Optoelectrical Molybdenum Disulfide (MoS2) â Ferroelectric Memories
SUPPORTING INFORMATION
Optoelectrical Molybdenum Disulfide (MoS 2) – Ferroelectric Memories Alexey Lipatov,† Pankaj Sharma,‡ Alexei Gruverman,‡,§ and Alexander Sinitskii*,†,§ †
Department of Chemistry, ‡Department of Physics and Astronomy, and §Nebraska Center for
Materials and Nanoscience, University of Nebraska, Lincoln, Nebraska 68588, United States *E-mail: [email protected]
1
Supporting Figure 1. Monolayer and bilayer MoS2-PZT FeFETs. (a,b) SEM images of the devices that were fabricated using monolayer (a) and bilayer (b) MoS2 flakes. Insets show optical images of the original MoS2 flakes on Si/SiO2. (c,d) Raman spectra of the MoS2 flakes shown in (a) and (b), respectively. (e,f) IDS - VG characteristics of the MoS2-PZT FeFETs shown in (a) and (b), respectively. Both FeFETs exhibit clockwise hysteresis loops.
2
Supporting Figure 2. Switching of ferroelectric polarization in the PZT film. (a) Polarization hysteresis loop measured in the PZT film with the SrRuO3 top electrode. (b) Topography AFM image of the PZT surface. (c) Local PFM spectroscopic phase and amplitude hysteresis loops. (d,e) PFM amplitude (d) and phase (e) images obtained on the bare PZT surface after poling it with a DC bias of + 5 and -5 V (rectangular stripes).
3
Supporting Figure 3. Electronic properties of a trilayer MoS2-PZT FeFET. In all measurements shown VDS = 0.1 V. (a) IDS-VG characteristics of the device shown in Figure 1e. The data are shown in linear (upper panel) and logarithmic (lower panel) current coordinates. (b) VG dependence of the leakage current for the same device. (c) IDS-VG characteristics of the same MoS2-PZT FeFET measured at different VG sweep rates. (d) Comparison of IDS-VG dependencies for the same device measured in the dark and under visible light illumination.
4
Supporting Figure 4. Electronic properties of FeFETs based on (a) monolayer, (b) bilayer and (c) trilayer MoS2 flakes. The IDS-VG characteristics of the devices shown in linear (top panels) and logarithmic (bottom panels) IDS coordinates. Two types of measurements are shown. Black data points show IDS values that were measured while VG was applied. Colored data points were collected in accordance with the measurement scheme shown in Figure 2c at various τwait, see text for details. The arrows indicate the directions of hysteresis loops.
5
Supporting Figure 5. Switching of the ferroelectric polarization in MoS2/PZT. (a) Measurements of switching current as a function bias applied to the PZT film through the MoS2 flake. (b) Local spectroscopic PFM phase and amplitude hysteresis loops measured on the MoS2 flake. (c,d) PFM phase images obtained after application of (-5 V, 1s) pulse (c) and (+5 V, 1s)
pulse (d) to the PZT film through the MoS2 flake. Dashed lines in images (c) and (d) show
a boundary of the MoS2 flake.
6
Supporting Note 1. Polarization Switching in the PZT Film Supporting Figure 2 shows results of electrical testing of the PZT film. A polarization loop (Supporting Figure 2a) measured in the film with the deposited SrRuO3 top electrode by the Sawyer-Tower method illustrates its symmetric switching behavior. Local PFM spectroscopic loops (Supporting Figure 2c) also reveal a hysteretic behavior that is typical for ferroelectric materials. PFM phase loop shows a phase change of ~ 180° indicative of the polarization reversal, while the PFM amplitude response displays a butterfly curve, which is symmetric with respect to the sign of the applied bias. A difference in the coercive biases detected by the Sawyer-Tower and PFM methods is due to the high tip-sample contact resistance in PFM. Supporting Figure 2d,e shows the spatially uniform switching of polarization in the PZT film by scanning its surface in the PFM contact mode under an electrical bias above the coercive value. Rectangular 2×1 µm2 regions have been poled and subsequently imaged using PFM. The directions of the polarization (left stripe – polarization down, right stripe – polarization up) have been indicated in the electrically-modified areas (Supporting Figure 2e). In addition, the PFM phase image (Supporting Figure 2e) also clearly identifies that the as-grown (outside the electrically written regions) state of the PZT has a polarization pointing downwards. It should be noted that in PFM experiments the tip acts as a top electrode, while in FET measurements the voltage is applied to the bottom gate electrode. Thus, when a positive voltage is applied in PFM experiments the polarization of PZT is downward, as shown in Supporting Figure 2d,e. On the other hand, when the same positive voltage is applied to the bottom gate electrode of a FET, the polarization of PZT is upward, as shown in Figure 2a.
