Memristive Behavior and Ideal Memristor of 1T Phase MoS2

Dec 14, 2015 - Memristor, which had been predicted a long time ago (Chua, L. O. IEEE Trans. Circuit Theory 1971, 18, 507), was recently invented (Stru...
2 downloads 0 Views 3MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Memristive Behavior and Ideal Memristor of 1T Phase MoS2 Nanosheets Peifu Cheng,† Kai Sun,‡ and Yun Hang Hu*,† †

Department of Materials Science and Engineering, Michigan Technological University, Houghton, Michigan 49931-1295, United States ‡ Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan 48109-2136, United States S Supporting Information *

ABSTRACT: Memristor, which had been predicted a long time ago (Chua, L. O. IEEE Trans. Circuit Theory 1971, 18, 507), was recently invented (Strukov, D. B.; et al. Nature 2008, 453, 80). The introduction of a memristor is expected to open a new era for nonvolatile memory storage, neuromorphic computing, digital logic, and analog circuit. Furthermore, several breakthroughs were made for memristive phenomena and transistors with single-layer MoS2 (Sangwan, V. K.; et al. Nat. Nanotechnol. 2015, 10, 403. van der Zande, A. M.; et al. Nat. Mater. 2013, 12, 554. Liu, H.; et al. ACS Nano 2014, 8, 1031. Bessonov, A. A.; et al. Nat. Mater. 2015, 14, 199. Yuan, J.; et al. Nat. Nanotechnol. 2015, 10, 389). Herein, we demonstrate that 2H phase of bulk MoS2 possessed an ohmic feature, whereas 1T phase of exfoliated MoS2 nanosheets exhibited a unique memristive behavior due to voltagedependent resistance change. Furthermore, an ideal odd-symmetric memristor with odd-symmetric I−V characteristics was successfully fabricated by the 1T phase MoS2 nanosheets via combining two asymmetric switches antiserially. KEYWORDS: Memristor, MoS2, 1T metallic phase, odd-symmetric

T

MoS2 has two main types of phases: 2H phase, in which two layers per unit cell stack in the hexagonal symmetry with trigonal prismatic coordination, and 3R phase, in which three layers per unit cell stack in the rhombohedral symmetry with trigonal prismatic coordination.17 The 2H phase, which is more stable than 3R phase, is dominant in nature. Furthermore, besides 2H and 3R phases, a metastable 1T metallic phase, which possesses one-layer unit cell in the tetragonal symmetry with octahedral coordination, 18 was observed in MoS2 nanosheets prepared by intercalation-assisted-exfoliation. However, 1T phase can be easily transformed to 2H phase under moderate-temperature (above 200 °C) annealing.19 Furthermore, Shirodkar et al. theoretically predicted the emergence of ferroelectricity at a metal−semiconductor transition in a 1Ttype monolayer MoS2.20 In addition, 1T phase MoS2 exhibited excellent performance for supercapacitors.21 On the other hand,

he discovery of graphene and the exploration of its superior properties have led to the identification of many other layer materials.8,9 One of them is transition metal dichalcogenides (TMDs).10 The most prominent TMDs are MoS2, WS2, MoSe2, and WSe2 that have been widely studied for applications in electronics due to their semiconductive properties. Those materials present a MX2 stoichiometry and are composed of 2D sheets stacked on top of one another. Each sheet is three atoms thick with a sandwich structure in which a metal atom at the middle is strongly bonded to above and below dichalcogenide atoms. The intralayer metal−chalcogen bond is predominantly covalent, while the sheets are held together by weak van der Waals interlayer interactions.10,11 The band gaps of these materials can be changed by stacking confinement and strong electric fields.12 Among those 2D materials, monolayer molybdenum disulfide (MoS2) is a rising star as a candidate for nanoelectronics and optoelectronics owing to its unique band gap, high mobility, and quantum confinement.4,5,7,10,13−16 © XXXX American Chemical Society

