Conductive Domain Walls and Resistive Switching in the Ferroelectric AlScN and NbOI2

Haidong Lu Author

05/20/2025 Added

0 Plays

Description

Haidong Lu presents "Conductive Domain Walls and Resistive Switching in the Ferroelectric AlScN and NbOI2" for this joint Grand Challenges (Quantum Approaches) IRG1 and EQUATE FRG1 meeting.

Searchable Transcript

Search:

[00:00:00.080]If we discuss our recent work on the conductive domain walls and resistive switching in the
[00:00:06.720]ferroelectric aluminum scandium nitride and also another ferroelectric niobium oxide diode iodide.
[00:00:15.920]So the story begins with the two-dimensional electron gas at the lanthanum aluminum oxide
[00:00:25.440]and strontium titanate oxide. So this discovery of 2D electron gas triggered the investigation
[00:00:33.680]for the conductive interfaces with the device concept of using such conductive interface
[00:00:40.720]compared to traditional 3D structure. So this mystery really lies in the polar discontinuity
[00:00:55.200]polar discontinuity. So because lanthanum and aluminum oxide it's a basically a polar material
[00:01:02.160]and strontium oxide is non-polar. So once you put these two materials two insulators together
[00:01:08.880]and suddenly there is in conductivity at the interface. So it's believed that it's this
[00:01:16.800]connectivity stems from the polar discontinuity that because of the increasing potential energy
[00:01:24.960]with the polar material and there is a charge transfer to the as to toward the interface in
[00:01:31.440]order to reduce such electrostatic potential energy making this interface conductive. And
[00:01:39.120]since then there are quite several materials showing such polar non-polar conductive interface
[00:01:45.360]and one particular example is a fellow between a ferroelectric bismuth ferrite and non-polar
[00:01:54.720]terbium scandate. So and here the ferroelectric polarization it's switchable so it provides
[00:02:03.120]additional degree of controlling the interface conductivity also possibly controlling the type
[00:02:09.920]of carrier carriers. So on the other hand polar discontinuity is it's basically routinely exist in
[00:02:24.480]in industrial materials as charges of the main walls. So let me use the
[00:02:30.640]laser point. So yeah basically once you have the polarization like oriented in head-to-head
[00:02:38.400]or tail-to-tail orientation so there will be charge polarization charge built up at the
[00:02:46.160]interface or the main wall. So this charge basically increases the electrostatic
[00:02:54.240]energy at the domain wall and which may cause a similar type of
[00:02:59.680]reconstruction in the electronic structure at the domain wall and there could be charge
[00:03:06.800]transfer to the domain wall if the material itself is like even non-conductive. So this
[00:03:15.920]could make this domain wall conductive and which is called conductive domain walls and
[00:03:24.000]or if the material itself is semiconducting so this polarization charge may attract free charge
[00:03:31.280]carriers to the domain wall so also make the connector the main wall conductive. And
[00:03:37.760]indeed there are quite several quite many observations of such domain wall conductivity
[00:03:45.360]and starting from the bithumous ferrite and then like to barium titanate pct
[00:03:53.760]and lithium niobate so quite a lot of material showing such charged domain wall conductivity.
