill the paper sign up sheet.

2. First step is to mount the tip onto the machine. First move optics up, unplug the wire powering the top piece (carefully hold both sides) and remove the springs left then right. Then lift the cantilever holder by rotating the back knob on the top piece, place the tip onto a brass tip holder using probe only tweezers. To unlock the brass holder, press the button onto table; the tip should have a T shaped line up instead of two parallel lines. Finally put the cantilever holder back, reattach the top piece but not the springs yet.

3. Next step is to align the laser. Plug in the power wire, move the entire top piece to a laser aligning stage. Rotate the bottom knobs on the top piece to align the laser. First go to positive x direction (clockwise) to hit the large chip, then go off to -x direction a little bit to find exactly the edge of the chip. Next rotate y knob to find when the laser spot is darkest to locate the tip. Finally go to -x to go to the edge of the tip. Put the top piece back on the machine and attach the springs right then left.

4. Final step for the hardware is to tweak the back knob to maximize SUM signal of the laser (around 5 ~ 6). Then change the machine to contact mode (will see a red light and the screen shows vertical and horizontal) Rotate the knobs on the left block to make both the vertical and horizonal readings 0.0~. Go back to tapping mode.

5. Now go to software. Select tapping mode, standard. Scan parameters: if homogenous sample, usually scan at 90 degree to avoid tip scratching damage, other depends on feature interested. Scan rate: usually 1Hz, minimum 0.5Hz maximum 2Hz. Resolution: usually 256, maximum 512. Turn on camera light at the power board or the white box below the machine. First move camera up then lower gradually to find the sample surface. Then move the bottom knobs to find the cantilever, then focus again on the sample surface. Turn on the motor to move the cantilever close to the sample as you see it just out of focus.

6. Next tune the tip to match the resonance frequency (tip may age and drive frequency can go up) finally click engage to engage the tip with the surface. Amplitude set point: notice feature height. Drive amplitude: may help get higher resolution but damage sample. Gain: controls feedback loop speed.

7. Finally after measurement click withdraw, move the top piece up through the motor, unplug the power wire, take off the cantilever holder, remove the tip. Remove the sample. Log out of the system.

2025/3/24

Attended the lab meeting at 9:30, mentioned learning basic laser alignment. Visited lab during afternoon to ask for a sample for AFM measurement. Lab uses AFM to confirm the growth surface is homogenous and smooth, estimate growth rate by measuring step size, etc.

Evening thought again about the sample holder design. Drew a engineer graph of the original design to scale. Currently interested in making a cavity in the plate center, attach a frame with bimetal strips that touch the sample when cold while bend away from it when heated, like a thermal circuit breaker. Reduce the height of the stage by removing 4 nuts. Or should even bring the stage plates directly into contact with the bottom plate. Concerned about non-uniform heating.

Not very optimistic about other designs involving mechanically moveable parts due to the small scale. Whatever lift or rotation seems unrealistic to handle manually.

In fact a good place to start is probably add a ceramic block contact right on sample center to verify if adding contact is helpful and to what extent. Should try metal contact as well.

Ceramic choices include alumina, aluminum nitride, boron nitride.

2025/3/26

Talked to G about sample holder design, agreed on trying ceramic blocks first, should start with alumina (sapphire. I actually just know that sapphire is just Al2O3 which is... disappointing. It is also true that sapphire comes in different colors due to impurities... I thought they were all blue) Searched for more information on alumina, aluminum nitride and boron nitride during evening, at room temperature it seems aluminum nitride has the best thermal conductivity but still, alumina provides the best electric insulation. It is fairly interesting to read that metals conduct heat through free electrons and thus have thermal conductivity proportional to absolute temperature times electrical conductivity; on the other hand non-metals conduct heat through lattice vibrations (phonons).

2025/4/2

Agreed upon buying a sapphire rod and slice into desired thickness (around 2mm) to make heat conducting pads. Will conduct a thermal test, G suggest mounting a substrate chip onto the sapphire pad, another directly on the sample plate, with thermal epoxy applied. Through ARPES measurement, map the electron energy distribution, which should follow Fermi-Dirac distribution (function has the form: f(E) = \frac{1}{e^{(E - \mu) / k_B T} + 1} ) which can give information of the substrate temperature through the slope. A sharper slope corresponds to a lower temperature.

2025/4/9

Today started the thermal test (preparation). Cut sapphire rod into 4 pieces (with a diamond spinning wheel blade!! In fact it was not exciting: the blade is not thin enough, the cut position can only be estimated, and when finally the rod is cut it flies off due to the clamp pressure...), the rod dimension is diameter 3.048mm, total length 9.398mm. The pieces have length 2.48mm, 2.3mm, 1.92mm, last piece broke in half, not measured. The ideal height is 2mm, if we should apply this to a real direct heating holder.

A holder plate is cleaned in alcohol, some silver epoxy was prepared (mix the two part glue together fully) and applied to the plate. The sapphire cylinder attached, then heated up (around 120 degree, test the temperature with a thermal couple) to help the epoxy solidify. Epoxy applied on top of the cylinder to attach a small piece of Bi2Se3, as well as to the side of the sapphire to form a thin line to ground Bi2Se3. Tested the grounding through a multimeter. Finally another piece of Bi2Se3 is glued to the plate beside the sapphire set to serve as reference. Heat up again, finally two ceramic rods are attached to the Bi2Se3 pieces for cleaving purpose. Heat up for 20 minutes to attain the best property of the epoxy.