Received: October 19, 2015 Revised: December 12, 2015

A

DOI: 10.1021/acs.nanolett.5b04260 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. Characterization of exfoliated MoS2. (a) FE-SEM morphology image of exfoliated MoS2 nanosheets. (b) TEM image of MoS2 nanosheets. (c) XRD patterns of MoS2 nanosheets (red) and bulk MoS2 powders (black). (d) AFM image and corresponding height profile of MoS2 nanosheets (on cleaned silica glass substrate). (e, f) Raman spectra of MoS2 nanosheets (red) and bulk MoS2 powders (black) using 632.8 nm laser line. (g) HAADF-STEM image of MoS2 nanosheets (blue dots displayed by Mo atoms exhibit their hexagonal lattice arrangement). (h) XPS spectra of Mo 3d species in MoS2 nanosheets. (i) XPS spectra of S 2p species in MoS2 nanosheets.

of all other peaks confirms the exfoliated structure, and the very small retained broad peak (002) is due to the presence of fewlayered MoS2. Furthermore, the atomic force microscopy (AFM) image shows its thickness of 2 nm (Figure 1d), indicating that it contains three layers since the thickness of a single-layer MoS2 is ∼0.65 nm.10,23 The structure of the MoS2 nanosheets was further evaluated by Raman spectra. Figure 1, panel e shows in-plane E12g mode, which is resulted from opposite vibration of two S atoms with respect to the Mo atom,24 and out-of-plane A1g modes, which are associated with the vibration of only S atoms in opposite directions,24 for both bulk and nanosheet MoS2. However, the exfoliation of bulk MoS2 to nanosheets led to a frequency increase for the E12g mode (from 375.9 to 377.8 cm−1) and a decrease for the A1g mode (from 402.8 to 401.8 cm−1), which were also observed for four-layer MoS2.24 Furthermore, an E1g peak at 280 cm−1, which is associated with the octahedral coordination of Mo atoms in the 1T phase of MoS2,25 was observed in the Raman spectrum of exfoliated MoS2 (Figure 1f). This indicates the presence of 1T phase in the exfoliated MoS2, which is further supported by the observation of other active modes (J1, J2, and

Hersam et al. recently revealed that grain boundaries in singlelayer MoS2 generated the memristive phenomena.3 Furthermore, Bessonov et al. demonstrated that the MoOx/MoS2 heterostructure created the tunable electrical resistance from 102 to 108 Ω.6 However, a phase-dependent memristive phenomenon has not been observed for MoS2. Herein, we found that the 1T phase of exfoliated MoS2 nanosheets exhibited the memristive behavior. Furthermore, we designed and successfully fabricated an ideal memristor with two 1TMoS2/Ag active layers antiserially, which exhibited oddsymmetric I−V characteristics. We exfoliated MoS2 bulk material with the method that was developed by Morrison et al.22 with a slight modification. The nanosheet structure of highly exfoliated MoS2 was observed by field emission scanning electron microscopy (FE-SEM) (Figure 1a) and transmission electron microscopy (TEM) (Figure 1b), which was further confirmed by disappearance of diffraction peaks in X-ray diffraction (XRD) pattern (Figure 1c). Different from bulk MoS2 crystal, the exfoliated MoS2 showed only one very small broad-diffraction peak corresponding to (002) plane of 2H-type MoS2 (JCPDS file number: 37−1492). The absence B

DOI: 10.1021/acs.nanolett.5b04260 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. I−V characteristic of MoS2 nanosheets based switch. (a) Schematic structure of Ag/MoS2/Ag switch. (b) Typical I−V characteristic of Ag/ MoS2/Ag switch at room temperature. (c) Typical I−V characteristic of Ag/MoS2/Ag switch in the logarithmic scale at room temperature. (d) Typical I−V characteristic of Ag/MoS2/Ag switch at the 1st (red), 500th (green), and 1000th (blue) cycle at room temperature.

