Fabrication of Superconducting Resonators on hBN Thin Films
River Chen - Parallel I
Author
09/26/2024
Added
4
Plays
Description
Student’s name: River Chen
Home Institution: University of Illinois at Urbana-Champaign
NNCI Site: CNF @ Cornell University
REU Principal Investigator: Professor Zhiting Tian
REU Mentor: Joyce Christiansen-Salameh
Abstract: 2D materials have attracted much attention due to their unique physical and chemical properties. Here we investigate the integration of a thin film grown of one of these 2D materials, hexagonal Boron Nitride (hBN), into a superconducting device. We believe that hBN will be compatible with the device because of its low dielectric loss tangent. The device, a coplanar waveguide resonator, was designed for high sensitivity to loss at the metal-substrate interface, allowing us to compare loss between a device with hBN and without. We fabricate a Nb on hBN on sapphire resonator and compare it to a standard Nb on sapphire resonator. The fabrication process was performed at the Cornell Nanoscale Facility at Cornell University, using the AJA sputter deposition tool, ABM Contact Mask Aligner, and Plasmatherm 770 Etcher. In this work, we explore the various adjustments made in order to accommodate and prevent damage to the hBN substrate. Additionally, certain adjustments to the device design were made in order to accommodate the testing platform at Fermilab.
Searchable Transcript
Toggle between list and paragraph view.
- [00:00:00.000]Hello everyone, my name is Rupert Chatt.
- [00:00:06.240]I study Material Science and Engineering at the University of Illinois Organic Champaign.
- [00:00:10.880]This summer I will conduct an aggregation of superconducting resonators on HVM thin foams.
- [00:00:17.440]A big thank you to my mentor, HG Kennedy Joyce Christensen-Salema,
- [00:00:23.200]and to the Cornell Nanoscale Facility for supporting my work.
- [00:00:29.520]So as sort of a broad background introduction to what I worked on this summer,
- [00:00:38.960]quantum devices have a problem, and that problem is that they experience a lot of
- [00:00:47.600]loss because of the inherent fragility of these quantum systems.
- [00:00:51.840]And this loss leads to loss in information, which is what quantum systems operate on.
- [00:01:00.560]So what we are specifically focusing on in our research are the dielectric losses from
- [00:01:07.840]the metal-substrate interface, which have been found in previous
- [00:01:13.920]research to be a significant source of loss in superconducting platforms.
- [00:01:19.520]So how will we do this? We will use a design for a superconducting resonator,
- [00:01:29.680]and what these do is you sort of have this line, and an electromagnetic wave will resonate.
- [00:01:37.920]A standing electromagnetic wave will occur along this line,
- [00:01:42.960]and what's special about these superconducting resonators is they have a very high Q-factor
- [00:01:49.360]and just a big word that means inversely proportional to loss.
- [00:01:53.760]And so when you are able to measure that very minuscule amount of loss and you compare them
- [00:02:02.480]between eight differently sized resonators on our mask design, you can determine,
- [00:02:08.480]calculate the dielectric loss that occurs from the metal substrate intruders.
- [00:02:12.240]Now hexagonal boron nitride, we do not know if it has low dielectric
- [00:02:19.280]loss of the metal substrate interface. However, we do know that first of all it's chemically stable
- [00:02:24.720]and stability is an important thing on systems. Another thing is that it has atomically flat
- [00:02:31.120]surfaces free of dangling bonds. What that means is, as you can see in this sort of xy plane,
- [00:02:35.840]you have strongly bonded boron and nitrogen atoms, whereas in the z plane you have sort of weaker
- [00:02:43.440]better wall bonds and that sort of inhibits dangling bonds just from happening all around the surface.
- [00:02:49.840]Another thing is that HBM has low dielectric loss, hexagonal boron nitride, and what that means,
- [00:02:55.760]how is that different from low dielectric loss at the metal substrate interface?
- [00:03:01.360]What that means is that in this sort of bulk material there's low dielectric loss,
- [00:03:07.040]whereas the interface is what we're trying to study.
- [00:03:09.760]And all of these properties make HBM an attractively compatible platform for superconducting circuits.
- [00:03:19.120]So a sort of previous work that was done was where they took a flake of HBM,
- [00:03:25.440]hexagonal boron nitride, and they placed it on top of a superconducting device,
- [00:03:29.840]and they found that the Q factor, that thing that's inversely proportional to loss,
- [00:03:35.360]was similar or even slightly lower when they introduced this new material.
- [00:03:39.680]However, this was only studying the dielectric loss in bulk, as I talked about before, and the reason
- [00:03:49.040]why our research goal for this summer was to employ a specific design that designed with the
- [00:03:55.360]eight resonators going across a feed line. And it's sensitive to dielectric loss at the metal
- [00:04:02.000]structure interface, and through this design we will compare a control niobium on SAFIRE
- [00:04:09.120]and a niobium on HVN on SAFIRE device. And how did we make it? First we grew about five nanometers
- [00:04:18.960]of hexagonal boron nitride on a 10 millimeter by 10 millimeter SAFIRE substrate.
