Evan Yamaguchi

1/22/18

I met with Jon, a graduate student in Dr. Mohammad Hafezi's group in the JQI. I will initially be working on constructing the procedure for fabricating a two-dimensional hexagonal boron nitride (hBN)device. This procedure will involve optimizing the parameters in both electron-beam lithography and argon etching to deduce previously unknown properties of the material.

1/24/18

I met with Jon and worked through various paperwork and online trainings I needed to get access to the FabLab. Unfortunately, the paperwork would take a day to process, so I couldn't start technical training on this day. Instead, I looked through hBN flakes collected by another graduate student in the group and analyzed them to assess which ones may be best to first test out the techniques listed above on. I primarily characterized the flakes according to color, which corresponds to the thickness of the flake. An example of what these hBN flakes look like is pictured below:

1/29/18

To access the UMD Nanocenter FabLab, where the etching and lithography tools are housed, I needed to go through standard laboratory training and that's what I did today. I also scheduled time tomorrow for me to get training on Denton argon etcher, where I will assess the properties of hBN under certain etching conditions.

1/30/18

Today I brought used the Denton argon etcher on a sample of hBN. This was a preliminary assessment of the etch rate of the material under these conditions.

1/31/18

I analyzed the samples I etched yesterday using a microscope and assessed the degree to which the argon etch thinned the flakes I had previously selected. All 3 flakes were at least slightly thinned, and one was completely gone, giving us an estimate of the etch rate of hBN. I then took the sample to the Denton to etch again for 1 minute.

2/2/18

I analyzed and documented pictures of the flakes I previously etched. Curiously, there was no real evidence of etching under the conditions used, so I will plan on etching for 5 minutes next time. I was also shown the technique of spinning glass slides in the FabLab. This simple procedure is a method of preparing sample slides that will stick flakes to them via van Der Waals forces. All that this technique involves is applying a vaccuum to the glass slide, dripping the polymer onto the sample at one end of the slide, and then spinning the slide for 40 seconds to evenly coat the entire slide. The slide is then heated to evaporate off the excess liquid.

2/6/18

Jon discovered a protocol(https://pubs.acs.org/doi/suppl/10.1021/nl504750f/suppl_file/nl504750f_si_001.pdf)for etching hBN that uses fluorine etching instead of argon etching. This involves the use of a different machine from which I have been using, so I will have to become fully trained on this instrument. I have reserved a training session on Wednesday morning (2/7/18). Jon also sent me some slides about the project (Encap and Ebeam.pptx). 

2/7/18

Today I was trained on the Oxford Etcher (Fluorine) and the FabLab. This instrument is very straightforward to use and is much more efficient than the Denton Etcher (Argon) (1-5mins/sample vs 1 hour/sample). The method we used, ICP, was very powerful, and apparently etched into the SiO2 layer that the hBN flakes rest on. This is not ideal and I will investigate alternatives.

2/8/18

Mark, the FabLab staff who trained me on the fluorine etcher, came up with a true reactive ion etching (RIE) method what have a lowered overall power (75W). It is saved under 'hBN test RIE'. I tested a variety of flake samples and deduced that the etch rate is about .2-.33nm/sec. I plan to affirm this estimate with further etches on slightly thicker samples.

2/12/18

Now that I have full independent access to the fluorine etcher I can schedule time on the machine whenever I want to. I selected 3 more samples to verify the etch rate under this specific recipe(attached below). Qualitatively, the amount of hBN etched seems to be consistent with the previous rate. I had to do this analysis quantitatively because the heights of these samples had not been experimentally determined by atomic force microscopy (AFM).

I also learned today how to clean samples.

2/13/18

The next step in the process of the fabrication of this device is to determine the etch rate of PMMA under the conditions that I've been etching hBN at. This step is critical because after I write in the device pattern using electron-beam (e-beam) lithography, I will need to etch away the hBN not enclosed by the specified pattern.