J3 peaks at 145.2, 234.1, and 336.1 cm−1, respectively) being attributed to a zone-folding mechanism due to the formation of a superlattice.25 To unambiguously verify the 1T-type phase structure of the exfoliated MoS2, the atomic resolution images were obtained with high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM). As shown in Figure 1, panel g, one can see a clear hexagonal lattice, which is a direct evidence of 1T crystal phase.19,21 X-ray photoelectron spectroscopy (XPS) is an efficient technique to evaluate the composition of 1T and 2H for 2D MoS2 materials.19,21 The quantitative analysis of 1T phase in the exfoliated MoS2 was carried out with XPS (Figure 1h,i). Namely, the binding energies of Mo 3d5/2 and Mo 3d3/2 are 228.25 and 231.35 eV for 1T-phase and 229.15 and 232.25 eV for 2H phase, respectively (Figure 1h). Similarly, the S 2p spectra can also be deconvoluted into four peaks, which are assigned to 1T-type MoS2 (S 2p3/2 at 161.0 eV and S 2p1/2 at 162.2 eV) and 2Htype MoS2 (S 2p3/2 at 161.8 eV and S 2p1/2 at 163.0 eV). Furthermore, the composition (92.5% 1T phase and 7.5% 2H phase) obtained from those XPS peaks indicated that the exfoliated MoS2 mainly consists of 1T phase MoS2 nanosheets. To fabricate a MoS2/Ag active layer, the obtained 1T phase MoS2 nanosheets were dispersed in 2-propanol solvent at room temperature and then spin-coated on the surface of a silver foil, followed by heating at 130 °C for 12 h. The MoS2 surface of the obtained MoS2/Ag device was further coated with a thin layer of silver conductive paste, followed by curing at 100 °C for 1 h. The obtained Ag/MoS2/Ag device was subjected to I− V measurements (Figure 2a). During a continuous sweeping of a bias voltage from 0 mV → 200 mV → −200 mV → 0 mV, a pinched hysteresis loop was obtained (Figure 2b). Initially, the current was relatively low due to the relative high resistance (OFF state). However, the current suddenly increased at bias voltage of 66 mV, indicating a switch to a low resistance state

(ON state). A subsequent sweeping bias (to 200 mV and then back to negative values) led to the device switched back to the OFF state at bias voltage of −98 mV. The alternation between ON and OFF states and the feature of a pinched hysteresis loop constitute an asymmetric switch. The logarithmic scale shown in Figure 2, panel c indicated an ON/OFF ratio with three orders of magnitude. The negligible change during 1000 cycles of ON/OFF indicates that 1T phase MoS2 nanosheets possess excellent memristive behavior (Figure 2d). In contrast, a switching behavior was not observed (Supporting Information Figure S1) for the similar device fabricated with 2H phase bulk MoS2 (Supporting Information Figure S2). Therefore, we can conclude that the 1T phase MoS2 plays a critical role in its memristive behavior. The 1T MoS2 phase is 107-times more conductive than 2H phase.21 It was revealed from density functional theory (DFT) calculations that both valence (VB) and conduction bands (CB) in 1H phase of MoS2 (with trigonal prismatic coordination of Mo atoms) are mainly composed of Mo dx2−y2, dz2, and S px, py orbitals, generating a VB−CB energy gap of about 1.7 eV.26 In contrast, 1T phase is a distorted structure with octahedral coordination of Mo atoms. The distortion can cause a strong interaction of Mo dz2, dxy, and pz orbitals with S pz and px orbitals to create the hybridization of orbitals, leading to the overlapping of VB and CB without a gap. As a result, electrons are delocalized to stabilize the structure of 1T phase MoS2, leading to its metallic feature. Such a metastable phase possesses a changeable structure. If an external electrical field is high enough to cause the displacement of Mo and S ions in 1T phase, a lattice distortion should occur. Such an electrical-fieldinduced distortion could enhance the delocalization of electrons and thus significantly increase the conductivity. This can explain why the state switch of OFF-to-ON in 1T phase occurred by a bias voltage of 66 mV. In contrast, if a reverse C

DOI: 10.1021/acs.nanolett.5b04260 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. I−V characteristic of MoS2 nanosheets based memristor. (a) Schematic structure of Ag/MoS2/Ag/MoS2/Ag memristor. (b) Typical I−V characteristic of Ag/MoS2/Ag/MoS2/Ag memristor at room temperature. (c) Typical I−V characteristic of Ag/MoS2/Ag/MoS2/Ag memristor in the logarithmic scale at room temperature. (d) Typical I−V characteristic of Ag/MoS2/Ag/MoS2/Ag memristor at the 1st (red), 500th (green), and 1000th (blue) cycle at room temperature.