- [00:04:25.360]Now we used these techniques such as electron diffraction, Raman,
- [00:04:32.960]and AFM just to make sure that our film was first of all epitaxial, which means that the
- [00:04:41.040]HVN formed in the ordered layers on top of our substrate. And then we wanted to use Raman
- [00:04:48.880]to make sure that was HVN and not a mystery material. And we also used AFM just to make
- [00:04:55.040]sure the surface of our film was smooth. All right, so after we deposited that layer of HVN,
- [00:05:08.160]we sputtered 60 nanometers of niobium, a superconducting metal, onto our BN film. Now,
- [00:05:17.760]originally we did this
- [00:05:18.800]at the default for the sputtering machine, a high power. And when we performed TEM,
- [00:05:24.800]transmission electron microscopy, we found that the metal ion bombardment from the sputter had
- [00:05:32.560]damaged our BN thin film. So you can see a sort of wishy-washy area there. It's neither just BN
- [00:05:38.560]or niobium. And we were like, how will we solve this? We can sputter at a lower power. If you
- [00:05:44.080]hit it less hard, it's probably not going to have chunks of BN flying around.
- [00:05:48.720]So that is what's shown on the right here. You can see a huge contrast between niobium and BN.
- [00:05:56.640]And you might be wondering, why is this niobium so bright? No, it's blinding. And that's because
- [00:06:05.040]when we cut the cross-section of our niobium on BN, the niobium had smeared over the BN thin film.
- [00:06:12.400]So we just turned up the brightness to establish that contrast.
- [00:06:18.640]All right, photolithography was the next step in our process. We spin photoresist,
- [00:06:26.720]photoresist, and use a contact mask to essentially weaken the photoresist
- [00:06:34.080]by a UV light so that the weakened areas would wash away in development.
- [00:06:39.440]And at this point, I'd like to mention that our original device design was a
- [00:06:48.560]design for a 10 by 10 millimeter platform, but our testing parameters at
- [00:06:53.520]Fermilab only operate on 6x6, so I just rescaled everything to a 6x6, moved everything closer
- [00:07:03.040]together. But the important part about this is you have to make sure that they aren't too close
- [00:07:08.800]together because then the resonators can have crosstalk with each other and interfere, and
- [00:07:14.400]that's not good because the whole point is that each resonator is its own system and we're trying
- [00:07:18.480]to measure what happens in each system. So when we did photolithography, when we exposed the
- [00:07:28.560]photoresist, we varied the times, the dose rate of this exposure because we wanted to get the
- [00:07:38.240]optimal feature definition for our superconducted device. And we found that at 1.6 seconds,
- [00:07:44.400]we got what we wanted. We got these
- [00:07:48.400]three micron gaps and a six micron conductor width.
- [00:07:52.560]After performing that photolithography step, we moved on to etching.
- [00:08:02.960]Reactive ion etching is this process where you excite a plasma, you use an electric
- [00:08:11.280]field to move ions in a plasma towards its heart. And this is great for our application
- [00:08:17.280]because reactive ion etching allows for high aspect ratio features. And our resonator is very long and narrow.
- [00:08:24.070]So that's sort of just a process of how reactive ion actually happens. Then after that we strip
- [00:08:33.910]with photoresist and the pattern has appeared on our niopium. However, when we were doing that
- [00:08:43.430]etching stuff, we realized that what happens after you get through the niopium, there's Bn.
- [00:08:48.870]And how we were etching our material was this chlorine chemistry. So Cl2 has been shown to
- [00:08:56.870]react with boron nitride. So we wanted to know how to not etch away our Bn and also
- [00:09:05.750]how much the Cl2 would react with it. So firstly we determined the etch rate.
- [00:09:12.310]of our niobium by just etching a little bit of the niobium away and measuring it by
- [00:09:18.470]coulometry. And that picture on the right just shows a really nice vertical sidewalk that we got.
- [00:09:22.950]And then we figured out hey if our chlorine gets around our Bn how long do we have. We found that
- [00:09:35.270]about one nanometer of Bn would be etched every three seconds of our process which is pretty good.
- [00:09:40.310]You know
- [00:09:42.230]not disastrous and if we found our niobium etched time properly
- [00:09:48.950]then the Bn film would not be affected too much.
- [00:09:53.110]Next we verified that the niobium was once again completely etched through
- [00:10:02.470]using the probe station and basically when the gap trench shown in the red and the
- [00:10:12.150]conductor band shown in the black we used the probe station and we measured these two different
- [00:10:18.230]areas and if the resistance was zero that we know is shorted and there's still some niobium left over
- [00:10:24.630]we have to increase our etch time maybe go back over our niobium etch test. After verifying that
- [00:10:35.430]trench width we would use a physical property measurement system which can measure the super
- [00:10:42.070]conducting transition temperature of a given metal and you might be wondering I think niobium
- [00:10:49.110]is a very common superconducting metal right why would you want to verify this transition
- [00:10:55.830]temperature and the reason for that is the hdm is another you know electronic component of our
- [00:11:03.670]system that can shift the superconducting transition temperature in a good direction.