7
Supporting Note 2. Relaxation of the IDS in MoS2/PZT FeFETs While τwait = 20 sec is sufficient to reverse the MoS2 conductivity hysteresis (see Supporting Figure 4a), it is unclear to which extent the interfacial charges have dissipated during that period of time. To answer this question, we have used the same measurement procedure (Figure 2c) to study the electronic behavior of the same MoS2-PZT FeFET at different τwait. Supporting Figure 4 shows a series of polarization-controlled conductivity hysteresis loops measured with τwait ranging from 20 to 120 sec with a 20 sec increment for monolayer and bilayer MoS2 devices, and from 5 to 30 min with a 5 min increment for a trilayer MoS2 device. These data show that some residual charges remain at the MoS2/PZT interface even after τwait = 5 min, as longer τwait results in further evolution of the hysteresis’ shape. However, these changes are rather small compared to the original transformation of the hysteresis that occurred in the first 5 min.
8
Supporting Note 3. Switching of the ferroelectric polarization in MoS2/PZT Effective control of the ferroelectric polarization of the PZT film by means of an electrical bias applied to the MoS2 flake on its surface using a conductive AFM tip has been verified by measuring the switching transient current (across the loading resistor in series with the film), and visualizing the remnant state of polarization via PFM. Supporting Figure 5a shows current as a function of DC bias measured using spectroscopic conducting-AFM (C-AFM). The presence of the peaks (in both forward and reverse directions) in the measured current-bias curves (Supporting Figure 5a) reveals switching current associated with reversal of the ferroelectric (PZT) polarization. Local PFM hysteresis loop measurements (Supporting Figure 5b) reveal asymmetric switching behavior due to the semiconducting nature of the top electrode (MoS2). Finally, Supporting Figure 5c,d illustrates the reversal of ferroelectric polarization of the PZT film under the MoS2 flake. As can be seen in Supporting Figure 5c,d, application of electrical bias of opposite polarity to the MoS2 flake results in reorientation of ferroelectric polarization in the PZT area covered by the flake, which is evidenced by the PFM phase change of 180°, while the area outside the flake (bare PZT surface) remains unperturbed.
9
Optoelectrical Molybdenum Disulfide (MoS 2) – Ferroelectric Memories Alexey Lipatov,† Pankaj Sharma,‡ Alexei Gruverman,‡,§ and Alexander Sinitskii*,†,§ †
Department of Chemistry, ‡Department of Physics and Astronomy, and §Nebraska Center for
Materials and Nanoscience, University of Nebraska, Lincoln, Nebraska 68588, United States *E-mail: [email protected]
1
Supporting Figure 1. Monolayer and bilayer MoS2-PZT FeFETs. (a,b) SEM images of the devices that were fabricated using monolayer (a) and bilayer (b) MoS2 flakes. Insets show optical images of the original MoS2 flakes on Si/SiO2. (c,d) Raman spectra of the MoS2 flakes shown in (a) and (b), respectively. (e,f) IDS - VG characteristics of the MoS2-PZT FeFETs shown in (a) and (b), respectively. Both FeFETs exhibit clockwise hysteresis loops.
2
Supporting Figure 2. Switching of ferroelectric polarization in the PZT film. (a) Polarization hysteresis loop measured in the PZT film with the SrRuO3 top electrode. (b) Topography AFM image of the PZT surface. (c) Local PFM spectroscopic phase and amplitude hysteresis loops. (d,e) PFM amplitude (d) and phase (e) images obtained on the bare PZT surface after poling it with a DC bias of + 5 and -5 V (rectangular stripes).
3
Supporting Figure 3. Electronic properties of a trilayer MoS2-PZT FeFET. In all measurements shown VDS = 0.1 V. (a) IDS-VG characteristics of the device shown in Figure 1e. The data are shown in linear (upper panel) and logarithmic (lower panel) current coordinates. (b) VG dependence of the leakage current for the same device. (c) IDS-VG characteristics of the same MoS2-PZT FeFET measured at different VG sweep rates. (d) Comparison of IDS-VG dependencies for the same device measured in the dark and under visible light illumination.
4
Supporting Figure 4. Electronic properties of FeFETs based on (a) monolayer, (b) bilayer and (c) trilayer MoS2 flakes. The IDS-VG characteristics of the devices shown in linear (top panels) and logarithmic (bottom panels) IDS coordinates. Two types of measurements are shown. Black data points show IDS values that were measured while VG was applied. Colored data points were collected in accordance with the measurement scheme shown in Figure 2c at various τwait, see text for details. The arrows indicate the directions of hysteresis loops.