[00:04:00.240]And another example is the improper ferroelectric urban magnet so basically you have tail-to-tail
[00:04:09.120]and head-to-head domain wall it's showing different type of conductivity till the tail here it's increasing
[00:04:15.680]the connectivity and head-to-head reducing the conductivity. Of course this material itself is
[00:04:23.520]semi-conducting so but basically the idea is that this kind of charge domain wall can modify the
[00:04:31.040]local connectivity as the domain wall so this involves a potential application use such
[00:04:39.440]to the interfaces employing such conductive domain walls as functional elements in the
[00:04:45.920]device structure and advantages advantages of using such the main wall connectivity
[00:04:53.280]as uh divide in the devices is basically these domain walls are routinely exist
[00:04:59.360]in the direction materials and they can be created electrically or can be erased by by voltage
[00:05:07.360]so and also we can control the density of such domain walls in the material by creating more
[00:05:16.160]domain walls basically you can we have a additional degree of controlling or erasing such
[00:05:23.040]conductive interfaces
[00:05:24.240]and so here we work on the aluminum cyanium nitrite so this is a belongs a type of the three
[00:05:36.320]nitrites uh semiconductor semiconductor materials so then basically this type of material has been
[00:05:43.840]long been used in the 3-5 semi-tactile industry and the advantage is also this material is seamless
[00:05:52.800]compatible so um can be easily integrated into the uh with sinus technology and so the host material
[00:06:03.600]aluminum nitrite it has the um wall site structure so depending on determination one can have the end
[00:06:15.360]polar or nitrogen polar with polarization down and metal polar with polarization up state however
[00:06:22.560]the aluminum nitrite without doping it's polar but it's non-ferroelectric because polarization
[00:06:29.440]is not switchable however with scandium doping and it makes uh the it reduces the switching voltage
[00:06:38.160]or reduces the cursive field so making this material switch so you're shown here is uh
[00:06:45.120]literature showing that with the increasing doping of scandium so there is also a reduction of the
[00:06:52.320]cursive field making it uh thermoelectrically switchable of course the cursive field is still
[00:06:58.560]much higher than conventional ferroelectrics like pzt at least one other magnitude higher
[00:07:05.440]and also another property of this material is that it shows pretty high remnant polarization
[00:07:12.640]it could be above 100 microchrom per centimeter square it's one of the highest materials with
[00:07:22.080]remnant polarization so on the other hand so recent investigation shown here is a tm
[00:07:30.000]image of a thin film of couple of nanometers of scanning aluminum scanning nitrite
[00:07:37.680]so this investigation have shown that there could be stable charge of the domain walls shown here
[00:07:44.960]dashed line is the domain wall between the m polar and n polar phase so basically you
[00:07:51.840]have stable charge to the main walls in the material so this provides the basis to study
[00:07:58.720]those charges of the main wall properties in this material because they are stable and they can be
[00:08:07.680]stably exist in it's not relaxing at least so our investigation uses the 20 nanometer
[00:08:21.600]of aluminum titanium nitride was 28 percent doping and we have platinum as top and bottom electrodes
[00:08:30.320]so basically we have a capacitor structure device and um preliminary study shows a good
[00:08:38.000]polarization switching shown here is the switching current as a functional voltage
[00:08:43.360]and also pfm switching spectroscopy shows good polarization switching behavior and um
[00:08:51.360]pfm imaging in this capacitor structure provides some additional details of the polarization
[00:08:58.880]switching dynamics and shown here is an example of the capacitor structure and we have the
[00:09:05.600]initial downward polarization state and with application of volatile voltage pulses we can
[00:09:14.560]basically create like intermediate state where it's part of the main state or a four-way
[00:09:21.120]switched upward the domain state and shown in the bottom is a detailed study of the
[00:09:27.840]high resolution pfm images of the domain structure so basically we have the pristine
[00:09:34.880]downward state and gradually switching to the upwards the state as a different voltage
[00:09:41.520]amplitude or and different um posturation so basically with this we can
[00:09:50.880]of course we can get quite some information of the switching dynamics like
[00:09:58.080]the main density or nucleation density and the main wall motion speed
[00:10:06.960]or and also basically with different voltage we can switch the capacitor at a different time scale
[00:10:14.560]shown here like nine volts we can switch it in microseconds 12 volts
[00:10:21.200]sorry millisecond 12 volts microseconds and 13 volts nanosecond basically we have