- [00:11:11.990]The next step would just be dicing to fit in the six by six millimeter testing specifications
- [00:11:19.030]and then we would send the final devices to Fermilab where they would be wire bonded and
- [00:11:24.950]tested in a quiet controlled environment. So yeah thank you for listening to my presentation.
- [00:11:33.270]Is there any questions?
- [00:11:42.310]Yes. Well you mentioned that there's a weak and walls forces like vertically in the atmosphere
- [00:11:49.270]so I'm wondering why that's important and also when it's interacting with an IOP
- [00:11:54.630]at the interface of those two is that important in any way?
- [00:11:59.910]So the question was what are the what's the significance of the van der Waals bonds and also
- [00:12:11.830]the significance of the interaction at the interface? So let's go back to the beyond slide.
- [00:12:19.510]The significance of the van der Waals bonds in that z-plane is that it's less likely for
- [00:12:30.310]there to be dangling bonds in the z-plane so on the top layer right you can imagine we have
- [00:12:36.630]our niobium on that top layer you have a bunch of van der Waals bonds there's unlikely to be
- [00:12:41.750]what we call parasitic two-level systems so like an OH group just dangling from the top
- [00:12:47.750]that is less likely to be there because the van der Waals bond is not going to let the OH stick
- [00:12:51.750]and if I did my job right with acetone in the IPA and it should be a relatively pristine surface
- [00:12:58.630]as for the interaction at the interface that is like sort of the crux of our
- [00:13:04.070]our research is just determining the interfacial loss between the
- [00:13:11.670]BN and the residue are you asking about specific mechanism that happens there yes
- [00:13:16.710]all right so there's a lot of time back there basically when you have a superconducting metal
- [00:13:25.270]the electrons are sort of all in all occupying multiple states at one time
- [00:13:33.510]but the thing is that that's not happening in the BN so you have this sort of
- [00:13:37.830]mismatch between the electronic states and the BN at the residue
- [00:13:41.590]resonator and a lot of loss can occur there and we're just trying to find
- [00:13:46.950]oh are the electron states the cooper pairs that allow electrons to superconduct are those going
- [00:13:52.790]to be disrupted or not you mentioned that having the resonators to post together cause like some
- [00:14:04.390]issues is there any worry that with like going between the resonator and the sapphire even though
- [00:14:11.510]yes um the question was about whether the sapphire could interfere with the resonator yeah or like
- [00:14:25.590]still is the boron nitride like completely blocking out any dielectric loss or it could
- [00:14:32.070]still have lost anything that the electric field goes into can cause loss but that's very low
- [00:14:41.430]very low loss is also undesirable but um it's sort of a balance thing so we will be using various high
- [00:14:48.150]quality seaplane sapphire yes this is a follow also super cool to learn about thanks for sharing
- [00:14:58.190]but i was wondering like how well characterized like loss between like the hbn and sapphire
- [00:15:04.990]has been in like literature and stuff or if there are any plans on like figuring out how to decouple
- [00:15:11.350]the loss from the metal interface from the loss from the sapphire interface
- [00:15:15.830]i will say that this is the first time they put hbn on it oh the question was um
- [00:15:24.710]first of all figuring out if there's a lot of horizons from the hbn on the sapphire
- [00:15:31.750]um and my answer to that is uh this is a relatively new research so i guess um
- [00:15:41.270]this would be sort of person so we could study that um the other question was about decoupling
- [00:15:49.190]what was it again the or like if you're measuring if your goal is to measure the loss between like
- [00:15:55.270]from the metal interface if there are any ways or there are any thoughts you have on like decoupling
- [00:16:01.190]the loss from this sapphire so you know like what losses from the metal what losses from the time
- [00:16:06.790]um that would be a question
- [00:16:11.190]for oh uh how to
- [00:16:13.030]uh separate the loss that arises at the metal structure interface and that sort of end up
- [00:16:20.870]sapphire um that would be a question for future research i think um it would be pretty tough to
- [00:16:28.230]just get a freestanding layer of boron nitride um but yeah
The screen size you are trying to search captions on is too small!
You can always jump over to MediaHub and check it out there.
Log in to post comments
Embed
Copy the following code into your page
HTML
<div style="padding-top: 56.25%; overflow: hidden; position:relative; -webkit-box-flex: 1; flex-grow: 1;"> <iframe style="bottom: 0; left: 0; position: absolute; right: 0; top: 0; border: 0; height: 100%; width: 100%;" src="https://mediahub.unl.edu/media/23117?format=iframe&autoplay=0" title="Video Player: Fabrication of Superconducting Resonators on hBN Thin Films" allowfullscreen ></iframe> </div>
Comments
0 Comments