I first learned how to spin PMMA onto clean samples. This process was very straightforward and simply involved my using the spinner to spin PMMA uniformly across the sample. I then baked the sample on a hot plate for 5 minutes and repeated the process to form a PMMA bilayer. I did this to 6 samples to assess the etch rate across a number of different etching times using the etching recipe specified previously. I was able to determine that the PMMA layer applied requires an etching time of greater than 5 minutes to etch away. This is good news for me because the hBN only requires about 20-30 seconds to etch away, leaving me lots of breathing room.

2/14/18

Today I selected a flake to write an e-beam pattern into so that I can create an hBN device. I selected flake .001.044 A3U0R2 due to its thickness and width. I went through the same PMMA spinning protocol as yesterday with this sample and then brought it over to the e-beam lithography machine. Jon had made a CAD file that specifies the design and walked me through the fairly complex process of setting up the e-beam writer. Once the writing process was done I learned how to clean off the residue PMMA from the sample by submerging the sample of MIBK for 1 minute. I baked the sample for 30 minutes on a hot plate and then went back to the FabLab to etch the sample for 30 seconds, which should be slightly more than enough time for the hBN surrounding the device pattern to be etched away.

2/19/18

Since it seems like the PMMA mask didn't quite work, Jon thinks that we should experiment with an Aluminum mask. To make this work, I now have to figure out the etch rate for aluminum. Jon had already made up a sample with different layers of the aluminum mask on it for me to etch. I fluorine etched the sample for 60 seconds and recorded the differences before and after the run.

2/20/18

I now have to use an acid mixture to wash away the excess aluminum on the sample. To learn how to do this, I contacted Tom at the FabLab, who walked me through the simple procedure of etching away aluminum under the fume hood. 15secs of applying acidic etchant to the sample was enough to completely etch away the aluminum without significantly damaging the SiO2 layer underneath. (pictures)

2/26/18

It seems that the etching protocol reduces the thickness of the SiO2 layer of the Si at a significant rate. To quantify the etch rate of SiO2 I used both the N&K Spectrophotometer and the flourine etcher mentioned previously. The spectrophotometer is a very easy piece of equipment that assesses the thickness of the SiO2 layer on the Si wafer. I initially had 6 samples each with ~2950 Angstroms of SiO2 and fluorine etched them at the time intervals 40s, 60s, 90s, 120s, 150s, and 180s (data shown below). It was in this way that I determined that the etch rate is about 4 Angstroms/second and steadily increases as the etch time increases. After I had safely determined this rate, I selected several hBN samples that contained flakes that I plan to use eBeam lithography on. This will be the first attempt at making this device using an alimunim mask instead of PMMA. 

2/28/18

Yesterday, Jon wrote CAD files outlining the hall bar pattern on a couple hBN flakes (pictured below). He then fed these flakes and patterns to the eBeam to pattern the PMMA that was applied previously. Today had I planned to fluorine etch the sample after Julia had deposited a layer of aluminum onto the sample, but the aluminum deposition procedure malfunctioned and was delayed until tomorrow. Instead, I installed AutoCAD and began playing around with the program in preparation for learning how to write the eBeam lithography files myself.

3/1/18

Today, Jon showed me the general process of aluminum deposition. The overall purpose of the aluminum mask is to further protect the hBN flake from the etching process. This was needed because the electron beam will create rebound electrons that affect the surrounding areas of eBeam contact. This will reduce the amount of PMMA in those areas past what was planned. When this sample would be taken to the fluorine etcher, these areas that were supposed to be protected were no longer protected to the extent that they should've been, and they are etched away more quickly. The aluminum mask, which is applied after eBeam lithography serves to protected the flake during the etching process, and since we know the rate that aluminum is etched away, we can calculate the exact amount of time that the sample needs to be etched for. 

So today I deposited aluminum on the sample using the Angstrom Aluminum Deposition machine. I then left the samples in acetone for one hour to dissolve away the PMMA in the areas that the eBeam patterned it (around the flake). This left aluminum residue on the flake, so I sprayed a little more acetone, followed by IPA on the samples and it all came off nicely. I took pictures of these, saved them under 'After Liftoff', and then fluorine etched them. I etched sample .001.009 for 150s and sample .002.003 for 90s. I took pictures of the resulting etches, but have to wait until I can acid etch the rest of the aluminum away before I can observe if the process was successful.