[00:10:28.240]a control of several orders of magnitude for the switching speeds with different voltage
[00:10:33.920]and resistive switching in this device structure is characterized by the
[00:10:43.440]IV characteristics so here basically we measure the IV characteristics as
[00:10:50.400]a function of the polarization state shown here we have seven different stage
[00:10:56.560]initial polarization down stage so we have IV shown here and basically here is the summary so
[00:11:04.400]basically initially it's starting from the downward and then whereas and the upward domain
[00:11:12.640]starts to grow at stage two and there is a significant increase in the current density and then
[00:11:20.160]it's in the polar domain state and it reaches maximum of the device conductivity and stage five
[00:11:29.920]once it's fully polarized upward there is a drop of the conductivity and of course it's not
[00:11:36.560]at this same level but still it's much lower than the
[00:11:41.040]polar domain state and then stage six again it goes back to quite the main state and there's
[00:11:49.920]again a jump up of the device conductivity and stage seven it's a fully downward state and the
[00:11:59.760]device connectivity returned to the initial state so basically polarization
[00:12:05.680]up or down the fully polarized state it corresponds to a relatively lower conductivity
[00:12:13.440]of the device or the highest highest conductivity is
[00:12:19.680]it's corresponding to the uh poly domain state so this indicates that it's possibly
[00:12:26.880]the main walls will be conductive because in the polydomain state we have a higher density
[00:12:34.080]of the main walls compared to the like lower kind of creativity state of the upward or downward state
[00:12:41.280]so and based on this property uh we can basically create
[00:12:49.440]a uh name register or like a multiple level storage device um so shown here we have basically
[00:12:59.760]we can we can circle the device from the downward polarization state and with increasing voltage
[00:13:06.960]powers and we stop at the polydomain state because it shows the highest conductivity and we can of
[00:13:13.920]course go back to the initial state so it shows the initial state
[00:13:19.200]it stresses of the demand sorry the device conductance as a function of voltage and of
[00:13:27.840]course we can also measure the device conductance as a function of pulse time so here we fix the
[00:13:35.520]voltage at eight volts and um eight volts uh it creates a polydomain state of course
[00:13:43.360]and where the increasing pulse time can continuously tune the device connectivity
[00:13:48.960]um so so more than one order of magnitude basically and with the minus voltage we can um
[00:13:58.240]device connectivity returns to the initial low connectivity state
[00:14:04.000]so basically we can program this device with a specific pulse time and
[00:14:11.120]to make it at this multiple level also the device resistance states and um
[00:14:18.720]so this is basically um and it's also post time dependence so basically it's a prehistory
[00:14:25.840]dependent uh resistance multiple resistance state and that's uh basically a memorizer so uh using
[00:14:33.520]the conductive domain walls we can create such a memorizative device in this uh i'm scanning nitride
[00:14:40.960]capacitor structure and of course uh we can tune the switching speeds as previously mentioned uh
[00:14:48.480]we can basically turn the switching speed at with different voltage amplitude so uh shown here um
[00:14:56.160]it's basically we can tune the switching speed or turn the device connectivity at different
[00:15:05.760]pulse duration uh with different different voltage amplitude and for different curves and like this
[00:15:14.240]green one it's at six volts and uh
[00:15:18.240]sine one it's eight volts so they apparently can turn the device um at different uh pulse
[00:15:26.720]durations so basically um voltage amplitude also can turn the device working at different time scale
[00:15:35.440]so um in order to uh observe directly the domain wall conductivity so of course previously we used
[00:15:48.000]the uh here we used the polycrystalline um films so given the domain the grains are pretty small
[00:15:58.400]so um basically the domains are most right most likely not very um homogeneous so uh we switched
[00:16:09.600]to here we switched it to single crystalline comes from by mocvd so the so this provides
[00:16:17.760]a like a homogeneous film for for the domain switching and also um for the domain walls
[00:16:24.720]uh to observe the domain walls and domain wall connectivity and um of course we have a much
[00:16:31.680]thicker film of uh 230 nanometers this is uh it's limited by the experimental condition
[00:16:40.240]and we have the same type of structure just with different type of uh
[00:16:47.520]uh top and bottom electrodes and um so shown here is a tm image and so we have a top electrode and
[00:16:59.360]this is the fm um element scanning nitrite film and this capacitor has been polarized um downward