Next transfer the whole thing into the fast intro chamber (it is not fast at all.) First stop the vacuum pump, wait for pressure to increase. Remove the view port (just unscrew the screws)(and beware of the copper seal ring). Insert the sample holder into its slot, do the lab journal, put on the view port again (make sure the screws are very tight!) Start pumping the chamber down to vacuum!! (can take 10hrs)

In summary, G is being very nice in leading this test. Always explained stuff clear and offer me opportunities to try things. I am just worried that he is too busy, as he always help with instrumentation stuff in the lab...

2025/4/10

Today afternoon from 12:30pm to 3:30pm did my first individual AFM scan, the process is rather difficult... I mounted a new tip (now it is stored on the bottom right corner of the gel box) and started the laser aligning stuff, all was well but when I started looking for the overlapped film edge I realized that my area of interest was not centered! Removed the whole top piece and nudged the sample film to the center. The trick is that the bottom platform moves the whole sample, the top piece moves the AFM tip, while the camera is always stationary. Basically it is always best to align everything (camera, sample area of interest, and AFM tip) to the center! So I did the laser alignment again, tuned the tip (which again caused confusion as it would not autotune! Nearly thought I broke the tip, but in fact it is just because the last operator used a softer tip which had a different frequency to mine, mine has 200kHz to 400kHz)

First did a scan of 2um*2um on the double-layer film region on the first square, bottom half, near the edge of overlapped films. Surface is rather smooth, some film bumps/dust are present as circles, height around 3nm to 5nm.

Then at the same region, zoomed in down to 200nm*200nm. Then zoom down to 20*20nm, things start look weird. And I realized my files are saved in somebody else's folder... quickly cut and paste, now I finally see my crappy files.
Changed to a region closer to the film overlapping edge, scanned 2um again, looks normal. However did not see the edge?? I tried to scan 10um then 100um, at 100um finally see the edge. I am seriously concerned right now because I did so many images and their purposes are kind of mixed up... I am also very concerned about what these jumps do to my scanning tip. The edge has a height of uhm,,, on scale of 100nm. Okay the output picture is very confusing, seems like color scale messed up due to crazy height difference. I zoomed down to 10um just at the previously saw edge and still the feature seems to be on the scale of 100nm.... no... but at least that is some feature....

Now zoomed in around (maybe it seems like on the edge) the edge and saw blobs... 760nm, blob height on 50nm level. Noticed the trace and retrace are the opposite of each other in the amplitude error data, which is what is should be I think, I mean when go up during trace, you go down during retrace.

Last final try since time is running out, 100nm, on the double film side, zero angle scan. just see some smooth, fuzzy surface. I don't know what to say. Finally did the same area 10nm, I don't know what to say, still fuzzy smooth surface. How can I make the resolution better? I played with the gain and drive amplitude a little bit, but I don't see what it really does.

There were just a lot of mysteries. I really need some better judgement and experience about all the parameters. Nevertheless it was my first measurement, worth celebrating that I didn't break anything. This is good enough.

2025/4/17

Today we did ARPES measurement on the sample, the final result was that we didn't find signal for Bi2Se3 on sapphire rod due to cleavage failure.

First, once the cool down is finished, the sample was transferred into the ARPES manipulator, ARPES shutter opened, sample cleaved by pushing off the ceramic posts.

Lower the manipulator, laser turned on (power and shutter), start locating the control group BS. The camera was very fuzzy and out of focus, had a hard time locating the cleaved area, so instead we did a spatial map of the ARPES data (step size?, steps on each side?), then successfully located the BS spectra (fermi level? electron energy?) We noticed some cutoff on one side, G's hypothesis was that the electrons might be blocked or interfered by local fields as the sapphire post and the control group BS are horizontally side by side. Next time should do vertical separation, and be mindful of aligning the ceramic posts better so the cleaved area align, easier to find. Besides moving the manipulator, G moved the laser position several times. In the end the spectra is decent, take a more homogeneous middle region, do integral across the spectra edge, we get the Fermi Dirac distribution, set the temperature by reading off the thermocouple (?K), we obtain a fit and the resolution (? ev). Use this resolution for the other test group fit so that we obtain temperature. Remember both resolution and temperature affect the slope in Fermi Dirac distribution, it can be difficult to tell apart their contribution, thus we do the control group first.

Then we continue with the sapphire rod group. Tried very hard to move around to find the cleave surface, not successful. Did a rectangular spatial map (? size), saw the edge of a circle, which we believe is the epoxy around the sapphire rod to keep it attached to holder. We did a spatial map for the whole sapphire rod upper view, see only a circle of epoxy (and even the line of epoxy for grounding purpose) but no BS on top of sample. It turned out that as we withdraw the manipulator from the ARPES chamber, our BS along with the epoxy to attach it to the sapphire are gone. I believe it is during cleaving, both are knocked off.

In the end we transferred the sample out back to intro chamber. But the transfer arm stuck and G resolved it later, which I had no idea how and eventually where the sample went.