electrical field (namely, a reverse bias voltage) is high enough to redisplace Mo and S ions from ON state to OFF state, a reverse-state switch of ON-to-OFF would occur. This is why a reverse switch of ON-to-OFF was observed by a reverse bias voltage of −98 mV. Memristors are memory resistors, which can remember internal resistance state associated with the history of applied current and voltage.1,2 Furthermore, the term of memristor is often used for two-terminal nonvolatile memories exhibiting resistance switching behavior with asymmetric I−V characteristics in which set/reset courses are determined by OFF/ON states alternations.27 However, an ideal memristor, which was defined by Chua, should exhibit odd-symmetric I−V Lissajous figures.1 Furthermore, an ideal memristor possesses a nonlinear resistance variation, which constitutes a unique property to retain the history of applied voltage via tiny read charges. To fabricate an ideal memristor, we combine two asymmetric Ag/ MoS2/Ag switches, namely, both sides of a silver foil were coated by exfoliated MoS2 with the same approach as used for the above one side switch, forming an Ag/MoS2/Ag/MoS2/Ag memristive device (Figure 3a). When a bias voltage was applied on this device from 0 mV → 1000 mV → −1000 mV → 0 mV, a symmetric pinched hysteresis I−V loop was obtained (Figure 3b,c). At the beginning of the sweep, the current was low, indicating a high resistance state. Then by sweeping the bias voltage to a more positive value on the top Ag electrode, the current increased sharply at 114 mV but then dropped at 200 mV. This indicates the existence of negative differential resistance (NDR).28,29 The current started to rise again at 238 mV. Then, with reverse sweeping the bias voltage from

1000 mV to 0 mV, the current decreased, generating a hysteresis loop between scanning up and scanning down. Subsequent 0 mV → −1000 mV → 0 mV sweep was completely symmetric with respecting to the previous 0 mV → 1000 mV → 0 mV sweep. The hysteresis I−V curves of the Ag/ MoS2/Ag/MoS2/Ag memristive device constituted an oddsymmetric “pinched” loop, which is the “fingerprint” of an ideal memristor. Furthermore, this excellent stability of the memristor was reflected by the negligible change in 1000 sweeping cycles (Figure 3d). In summary, it was found that 1T phase of exfoliated MoS2 nanosheets exhibited a memristive behavior, whereas 2H of bulk MoS2 possessed an ohmic feature. This happened because electrical field could induce lattice distortion in 1T phase, leading to resistant change. Furthermore, an ideal memristor with odd-symmetric I−V characteristics was designed and successfully fabricated with 1T phase MoS2 nanosheets via combining two asymmetric switches antiserially.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b04260. Information on experimental section, characterization of bulk MoS2, and I−V characteristic of bulk MoS2 based device (PDF) D

DOI: 10.1021/acs.nanolett.5b04260 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

■

(23) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Nat. Nanotechnol. 2012, 7, 699−712. (24) Li, H.; Zhang, Q.; Yap, C. C. R.; Tay, B. K.; Edwin, T. H. T.; Olivier, A.; Baillargeat, D. Adv. Funct. Mater. 2012, 22, 1385−1390. (25) Jiménez Sandoval, S.; Yang, D.; Frindt, R. F.; Irwin, J. C. Phys. Rev. B: Condens. Matter Mater. Phys. 1991, 44, 3955. (26) Nayak, A. P.; Pandey, T.; Voiry, D.; Liu, J.; Moran, S. T.; Sharma, A.; Tan, C.; Chen, C.-H.; Li, L.-J.; Chhowalla, M.; Lin, J.-F.; Singh, A. K.; Akinwande, D. Nano Lett. 2015, 15, 346−353. (27) Chua, L. Appl. Phys. A: Mater. Sci. Process. 2011, 102, 765−783. (28) Yoon, S. M.; Warren, S. C.; Grzybowski, B. A. Angew. Chem., Int. Ed. 2014, 53, 4437−4441. (29) Cheng, P.; Hu, Y. H. J. Mater. Chem. C 2015, 3, 2768−2772.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

Y.H.H. supervised the project. P.C. synthesized materials, fabricated devices, and performed material characterization (XRD, SEM, and Raman) and device measurements. K.S. carried out HAADF-STEM and XPS characterization. All authors were involved in analysis and discussion of the results. Y.H.H. finalized the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the U.S. National Science Foundation (NSF-CBET-0931587). Y.H.H. also thanks Charles and Carroll McArthur for their great support. The JEOL JEM 3100R05 double Cs-corrected AEM and the Krato XPS were supported by NSF Grant No. DMR-0723032 and Grant No. DMR-0420785, respectively.