3/5/18

Today I acid etched samples .001.009 and .002.003.This was then followed by cleaning them for 5 minutes in acetone followed by 5 minutes in IPA, and then 5 minutes on the hot plate. I initially etched both samples for 15 seconds each, but when I returned to inspect them under the microscope, it seemed as though the flake on .001.009 still had aluminum on it. I then acid etched this sample for another 20 seconds and repeated the cleaning process on it. (need pictures) This process suggests that the aluminum mask proved successful and now we are going to try to repeat this process one more time to confirm the reliability.

3/6/18

Since we don't really have many more flakes to experiment on, Julia put more hBN on the new samples for me to use. I then flake hunted for a while to characterize any flakes on these samples that we can use. In this process, I identified 1 flake that looked thin enough and large enough to craft into a device element. I documented it (.002.052 A4U0R3) and Jon will teach me how to create CAD files to run the samples through the eBeam writer.

I also completed the Research Conduct training I needed to take to work in the lab.

3/7/18

Today I prepped the sample .002.052 for CAD file writing by spinning PMMA onto it. The procedure for this is as follows: Bring PMMA, samples, and tweezers to the FabLab. Make sure to find an appropriately sized spinning holder (smaller is better) and set the speed to 5000rpm and the timer to 45 seconds. Put the sample on the spinning holder, drop a single drop of PMMA onto the sample, and spin for 45 seconds. Bake the sample on the hot plate at 180C for 5 minutes and repeat the process again, this time baking the sample for 10 minutes. This produces a homogeneous double layer of PMMA on the sample. When tilted at a certain angle, the sample should look homogeneously and uniformly green.

The small blue inscriptions are the patterns that I wrote into the file to indicate to the eBeam writer where to write eBeam. The areas over which eBeam writes are area with PMMA.

Jon then taught me how to write the CAD file that describes how to write eBeam onto the PMMA in the desired pattern. I then proceeded to make these files for both sample .002.051, which contains 4 flakes, and sample .001.052, which contains 1 flake (pictures of CAD files).

3/8/18

Today I spun PMMA onto sample .002.051 with the intention of also using the eBeam writer to inscribe the desired pattern into it. The day mainly consisted of Jon walking me through the entire process of using the eBeam lithography writer. It is a long a seemingly complex process. Following the eBeam lithography protocol, the sample has to be cleaned up (the is the development process). I went to the lab in the basement and put the samples in MIBK for exactly 1 minute, then rinsed each one off with IPA. I then imaged all flakes and labeled them "After Development."

This is a picture of flake .002.052 A4U3R2 After Development. PMMA has been spun onto it and eBeam has written the pattern.

3/9/18

Today Jon deposited 30nm of aluminum onto each sample (.002.051 and .002.052) using the Angstrom Aluminum Deposition machine. He then left the samples in acetone for 1 hour and rinsed them off with acetone followed by IPA and the N2 gun. This process is called "Liftoff." I took over from here. I imaged the samples, labeling them under "After Liftoff." I noticed that some flakes had some aluminum residue remaining, so I sonicated both samples for no more than 5 seconds each which they were in acetone.

This is a picture of flake .002.052 A4U3R2 After Liftoff. Aluminum had been deposited across the entire sample and liftoff is the subsequent process that removes PMMA and everything on top of it, leaving only aluminum where eBeam wrote (plus some odd places as shown in the right half of the picture).

Flake .002.052 A4U3R2 after 90s Fluorine Etch 1. The hBN not covered by the aluminum has been etched away.

Flake .002.052 A4U3R2 after 15s Acid Etch 1. The aluminum has been removed, leaving only the desired sample. The purple square region around the sample is SiO2 at a different height than the rest of the sample due to fluorine etch exposure.