[00:17:09.120]with the initial polarization upward or in polar state so from the tm image we already see this the
[00:17:17.280]uh kind of domain walls so basically in the port region we still have those um zigzag domain walls
[00:17:26.640]buried underneath the domain wall and on the edge of course we have
[00:17:30.720]also the domain wall and from the high resolution tm we can see that those domain walls
[00:17:40.400]are inclined with the maximum inclination angle of around 15 degrees so basically they
[00:17:47.040]are also charged there because um with the inclination we have the head-to-head oriented
[00:17:54.480]oriented um domain walls charge of the main walls in this material so this provides a basis to
[00:18:04.160]directly investigate the charge of the main wall properties using local probe local probe techniques
[00:18:16.800]uh so yeah and uh of course we also checked the resistive switching behavior in this uh single
[00:18:23.840]crystal single crystalline genres of the nanometer in scanning nitrite capacitor structure so
[00:18:30.240]basically we have uh part of the main state corresponding to the highest uh domain highest
[00:18:38.480]device conductivity the red one compared to the fully polarized upward or downward state and um
[00:18:46.560]of course we can turn the device connectivity with uh pulse amplitude from four poles fully
[00:18:54.800]polarized upstate to fully polarized down state so the current density increases first due to the
[00:19:03.360]creation of those domains or increasing domain density domain wall density and then gradually
[00:19:10.720]reduce to the fully downward state and we can also turn this device
[00:19:16.320]uh using the number of pulses or basically um postulation so basically it's showing the
[00:19:24.720]similar kind of behavior uh current density increases first to some intermediate state
[00:19:31.600]and then starts to decrease so basically we have the point of main state corresponds to the
[00:19:37.120]highest device connectivity and very similar to this the previous device we have in the political
[00:19:46.080]crystalline case
[00:19:47.040]and in order to observe the probe the domain walk activity with local probe technique so we have to
[00:19:58.880]remove the top electrode so we have the polarized capacitor and we etched away the top electrode
[00:20:05.760]so that we can probe the local local information instead of the collective
[00:20:12.800]information when the top electrode is covered
[00:20:16.400]and of course there is some over edge of the free surface about 15 to 20 nanometers but compared to
[00:20:25.200]the thickness this action is still negligible so shown here is the removed capacitor edge and we
[00:20:37.760]have the pfm image for this was initially polarized with the downward polarization state and the free
[00:20:45.600]surface was initially polarized up and we use conductive afm imaging which probes the local
[00:20:54.720]conductivity in this domain boundary and what we see is a bright contrast
[00:21:04.000]corresponding to the domain edge
[00:21:08.140]And while on the fully polarized state,
[00:21:10.780]the signal is lower compared to the signal on the domain wall.
[00:21:16.720]And of course, the IV on the domain wall and on the domains
[00:21:22.640]are also showing similar kind of trend.
[00:21:25.360]Of course, this is a pretty thick film.
[00:21:27.700]It's generally in nanometers, so the signal is pretty low.
[00:21:32.020]It's in several picameter amps up to like 50 volts.
[00:21:38.840]So the signal is very low.
[00:21:40.480]But still, this higher contrast is very clear
[00:21:45.820]at the domain boundary.
[00:21:47.940]So this demonstrates that those head-to-head inclined
[00:21:53.300]domain walls are indeed conductive.
[00:21:56.340]And of course, we can control this.
[00:22:01.880]And domain wall conductivity by applying a local pulse.
[00:22:06.600]So shown here is a local pulse trying to erase this domain.
[00:22:12.160]And also this domain wall conductivity is erased.
[00:22:17.000]And shown here are traditional pulses
[00:22:21.760]to create local domains, upward domains.
[00:22:24.520]And shown this is a schematic cross-sectional drawing.
[00:22:31.660]So here, because of the non-uniformity
[00:22:36.520]of the electric field generated by the probe,
[00:22:39.400]we assume that those domains could be polarization up
[00:22:46.360]and down.
[00:22:47.140]So those could be like tail-to-tail domain walls.
[00:22:51.160]Of course, they may not reach to the bottom electrode
[00:22:54.040]because those domains are pretty small.
[00:22:57.340]And we know that domain walls could be inclined.
[00:23:01.440]But those domain walls are not showing conductivity.
[00:23:05.320]So implies that either tail-to-tail domain walls
[00:23:10.860]are not conductive, or those domain walls
[00:23:13.680]are just floating domain walls, and that we could not
[00:23:18.100]probe the conductivity.
[00:23:19.340]And so another material we investigate
[00:23:27.460]is the niobium oxide iodide.
[00:23:31.220]So this is a two-dimensional van der Waals ferroelectric
[00:23:36.440]material.