■

REFERENCES

(1) Chua, L. O. IEEE Trans. Circuit Theory 1971, 18, 507−519. (2) Strukov, D. B.; Snider, G. S.; Stewart, D. R.; Williams, R. S. Nature 2008, 453, 80−83. (3) Sangwan, V. K.; Jariwala, D.; Kim, I. S.; Chen, K.-S.; Marks, T. J.; Lauhon, L. J.; Hersam, M. C. Nat. Nanotechnol. 2015, 10, 403−406. (4) van der Zande, A. M.; Huang, P. Y.; Chenet, D. A.; Berkelbach, T. C.; You, Y.; Lee, G.-H.; Heinz, T. F.; Reichman, D. R.; Muller, D. A.; Hone, J. C. Nat. Mater. 2013, 12, 554−561. (5) Liu, H.; Si, M.; Deng, Y.; Neal, A. T.; Du, Y.; Najmaei, S.; Ajayan, P. M.; Lou, J.; Ye, P. D. ACS Nano 2014, 8, 1031−1038. (6) Bessonov, A. A.; Kirikova, M. N.; Petukhov, D. I.; Allen, M.; Ryhänen, T.; Bailey, M. J. A. Nat. Mater. 2015, 14, 199−204. (7) Yuan, J.; Lou, J. Nat. Nanotechnol. 2015, 10, 389−390. (8) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Science 2004, 306, 666−669. (9) Geim, A. K.; Novoselov, K. S. Nat. Mater. 2007, 6, 183−191. (10) Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A. Nat. Nanotechnol. 2011, 6, 147−150. (11) Novoselov, K. S.; Jiang, D.; Schedin, F.; Booth, T. J.; Khotkevich, V. V.; Morozov, S. V.; Geim, A. K. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 10451−10453. (12) Zibouche, N.; Philipsen, P.; Heine, T.; Kuc, A. Phys. Chem. Chem. Phys. 2014, 16, 11251−11255. (13) Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Phys. Rev. Lett. 2010, 105, 136805. (14) Yin, Z.; Zeng, Z.; Liu, J.; He, Q.; Chen, P.; Zhang, H. Small 2013, 9, 727−731. (15) Yoon, Y.; Ganapathi, K.; Salahuddin, S. Nano Lett. 2011, 11, 3768−3773. (16) Lee, H. S.; Min, S.-W.; Chang, Y.-G.; Park, M. K.; Nam, T.; Kim, H.; Kim, J. H.; Ryu, S.; Im, S. Nano Lett. 2012, 12, 3695−3700. (17) Beal, A. R.; Knights, J. C.; Liang, W. Y. J. Phys. C: Solid State Phys. 1972, 5, 3540−3551. (18) Wang, L.; Xu, Z.; Wang, W.; Bai, X. J. Am. Chem. Soc. 2014, 136, 6693−6997. (19) Eda, G.; Yamaguchi, H.; Voiry, D.; Fujita, T.; Chen, M.; Chhowalla, M. Nano Lett. 2011, 11, 5111−5116. (20) Shirodkar, S. N.; Waghmare, U. V. Phys. Rev. Lett. 2014, 112, 157601. (21) Acerce, M.; Voiry, D.; Chhowalla, M. Nat. Nanotechnol. 2015, 10, 313−318. (22) Miremadi, B. K.; Cowan, T.; Morrison, S. R. J. Appl. Phys. 1991, 69, 6373−6379. E

DOI: 10.1021/acs.nanolett.5b04260 Nano Lett. XXXX, XXX, XXX−XXX