I then proceeded to fluorine etch the samples under the standard 'hBN test RIE' recipe loaded onto the Oxford Fluorine Etcher. I etched sample .002.051 for 150s (2.5mins) because the thickest flake on this sample was comparable to the thickest one the last time I fluorine etched. I etched sample .002.052 for 90s (1.5mins) because the flake on this sample was relatively thin. I brought these back to the PSC, imaged them, saved them under "Fluorine Etches" and then proceeded back to the clean room to acid etch the aluminum away. Each sample was acid etched for 15 seconds, rinsed with DI water, rinsed with acetone, and finally by IPA and dried with the N2 gun.

3/10/18-3/11/18

Over the weekend I went into flake hunt. This just means that I documented more uncharacterized flakes that I can use in further experiments.

3/12/18

Today I acid etched sample .002.051 for another 20s to try to clean off mysterious debris on the sample. I then put it in acetone for 10 mins and washed it off with IPA. The gunk remained and Jon and I do not know why. To test of this gunk is fumbling oil from the fluorine etching process, I put the sample in acetone, put it on a hot plate set for 60 degrees Celsius and left it overnight. For the rest of my time in the lab I finished flake hunting on all the samples that have been made up. This includes samples up to .002.070.

3/15/18

Today I was trained to use the Angstrom to deposit aluminum on sample .001.002 (This sample has no significance to the project, it was just to use for my training). The steps for the aluminum deposition process are as follows.

Tape the sample to the wafer by the corners. Unlock the main chamber door and hit the vent/evacuate button. Wait ~5mins for the chamber to vent. Change the metal setting by switching to 'Manual' and rotating to 'Aluminum' (for aluminum deposition obviously). Take the wafer with the sample loaded onto it, carefully place it in a sample holder upside down. Check to make sure there's enough aluminum in the capsule. If not, go over to the metal cabinet and pour a little more into the holder. Once all this is done, closer and lock the chamber and pump down. Wait a couple minutes to see the pressure gauge decrease, but you still have to wait ~1hour for the chamber to pump down. Once the chamber has pumped down, follow the instructions on the paper by the Angstrom. The paper will (from memory) basically describe everything I'll need to do to operate the instrument.

Later today I successfully ran eBeam lithography on samples .002.054 and .002.056 and imaged them. The process seems to have worked as planned.

.002.054

.002.056

3/26/18

Broadly, I am perfecting my ability to use the eBeam writer so that I can operate the machine without even thinking about it. As of right now I am waiting for another part of the project to be completed so that I can proceed.

Today I set up the CAD files for eBeam lithography I plan to do tomorrow on hBN samples .002.055 and .002.058. I also helped out around the lab with various jobs that needed to be done.

3/27/18

Today I used the eBeam lithography writer to write two more samples. This was mostly practice as I am currently just repeating the procedure to test its resolve. I also spun Elvacite onto a sample for Jon to write eBeam onto it to test its viability as the base for a PMMA clean.

.002.055 C2U0R4 After Development

.002.058 A4U4R1 After Development

3/28/18

Today I wanted to test an etching recipe designed specifically for use with just a PMMA mask. This recipe was gathered from reference (1). Another benefit to this etching recipe is that it is reported to be a bit more tame than the previous recipe, meaning that we will have better control over the thickness of the sample. Unfortunately, the Oxford Fluorine etcher software did not allow me to compose a recipe with the two gases we need to use. This is due to a mechanical issue that we need to resolve with the Nanocenter. I will have to wait for this issue to be resolved before I can proceed.

I also spun PMMA down on the surface of a sample for Jon to test the efficacy of a polymer called elvacide as the base for a sample cleaning.

Lastly, I tried to learn more about the history of 2D materials and their properties by reading (1) and (4).

4/2/18

Today I helped to design some structural elements of a lab space we plan to move into in the future. The space is relatively limited and we are trying to maximize its use by building in storage space around the area where we plan to make measurements with the 2D device.

4/9/18

We got word today that the Trion Etcher will be hooked up with a CHF3 gas line on Wednesday so that we can apply our etching recipe to the samples I've eBeam'd already. I also toured the IREAP machine shop so that we could see what materials and tools they have to help us construct an acrylic box for us to put our mechanical exfoliation system in.