[00:23:37.100]It's a semiconductor.
[00:23:38.880]And we have layered structure.
[00:23:41.560]So basically, you can exfoliate this material into layer
[00:23:48.660]to layer.
[00:23:50.060]And of course, shown here are some PFM
[00:23:52.520]images on slide realistic films of some tens of nanometers.
[00:23:58.320]That's much easier for us to--
[00:24:01.000]image the domain structure and work--
[00:24:04.640]of course, also to work with.
[00:24:07.260]And what we see is that those striped domains--
[00:24:13.360]of course, the polar axis is along the B-axis.
[00:24:16.240]So it's in plane, basically.
[00:24:18.700]And so what we see is the striped domains
[00:24:22.500]with any parallel domain orientation in plane.
[00:24:26.300]However, there are also some flakes
[00:24:28.620]showing charger domain boundary.
[00:24:30.780]Like this one is basically showing tail-to-tail charger
[00:24:36.620]domain wall.
[00:24:37.860]And this is another example.
[00:24:40.240]We have those zigzag structure of domain wall
[00:24:44.480]with head-to-head domain orientation.
[00:24:48.500]So again, the polarization is in plane.
[00:24:51.420]So this provides easier configuration
[00:24:58.120]for us measuring the domain wall
[00:25:00.560]connectivity.
[00:25:01.800]So basically, we just need the connective AFM image directly
[00:25:06.820]on the surface.
[00:25:08.420]And shown here is the result that we pick this head-to-head
[00:25:14.240]domain wall with those zigzag domain structure.
[00:25:17.360]And the corresponding connective AFM image
[00:25:20.760]showing it's like much higher connectivity around those
[00:25:25.940]head-to-head charge domain walls compared to the--
[00:25:30.340]the single polarized state.
[00:25:33.660]So this means that these domain walls are, again, also
[00:25:37.680]connective.
[00:25:38.940]And given that this is a two-dimensional material,
[00:25:42.640]there is also a possibility that we can further exfoliate down
[00:25:49.680]to a monolayer limit.
[00:25:52.480]And potentially, you can create like one-dimensional connective
[00:25:57.220]interfaces with this material, which is also interesting.
[00:26:00.120]And so, yeah, so that's pretty much for our recent work.
[00:26:09.360]And as a conclusion, basically, we
[00:26:14.760]directly confirmed the domain wall connectivity
[00:26:17.940]in the aluminum-scanning nitrite.
[00:26:21.060]And we tried to use such domain wall connective--
[00:26:24.560]connective domain walls to create
[00:26:26.340]a memristor device based on different polar
[00:26:29.900]polarization states or based on the density of the domain walls.
[00:26:34.820]And another work is that we also observed the domain wall
[00:26:40.400]connectivity in the 2D van der Waals
[00:26:42.920]niobium oxide iodide.
[00:26:47.680]So with this material, there is additional chance
[00:26:54.080]that we can--
[00:26:57.020]this is a semi-conducting material, so basically,
[00:26:59.680]we can do additional studies on, say,
[00:27:05.040]light illumination or some further investigations.
[00:27:09.880]And yeah, so with that, that's pretty much my presentation.
[00:27:16.320]And I would like to take any questions.
[00:27:20.820]Yeah.
[00:27:22.080]Thank you, Haidang.
[00:27:23.180]If anybody has questions, you can ask.
[00:27:27.080]Haidang, I have a question.
[00:27:28.460]Yeah.
[00:27:29.460]About the niobium oxide iodide.
[00:27:31.720]So this domain wall you are showing,
[00:27:34.440]they are natural domain wall, right?
[00:27:35.820]So it's naturally formed.
[00:27:37.500]So you basically look for if you have a straight or zigzag
[00:27:43.440]shape to determine if it's neutral or charged.
[00:27:47.080]Is that right?
[00:27:48.560]Yeah.
[00:27:49.060]So those parallel domains are most frequently observed.
[00:27:55.800]But there is also a chance that we can see--
[00:27:59.240]this kind of head-to-head or tail-to-tail domain wall.
[00:28:04.760]Yeah, these are all S-ground domain walls.