5
Supporting Figure 5. Switching of the ferroelectric polarization in MoS2/PZT. (a) Measurements of switching current as a function bias applied to the PZT film through the MoS2 flake. (b) Local spectroscopic PFM phase and amplitude hysteresis loops measured on the MoS2 flake. (c,d) PFM phase images obtained after application of (-5 V, 1s) pulse (c) and (+5 V, 1s)
pulse (d) to the PZT film through the MoS2 flake. Dashed lines in images (c) and (d) show
a boundary of the MoS2 flake.
6
Supporting Note 1. Polarization Switching in the PZT Film Supporting Figure 2 shows results of electrical testing of the PZT film. A polarization loop (Supporting Figure 2a) measured in the film with the deposited SrRuO3 top electrode by the Sawyer-Tower method illustrates its symmetric switching behavior. Local PFM spectroscopic loops (Supporting Figure 2c) also reveal a hysteretic behavior that is typical for ferroelectric materials. PFM phase loop shows a phase change of ~ 180° indicative of the polarization reversal, while the PFM amplitude response displays a butterfly curve, which is symmetric with respect to the sign of the applied bias. A difference in the coercive biases detected by the Sawyer-Tower and PFM methods is due to the high tip-sample contact resistance in PFM. Supporting Figure 2d,e shows the spatially uniform switching of polarization in the PZT film by scanning its surface in the PFM contact mode under an electrical bias above the coercive value. Rectangular 2×1 µm2 regions have been poled and subsequently imaged using PFM. The directions of the polarization (left stripe – polarization down, right stripe – polarization up) have been indicated in the electrically-modified areas (Supporting Figure 2e). In addition, the PFM phase image (Supporting Figure 2e) also clearly identifies that the as-grown (outside the electrically written regions) state of the PZT has a polarization pointing downwards. It should be noted that in PFM experiments the tip acts as a top electrode, while in FET measurements the voltage is applied to the bottom gate electrode. Thus, when a positive voltage is applied in PFM experiments the polarization of PZT is downward, as shown in Supporting Figure 2d,e. On the other hand, when the same positive voltage is applied to the bottom gate electrode of a FET, the polarization of PZT is upward, as shown in Figure 2a.
7
Supporting Note 2. Relaxation of the IDS in MoS2/PZT FeFETs While τwait = 20 sec is sufficient to reverse the MoS2 conductivity hysteresis (see Supporting Figure 4a), it is unclear to which extent the interfacial charges have dissipated during that period of time. To answer this question, we have used the same measurement procedure (Figure 2c) to study the electronic behavior of the same MoS2-PZT FeFET at different τwait. Supporting Figure 4 shows a series of polarization-controlled conductivity hysteresis loops measured with τwait ranging from 20 to 120 sec with a 20 sec increment for monolayer and bilayer MoS2 devices, and from 5 to 30 min with a 5 min increment for a trilayer MoS2 device. These data show that some residual charges remain at the MoS2/PZT interface even after τwait = 5 min, as longer τwait results in further evolution of the hysteresis’ shape. However, these changes are rather small compared to the original transformation of the hysteresis that occurred in the first 5 min.
8
Supporting Note 3. Switching of the ferroelectric polarization in MoS2/PZT Effective control of the ferroelectric polarization of the PZT film by means of an electrical bias applied to the MoS2 flake on its surface using a conductive AFM tip has been verified by measuring the switching transient current (across the loading resistor in series with the film), and visualizing the remnant state of polarization via PFM. Supporting Figure 5a shows current as a function of DC bias measured using spectroscopic conducting-AFM (C-AFM). The presence of the peaks (in both forward and reverse directions) in the measured current-bias curves (Supporting Figure 5a) reveals switching current associated with reversal of the ferroelectric (PZT) polarization. Local PFM hysteresis loop measurements (Supporting Figure 5b) reveal asymmetric switching behavior due to the semiconducting nature of the top electrode (MoS2). Finally, Supporting Figure 5c,d illustrates the reversal of ferroelectric polarization of the PZT film under the MoS2 flake. As can be seen in Supporting Figure 5c,d, application of electrical bias of opposite polarity to the MoS2 flake results in reorientation of ferroelectric polarization in the PZT area covered by the flake, which is evidenced by the PFM phase change of 180°, while the area outside the flake (bare PZT surface) remains unperturbed.
9