4/11/18

Today I was finally able to use the Trion Etcher to see if our recipe works. I was given a tutorial on how to use the device (super easy) and proceeded to etch two samples at varying time intervals. I etched hBN sample .002.054 for 10s and hBN sample .002.056 for 60s and lifted off the PMMA for about 45 seconds each in aceetone. The etch recipe worked successfully and we plan to confirm the fidelity of the procedure tomorrow with 3 more etching tests.

hBN.002.054 C1U0R0, 10s Trion etch, after liftoff

hBN.002.056 D2U2R2, 60s Trion etch, after liftoff

4/12/18

Today I confirmed this procedure with 3 more etching tests, as described in my last entry.

hBN.002.058 A4U4R1, 20s Trion etch, after liftoff

hBN.002.062 B1U1R0, 30s Trion etch, after liftoff

hBN.002.062 C3U3R2, 30s Trion etch, after liftoff

hBN.002.063 B5U0R3, 40s Trion etch, after liftoff

hBN.002.063 C5U1R4, 40s Trion etch, after liftoff

4/16/18

Today I put together CAD files for a device pattern that we can then feed into the eBeam machine. [explain more about specifics]. I then spun PMMA onto the dozen or so device samples that I made some CAD files for. I had issues because the PMMA would not spin evenly onto the wafers due to residual polymer materials. Then slightly complicated the procedure but ultimately did not fundamentally change my procedure.

4/17/18

Today I finished writing the CAD files for g.008.010 and g.008.012 and then wrote eBeam onto both samples (1800uC;500um;5nA). I then etched sample g.008.010 for 20s in the Trion Etcher and took pictures. It seems to have worked. I then left g.008.010 in acetone overnight to liftoff the PMMA.

g.008.010 e-Beam, 20s Trion etch

4/18/18

At this point, I had one sample, g.008.010, that had been etched so as to expose the sides of the hBN-G-hBN stack.

4/19/18

I etched g.008.012 using the Trion etcher in the exact same manner as I etched g.008.010. I imaged the sample, and it looks like it turned out as desired. It turns out that the Angstrom deposition machine is not available for use until Saturday, so Jon is going to do the aluminum deposition and PMMA liftoff.

g.008.012 e-Beam, 20s Trion etch

4/23/18

As planned, Jon deposited gold contacts onto samples g.008.010 and g.008.012 and lifted off the PMMA by leaving the samples in acetone overnight. The results are pictured below. Since we do not have a crucial part to our preliminary measurement system, we cannot fabrication the connections to the system and see if the device conducts electricity as it should.

g.008.010, after gold contact deposition and PMMA liftoff

g.008.012, after gold contact deposition and PMMA liftoff

4/24/18

Today, since we cannot do measurements on the sample, I designed the breakout box for the measurement system. I was instructed to use company software that lets you specifically implement all necessary punch-out holes and labels on the box. This is essentially the system/human interface.

4/25/18

Today, 80/20 structure material arrived so I helped to put together structures for the new lab we are moving into. 

4/26/18

I used powertools to configure an large aluminum sheet to fit into the shelving structure previously constructed. This involved cutting the metal and drilling holes into it for screws.

4/27/18

Today I put together more 80/20 structures for the upper layer of shelving that will be positioned above the optical tables in our new lab.

4/30/18

References:

(1) One-Dimensional Electrical Contact to a Two-Dimensional Material: http://science.sciencemag.org/content/342/6158/614.full

(2) PDF discussing semiconductor physics: http://www.ece.umd.edu/~dilli/courses/enee313_spr09/files/supplement1_carrierconc.pdf

(3) New hBN device fabrication protocol: https://pubs.acs.org/doi/suppl/10.1021/nl504750f/suppl_file/nl504750f_si_001.pdf

(4) Graphene-Like Two-Dimensional Materials: https://pubs.acs.org/doi/pdf/10.1021/cr300263a