[00:28:07.100]But of course, we can electrically
[00:28:09.420]switch the polarization.
[00:28:11.420]So there are some issues with electrical switching,
[00:28:15.260]because usually we work with pretty thick dump.
[00:28:18.920]And those electrically created domains are relatively shallow.
[00:28:23.660]It's most likely not going to the--
[00:28:29.020]the bottom interface.
[00:28:30.760]So measuring the main wall connectivity
[00:28:33.340]might require much thinner films.
[00:28:36.680]So we haven't investigated those thin films yet.
[00:28:42.340]So these are relatively thick.
[00:28:45.760]OK.
[00:28:46.320]So basically, that is a paradox you were saying,
[00:28:50.140]like if you can electrically perturb your stomach wall.
[00:28:53.680]Is that an idea?
[00:28:55.220]Yeah.
[00:28:55.720]So yeah, we can successfully--
[00:28:58.800]create those head-to-head or total charge domain wall
[00:29:02.360]electrically, yes.
[00:29:04.180]OK, that would be cool.
[00:29:05.320]So is that-- I mean, but is that something by chance?
[00:29:10.420]So you actually can intentionally
[00:29:12.940]create one type domain wall versus the other type?
[00:29:16.040]We can create those domain walls intentionally anywhere
[00:29:21.460]and just need to apply the voltage.
[00:29:25.460]So basically, the tip voltage--
[00:29:28.580]of course, there is a non-uniform electric field.
[00:29:31.680]So it can still switch the in-plane polarization.
[00:29:35.820]It switches the polarization oppositely on the tip.
[00:29:42.800]So you can create head-to-head walls
[00:29:46.940]once the tip can draw a line across the ..
[00:29:51.940]But you also have shown those--
[00:29:58.360]neutral type domain wall.
[00:30:00.940]Is that something you can also write?
[00:30:04.120]Yeah, so neutral domain wall, so yeah, these are 180 degree
[00:30:12.060]neutral domain walls.
[00:30:13.020]And so we don't see domain wall connectivity
[00:30:17.380]for those neutral domain walls.
[00:30:21.360]OK, so when you're basically using field to write,
[00:30:24.860]you only see charged domain wall.
[00:30:28.140]So basically, on the side, you have neutral domain wall,
[00:30:32.560]and across the written line, you have charged domain wall.
[00:30:36.120]You can create both.
[00:30:37.360]Sorry, I was like, kind of, Mishra?
[00:30:44.920]Mishra, can you mute yourself?
[00:30:47.260]Sorry.
[00:30:48.760]Oh, OK.
[00:30:49.660]Yeah, yeah.
[00:30:50.480]So basically, you write a line, you
[00:30:54.260]create the charged domain wall along the line, but you also
[00:30:57.920]have neutral domain wall perpendicular to this written
[00:31:03.360]line, or neutral domain wall along this polar axis.
[00:31:08.700]OK, so it depends on the personal orientation,
[00:31:11.400]how you write it, along B or along--
[00:31:12.920]Yeah, yeah.
[00:31:13.600]OK, I see.
[00:31:14.840]Great, thank you.
[00:31:16.640]Yeah.
[00:31:17.140]Nice talk.
[00:31:25.420]I have a question.
[00:31:27.700]If one wants to really quantify the domain wall connectivity
[00:31:33.560]or resistance, it would be nice to know the--
[00:31:38.840]more than just the number of domains you create,
[00:31:41.340]but the actual area of the domain walls
[00:31:45.940]to come to a specific property of the wall.
[00:31:49.860]Is that something you can possibly quantify?
[00:31:54.420]Yeah, so there's some--
[00:31:57.480]further work.
[00:32:00.080]So what we can do is that we can basically place electrodes
[00:32:07.500]on top so that we can measure a more quantitative domain wall,
[00:32:14.400]like transport behavior.
[00:32:17.080]So like here, if we have this kind of domain wall,
[00:32:20.860]we can place two parallel electrodes, but this requires--
[00:32:27.260]requires some device fabrication.
[00:32:32.260]Yeah, so it's-- yeah, so far, we haven't
[00:32:37.660]been able to fabricate such devices.
[00:32:41.680]So it will be a good plan to further investigation
[00:32:52.080]of some further quantification of the transport behavior.
[00:32:57.040]I don't-- I have a question for the nominal scandium nitride
[00:33:15.860]part.
[00:33:16.360]Yeah.
[00:33:16.860]The single crystal, yes.
[00:33:20.580]I suppose the film has the c-axis pointing out
[00:33:24.080]of the planet, right?
[00:33:24.860]Right, right.
[00:33:25.620]So--
[00:33:26.120]Yeah.
[00:33:26.820]So the polarization is around c-axis,
[00:33:29.240]and the c-axis is out of the plane, yeah.
[00:33:33.560]So why is this tilted domain wall formed?
[00:33:37.400]I suppose the domain wall to be along the c-axis
[00:33:41.900]or something like that?
[00:33:45.060]Yeah.
[00:33:47.440]So yeah, this is very clear.
[00:33:51.780]So most likely, from their TM observations,
[00:33:56.600]those domain wall start to grow at certain interface.
[00:34:04.100]And once it's nucleate, it's like a triangular shape
[00:34:11.700]of the domains.
[00:34:13.560]And those domains start to grow toward the other interface.
[00:34:19.340]But during the switching process,
[00:34:22.540]those domain walls are still inclined.
[00:34:25.160]So they're still charged.
[00:34:26.380]And the thing here is that those inclined domain walls
[00:34:30.880]are pretty stable.
[00:34:32.860]So even with the TM imaging, they're still stably present.
[00:34:38.980]So this is probably a property of the material itself.
[00:34:44.620]Because earlier studies have shown
[00:34:48.320]that those charged domain walls can be stably
[00:34:53.420]present in the material, even with very thin
[00:34:56.160]films.
[00:34:56.660]This is only 4 nanometers.
[00:34:58.660]And there is 180 degree head-to-head or tail-to-tail
[00:35:03.800]domain.
[00:35:04.300]And this one is also highly inclined.
[00:35:06.220]So those domain walls are pretty stable.
[00:35:09.560]Yeah, usually for conventional ferroelectrics,
[00:35:15.920]those charged domain walls wouldn't be very stable.
[00:35:19.920]They would prefer to be neutral.
[00:35:22.380]But somehow, in this material, we
[00:35:25.940]have inclined stable domains.
[00:35:30.200]So from a charge point of view, it shouldn't be stable.
[00:35:37.180]And also from a structural point of view,
[00:35:41.360]I don't see any obvious reason that this tilted domain wall
[00:35:47.400]is stable.
[00:35:48.340]So is this something still people do not understand,
[00:35:52.840]or there's some theory that can
[00:35:55.720]explain this?
[00:35:57.940]We don't have a theory yet.
[00:36:02.360]Professor Belashchenko's group and his former student
[00:36:06.640]is working on the modeling.
[00:36:12.080]OK, I see.
[00:36:14.220]All right, all right, thank you.
[00:36:15.520]Hi, Han Dong.
[00:36:20.760]Thank you for your presentation.
[00:36:22.680]I also have one question for you.
[00:36:25.500]Regarding charges which are responsible for conductivity
[00:36:31.720]of these domain walls, so I guess
[00:36:33.720]depending on whether you have head-to-head or tail-to-tail
[00:36:37.600]domain wall, you would have positive or negative charges
[00:36:41.140]to compensate polarization charge, right?
[00:36:43.520]Yeah, right.
[00:36:44.620]Depending on this, you probably should
[00:36:46.200]have different type of conductivity, E-type or M-type.
[00:36:49.900]Yeah.
[00:36:51.300]And the first question is, and I'm
[00:36:52.880]doing orders and a significant difference between--
[00:36:55.280]I mean, if you have tail-to-tail or head-to-head domain wall,
[00:37:01.840]can you distinguish between the two?
[00:37:03.920]Is a significant difference in conductivity?
[00:37:06.980]And the second question, can you really
[00:37:10.520]quantify whether it's P-type or M-type conductivity?
[00:37:16.580]Yeah, so first question, in the aluminum scandium nitride--
[00:37:21.920]so here, those domain walls are--
[00:37:25.060]are predominantly head-to-head.
[00:37:28.840]So basically, communication always
[00:37:32.740]starts from the same interface.
[00:37:35.080]So we always have head-to-head oriented domain walls.
[00:37:41.700]Of course, in this schematic drawing,
[00:37:47.320]so once we apply a pulse from the top,
[00:37:51.300]we assume that the electric field is much stronger
[00:37:54.840]at the top interface, so there may
[00:37:57.420]be a chance to create tail-to-tail domain walls.
[00:38:03.620]But here, of course, we don't know
[00:38:06.080]if those domain walls are going all the way down
[00:38:09.700]to the bottom interface, because we can only
[00:38:13.200]create small domains.
[00:38:15.780]If we apply much higher voltage, there
[00:38:18.540]is a change to the top interface.
[00:38:20.960]So we cannot really create large domains.
[00:38:24.620]So--
[00:38:25.620]Excuse me.
[00:38:26.620]For the other material, you have
[00:38:29.500]shown tail-to-tail and head-to-head domain wall.
[00:38:35.420]For which one?
[00:38:37.000]This one?
[00:38:38.000]Yeah, for this one.
[00:38:39.020]Yeah.
[00:38:39.920]Yeah.
[00:38:40.420]So for this one, yeah, it would be
[00:38:43.120]good to investigate the direct evidence of the difference
[00:38:49.180]for head-to-head and tail-to-tail domain wall.
[00:38:52.180]So however, this--
[00:38:54.400]this investigation must be done on the same flake
[00:38:58.520]with the same thickness or same conditions.
[00:39:02.520]And this is quite challenging for us
[00:39:07.280]to find a flake with both head-to-head and tail-to-tail
[00:39:10.640]domain walls.
[00:39:11.940]Basically, on different flakes, we
[00:39:14.480]can see connectivity for both head-to-head and tail-to-tail.
[00:39:18.820]But a quantitative comparison will be still
[00:39:22.120]difficult at that point.
[00:39:24.180]So both are conductive.
[00:39:25.320]So you see that both are conductive.
[00:39:27.300]Yeah.
[00:39:27.900]And about the second question, so we
[00:39:31.600]need transport measurement, detailed transport measurement.
[00:39:35.100]So, so far, we don't have further information
[00:39:39.020]on the type of charge carrier.
[00:39:42.140]So we can only guess that the host material, if it's N-doped,
[00:39:47.600]we assume that it could be like free electrons.
[00:39:53.960]Otherwise, we don't have any evidence whether it's N-type
[00:39:57.300]or P-type conducting.
[00:40:00.220]So this requires some transport measurement
[00:40:03.680]with device structure.
[00:40:06.880]Thank you.
[00:40:08.160]One more question I have regarding
[00:40:10.320]the memoristic behavior.
[00:40:13.880]So when you are doing this intermediate configurations
[00:40:19.120]of domains, in order to have this numeristic performance,
[00:40:22.720]I guess, of this--
[00:40:23.740]the main configuration should be stable.
[00:40:26.560]So if you wait for some time, is that stable or it changes?
[00:40:33.560]The domains are pretty stable.
[00:40:35.420]So we have looked at the domains one day after polling,
[00:40:43.760]and we can still see the domains.
[00:40:46.700]They do have some relaxation for this connectivity state.
[00:40:51.740]Possibly, there is--
[00:40:53.520]like, a relaxation of the inclination angle,
[00:40:56.880]or there could be some relaxation.
[00:41:02.260]So this state, there is a small relaxation
[00:41:07.620]initially, and then it becomes stable.
[00:41:11.580]Yeah, so probably the main state, it's pretty stable.
[00:41:16.920]So we can keep this state for days, at least.
[00:41:21.760]Thank you.
[00:41:24.960]So does anybody has any other questions?
[00:41:44.360]I think I have the same, so let's take our speaker again.
[00:42:03.560]Yeah, all right, thank you.
[00:42:04.860]Thank you.

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.

Comments

0 Comments

Conductive Domain Walls and Resistive Switching in the Ferroelectric AlScN and NbOI2

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments