Probing the mechanism for bipolar resistive switching in annealed

Subscriber access provided by READING UNIV

Article

Probing the mechanism for bipolar resistive switching in annealed graphene oxide thin films Pooja Saini, Manjri Singh, Jyoti Thakur, Ranjit A. Patil, Yuan-Ron Ma, Ram Pal Tandon, Surinder P Singh, and Ajit Kumar Mahapatro ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b09447 • Publication Date (Web): 24 Jan 2018 Downloaded from http://pubs.acs.org on January 25, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Probing the mechanism for bipolar resistive switching in annealed graphene oxide thin films Pooja Saini,1† Manjri Singh,2† Jyoti Thakur,1 Ranjit Patil,3 Yuan Ron Ma,3 Ram P. Tandon,1 Surinder P. Singh2* and Ajit K. Mahapatro1* 1

Department of Physics and Astrophysics, University of Delhi, Delhi 110007, India.

2

CSIR-National Physical Laboratory, Dr. K. S. Krishnan Marg, New Delhi 110012, India.

3

Department of Physics, National Dong-Hwa University, Hualien 97401, Taiwan.

† Equal Contributions

* Corresponding Authors: E-mail: [email protected], and [email protected] KEYWORDS: Annealed graphene oxide thin films, bipolar resistive switching, low resistance state, high resistance state, metal filament formation.

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 39

Abstract:

The bipolar resistive switching (BRS) between a metallic low resistance state (LRS) and an insulating high resistance state (HRS) is demonstrated for annealed graphene oxide (GO) thin film based device structures with aluminum (Al) as one of the contact electrodes. An optimal switching of ~104 order is recorded for Al/GO(200°C)/ITO among the device structures in metal(M2)/GO(T)/metal(M1) configurations (M1 = Al, Au, ITO, and M2 = Au, Al), fabricated using GO(T)/metal(M1), annealed at different temperatures, T = 100, 200, 300, and 400 °C. The initial Ohmic conduction for electronic transport and presence of metal contents through GO thin films in the X-ray photoelectron spectroscopy, support the physical evidence of Al filament formation between the two electrodes as imaged by high resolution transmission electron microscopy. The speculated mechanism for BRS in repeated voltage sweep cycles is attributed to the current triggered breaking of metal filaments due to a combined effect of Joule’s heating and Peltier heat generation at LRS→HRS transition, and electric field induced migration of metal atoms leading to the formation of metal filaments through the GO film at HRS→LRS transition. The higher switching ratio exhibited in the current study could be translated to engineer simple and low cost resistive memory devices.


2

Page 3 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction The progress of semiconductor industry to meet the continuous demand of modern electronic gadgets with improved computing capabilities is attributed to the innovation of efficient basic components with enhanced electronic activities.1 Currently, semiconducting technology utilizes memory elements fabricated by adopting transition metal oxide based device structures with bipolar resistive switching (BRS)2-5 that promotes transition from one resistive state to another at one polarity and regains the original resistive state at reverse polarity. The miniaturization trend in electronic devices has motivated researchers for engineering in resistive random access memory (RRAM) devices utilizing nanostructured materials with BRS properties.6-11 The carbon electronics, an emerging technology, with high density device integration capabilities promises to overcome the inherent technical issues related to miniaturization of Si technology.12,13 Recent discovery of the two-dimensional (2D) allotrope of carbon, i.e, graphene, and its derivatives graphene oxide (GO) and reduced graphene oxide (RGO), have emerged as potential low-dimensional materials with promising electronic12-18 and optical19,20 properties. The easy processing for formation of ultra-thin and robust film, makes graphene derivatives the interesting materials for minuscule electronic devices including RRAM.21,22 The presence of various electron withdrawing and electron donating functional groups (-COOH, -OH, >C=O), and defect sites in the 2D carbon network of GO and RGO, play a crucial role in defining the electronic transport properties of the material and supports development of devices controlled by transportation of electrons in single GO/RGO sheet23 and through thin layers containing stacks of GO/RGO sheets.24 The unipolar and bipolar resistive switching (RS) behavior in GO have been demonstrated for two terminal M2/GO/M1 device structures with various electrode (M1 and M2,) configurations,


3


Page 4 of 39

and using GO synthesized by adopting different recipes. The available reports describing the switching mechanism are inadequate for understanding the origin of switching phenomena and exact electronic charge transport mechanism in metal(M2)/GO/metal(M1) structures25-34, and are still debatable in the scientific and research community. In general, the experimentally observed BRS behavior in M2/GO/M1 structures is analyzed using available models for RS-behavior in metal/transition-metal-oxide/metal structures.2,6,35-39 However, contrasting RS-behaviors have been reported through identical M2/GO/M1 device structures, including (i) unipolar switching,25 and BRS with either high resistance state (HRS)26 or low resistance state (LRS)27 as initial states in Al/GO/ITO, (ii) BRS in platinum electrode structures with initial HRS28,29 in Cu/GO/Pt and LRS30 for Pt/GO/ITO, and (iii) BRS31,32 and no switching,33 in Al/GO/Al structures. Few mechanisms have also been proposed for describing the BRS events in the electrical transport measurements in the GO based two terminal structures, including filamentary effect,30 mixture of filament formation and oxygen vacancy,27,28,32,34 oxygen vacancy migration.26,29 The appropriate electronic transport mechanisms to analyze the BRS phenomena in M2/GO/M1 structures could be understood by studying current-voltage (I-V) characteristics in M2/GO(T)/M1 device structures with detailed knowledge of the structural and optical properties of GO(T)/M1 films. The present work demonstrates investigation of optimized process parameters for BRS transitions between the LRS and HRS in annealed GO thin films, and analyzes the role of metal electrodes for achieving efficient BRS behavior by performing electrical measurements through GO thin film based device structures in metal(M2)/GO(T)/metal(M1), (M1 = Al, Au, ITO, and M2 = Au, Al) configurations, with GO thin films annealed at different temperatures (T = 100, 200, 300, and 400 °C). The electronic charge transport mechanisms for LRS and HRS, role of contact electrodes and fabrication processes in achieving BRS behavior, and origin of the initial resistive


4

Page 5 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


state (LRS or HRS), are investigated. In the current-voltage (I-V) sweeps recorded through various M2/GO(T)/M1 structures, the BRS-behavior is prominent only in devices with Al as one of the electrodes. The Al/GO(200°C)/ITO structure exhibits a high switching ratio of ~104 at the threshold voltage (

) of transition. The dominating electrical transport phenomena for

describing net electronic conduction through device structures with BRS behavior follow the Ohmic and space charge limited current flow at the LRS and HRS, respectively. The BRS mechanism is attributed to the breaking and formation of the initial filaments formed due to metal (Al) atom diffusion through the thin film of GO, as supported by X-ray photoelectron spectroscopy (XPS) analysis and captured image using high resolution transmission electron microscopy (HRTEM). At

, the breaking of Al-filament is triggered by the combined effect

of Joule’s heating generated by the current flow in M2(Al)/GO(T)/M1(ITO) device structure and Peltier heat generation due to different metal junction at the GO/ITO interface. Again, the formation of these filaments due to electric field induced metal atom migration at opposite polarity of the voltage sweep.

Experimental Technique GO sheets in dry form are prepared from the graphite powders using modified Hummer’s method40 and freeze-drying using vacuum assisted low temperature processing.41,42 A dispersion solution of GO in deionized (DI) water is ultra-sonicated using a probe sonicator and spin coated over various metal substrates M1 (Al, Au, ITO). The resulting as-coated GO/M1 thin films are characterized for purity and the top metals M2 (Al, Au) are thermally deposited as the second electrode. The detailed experimental procedure is explained below.


5


Page 6 of 39

Preparation of graphene oxide sheets: All the chemicals required for the synthesis of GO, graphite powder is purchased from SP-1 graphite, Bay Carbon, USA, and sulfuric acid (H2SO4), potassium permanganate (KMnO4), sodium nitrate (NaNO3), hydrogen peroxide (H2O2), with 99.9% purity are procured from Sigma-Aldrich Co., USA. The graphene oxide sheets are synthesized from the graphite powders using modified Hummers method. Graphite powder (0.5 g) is mixed with concentrated H2SO4 (23 ml) and stirred in a conical flask mounted in an ice-bath. Additional ingredients of NaNO3 (0.5g) and KMnO4 (1.5 g) are added to the mixture followed by heating with hot plate at 35 °C and stirred for 1 h. The resulting solution is diluted by pouring 100 ml DI water (Millipore Technology, USA) and kept on stirring for another 2 h at 35 °C. The reaction is quenched by adding 27% H2O2 (1 ml) and the solution is centrifuged at 1000 rpm for 10 min. to remove the un-reacted graphite powder. Graphene oxide collected as supernatant is further centrifuged at 12000 rpm for 20 min. and washed repeatedly with DI water till the pH of the GO mixture turns to 7.0. The resulting solution is freeze-dried using vacuum assisted low temperature processing to collect highly pure GO powder.

Formation of annealed GO layers on substrates: ITO coated glass substrates with resistivity of 10-15 Ω/sq are procured from SPI supplies, USA, and pre-cleaned by ultra-sonicating in toluene, acetone, and methanol, sequentially for 3 min. each. A dispersion solution of GO with a concentration of 1 mg/ml was prepared in DI water and ultra-sonicated using a probe sonicator to achieve uniform dispersion of the dried GO. Au and Al substrates are prepared by thermally deposition technique over glass substrates. The freshly prepared slush containing GO was spin coated at 3000 rpm for 1 min. on the substrate M1


6

Page 7 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(ITO, Au, Al). This spin-coating process is repeated for 5 times to achieve the desired thickness (~ 30 nm). The spin coated GO thin films on various substrates (GO/M1) are annealed at different temperatures of T = 100, 200, 300, and 400 °C, in a furnace by maintaining the environment at ambient conditions and are identified as GO(T)/M1, e.g, GO/ITO thin films annealed at different temperatures of 100, 200, 300, and 400 °C, are identified as GO(100°C)/ITO, GO(200°C)/ITO, GO(300°C)/ITO, GO(400°C)/ITO, respectively. Similar processes are followed to form annealed and un-annealed GO thin films over Au and Al deposited glass substrates and nomenclatured accordingly. Figure1(a) displays the representative field emission scanning electron microscopy (FESEM) image of an un-annealed GO/ITO surface, captured at 5k magnifications and represented with a 1 μm scale bar. All the annealed and un-annealed GO/ITO thin films are further characterized using the available techniques for materials characterization including Fourier transform infrared spectroscopy (FTIR), UV-VisNIR absorption spectroscopy, Raman spectroscopy (Figure 1b) and transmission electron microscopy, TEM (Figure1c and 1d).

Fabrication of M2/GO/M1 device structure: The top electrode M2(Al or Au) is thermally deposited through a shadow mask on the annealed GO(T)/M1 and un-annealed GO(RT)/M1 surfaces using vacuum deposition technique, and are identified as M2/GO(T)/M1 and M2/GO(RT)/M1, respectively. The rate of deposition is kept at 1 Å/sec and the chamber pressure is maintained at 2-3 x 10-6 Torr. The optical micropscopic image of a representative M2(Al)/GO/M1(ITO) vertical device structure (VDS) is captured in Figure 1(e). The nomenclature for the Al/GO/ITO device structures prepared by thermally depositing the Al top electrode on the spin coated GO thin films over ITO surfaces without annealing is


7


Page 8 of 39

Al/GO(RT)/ITO, and for the samples annealed at temperatures of 100, 200, 300, and 400 °C are termed

as

Al/GO(100°C)/ITO,

Al/GO(200°C)/ITO,

Al/GO(300°C)/ITO,

and

Al/GO(400°C)/ITO, respectively. In the text, similar nomenclatures are followed for the Al/GO/Al, Au/GO/Al, Al/GO/Au, and Au/GO/Au structures. The M2/GO/M1 structures are placed on a probe station placed inside a Faraday cage that provides a suitable shielding platform to avoid possibility of any external electromagnetic interaction with the electronic current signals of the device structures. The grounded bottom (ITO or Al or Au) and biased top (Al or Au) metal electrodes are connected through the point probe contacts, as shown schematically in Figure 1(f). The electrical signals through the M2/GO/M1 structures are recorded by pursuing the I-V characteristics using a Keithley 4200 Semiconductor Analyzer, USA. Different Segments during the voltage sweep cycle: I-V characteristics of an Al/GO(200°C)/ITO device structure recorded by sweeping the direct current (DC) voltage in a cycle,

, containing four different

segments. In all measurements, the voltage sweep cycle starts from 0 V and ramps towards the higher positive voltage till the negative voltage

(Segment-I,

). Sweeping back from

towards

, passing through 0V and without breaking the continuity of the

applied voltage (Segment-II,

and Segment–III,

V from

), making the I-V measurements recorded in a sweep

cycle of

(Segment–IV,

). Finally, it reaches at 0

defined by the direction of bias ramp as Segment-I →

Segment-II→ Segment-III → Segment-IV.


8

Page 9 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Results Figure 1a displays the representative FESEM image of an unannealed GO(RT)/ITO surface, indicating thin layer of uniformly distributed GO sheets of various shape and sizes with wrinkled morphology. The prominent D (1362 cm-1) and G (1584 cm-1) peaks in the Raman spectroscopy of the GO(RT)/ITO surface (as represented in Figure1b) are the signature bands corresponding to the presence of sp3 and sp2 hybridized carbons in the GO sheets, respectively.43,44 The D/G ratio of 0.9 evidences the formation of highly pure GO containing ordered carbon networking with defect sites.45 The TEM image of GO sheets, drop casted on the carbon coated copper grid is shown in Figure1c. The corresponding selected area electron diffraction (SAED) plot (Figure1d) indicates formation of circular rings consisting of hexagonally arranged same intensity spots arising from the ordered region of the GO flakes.46-48 The I-V characteristics for the structures containing only the possible pairs of electrodes (Al/M1 and Al/M1(200°C), M1 = ITO, Au) formed with unannealed or annealed substrates of ITO or Au, and top Al electrode, are recorded by ramping the voltage in a continuous sweep cycle of

(Figure

S1). All these structures show Ohmic behavior with no sign of inherent hysteresis or switching and suggest their suitability in selecting as contact pairs for the formation of device structures with sandwiched GO thin films between these electrode pairs. Figure 1e shows the optical micropscopic image captured for a representative M2(Al)/GO(RT)/M1(ITO) device structure with 220 µm diameter top Al electrodes. The root mean square surface roughness of 3.6 nm was observed from atomic force microscopy (AFM) topographical measurements (Figure S2). The observed roughness is much smaller compare to the film thickness of 30 nm, indicating that the wrinkles and roughness would not impact significantly on the electrical transport measurement


9


Page 10 of 39

through VDS. Figure 1f schematizes the electrical contacts for detecting electronic signals measured by the I-V characteristics through the M2/GO(T)/M1 structures. Figure 2 represents the typical room temperature I-V characteristics of an Al/GO(200°C)/ITO device structure recorded by sweeping DC voltage in a sweep cycle, . During Segment-I of the voltage sweep, initially the current increases linearly with applied bias and at a threshold voltage (

= 3.8 ± 0.8 V), an abrupt decrease of

four orders of magnitude in current value is noticed. Since, the initial current value in Segment-I is high compared to the current after

, the earlier high conductive state is termed as LRS, and

the later low conductive state as HRS. The I-V curve follows the new HRS state by further increasing the applied bias till

and during Segment-II of the sweep cycle from

to

0V. In Segment-III, the HRS persists in the low voltages region till another threshold voltage (

=

0.7 ± 0.4 V) is attained. An abrupt increase in current at

original LRS, as in Segment-I and continue in this state till in LRS while ramping back from

shifts the structure to the . The structure further remains

till the end of the voltage sweep cycle to 0V, following

Segment-IV. The vertical and horizontal shadowed lines in Figure 2 indicate the ranges of threshold values

and

for voltages (

and currents (

, respectively, at LRS 

HRS transitions for 60 repeated switching cycles in Al/GO(200°C)/ITO device structure. Multiple devices with identical structures show LRS  HRS transitions at threshold values within the same shadowed region, indicating high yield of the switching events within this range. The I-V characteristics through other different unannealed Al/GO(RT)/ITO and annealed GO thin films at T = 100, 300, and 400 °C based Al/GO(T)/ITO device structures, are represented in Figure S3a-d. Prominent BRS behavior is noticed in Al/GO(200°C)/ITO, Al/GO(100°C)/ITO,


10

Page 11 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and Al/GO(300°C)/ITO structures, and no switching in Al/GO(RT)/ITO and Al/GO(400°C)/ITO structures. I-V sweeps for devices with various electrode configurations (Figure S4) show distinctive BRS in Al/GO(200°C)/Au, absence of switching in Au/GO(200°C)/Au, and memristive behavior49-51 in Au/GO(200°C)/ITO structures. Similar measurements reveal absence of RS behavior in Al/GO(RT)/Au and Al/GO(100°C)/Au structures (Figure S5), Au/GO(T)/ITO (Figure S6), and Au/GO(T)/Au (Figure S7) device structures. The I-V characteristics through Al/GO(T)/Al structures show three kinds of switching behaviors during the voltage sweep cycles. First, the BRS with LRS HRS transition during voltage sweep and HRS LRS during

ve

ve voltage sweep direction (Figure S8a-d,) having

switching ratio (the ratio of the current values between the LRS and HRS,

) of ~101,

~104, ~103, and ~100, for Al/GO(RT)/Al, Al/GO(100°C)/Al, Al/GO(200°C)/Al, and Al/GO(300°C)/Al device structures, respectively. Second, the observation of LRS HRS at while ramping the bias from voltage sweep cycle of

that remains in the same HRS for rest of the , and the system regains the LRS during bias

ramp of

in the next sweep cycle (Figure S9). Third, the HRS LRS switching at

during

that remains in the LRS throughout the rest of sweep cycle, (Figure S10). The above observation infers inconsistent unipolar and bipolar

switching with noticeable

ratio in Al/GO(T)/Al structures. This inconsistency in

achieving similar switching events repeatedly during various scans and from multiple devices with identical structures remain the major issues in devices with Al/GO(T)/Al configurations. Generally, the quantitative estimation of switching is calculated from the

ratio at

each applied voltages. This would provide information about the optimal operating parameters


11


of

Page 12 of 39

and range of applied bias for observing distinct LRS and HRS, amount of distinctiveness

(

) of the states, and could lead to define the critical processing recipe for fabrication of

device structures with efficient BRS behavior. Figure 3 plots the estimated values for verses the applied voltage for various scans of representative Al/GO(100°C)/ITO and Al/GO(200°C)/ITO device structures. The are observed to be 3.0 ± 0.5 V and

at LRS

HRS and

HRS LRS transitions

1.0 ± 0.5 V for Al/GO(100°C)/ITO, and 3.8 ± 0.8 V and

0.7 ± 0.4 V for Al/GO(200°C)/ITO structures, respectively. The maximum estimated switching ratio (

) at

for devices with GO(T)/ITO layers processed at various annealing

temperatures (inset of Figure 3) indicate an optimal maximum switching of 1.1

104 in the

Al/GO(200°C)/ITO structure. The statistics of threshold parameters

, power

, and

at the

LRS  HRS transitions) for the set (HRS LRS) and reset (LRS HRS) transitions in repeated scans (Figure S11) of an individual representative Al/GO(200°C)/ITO structure are analysed to be (i) independent of the resistance (RLRS and RHRS) values calculated from low bias region of the I-V curves (Figures 4a-d). The histogram plots for the threshold parameters reveal Gaussian distributions for consistent values for

centered at 3.8 V (Figure 4e) and

centered at

= (0.085 ± 0.007) A (Figure 4g) and

Gaussian distributions for

= (9.2 ± 0.7) μA- (Figure 4h),

centered at 0.3 W (Figure 4i) and

(Figure 4j), and consistently observed switching (

0.7 V (Figure 4f),

centered at 5.0 μW

) in the range of 2

103 - 1

104

(Figures 4k,4l). The switching stability is tested for reproducibility and repeatability in the set and reset values of BRS devices operation are studied by plotting the cumulative probability distribution for the threshold parameters (Figures 4m-p). Also the set and reset events doesn’t depend on the history of the structure (i.e., dependence of the values for set on the previous reset


12

Page 13 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


event and vice versa, Figure S12a-d). The narrow straight line in cumulative probability indicates stability of the device. The above statistics suggests occurrence of consistent BRS behaviour with transitions between two definite stable resistance states (LRS  HRS) in Al/GO(200°C)/ITO device structure, is triggered at constant

and

, achieved at random

, suggesting a current triggered switching mechanism responsible for observing the switching events of LRS  HRS transitions. The influence of material contents of GO layer in presence of the contact electrodes for Al/GO(200°C)/ITO structure are logistically identified by analysing the combined XPS profile of GO(RT)/Al, GO(200°C)/ITO, and compared with profile for Al/GO(200°C)/ITO. The XPS spectra are recorded at surface (d1), and at various depth positions at middle of the GO layer at a distance (d2) from the surface and at the interface (d3), as represented in the inset of Figure 5. The XPS of C1s, O1s, and Al2p core levels at various depths (d1, d2, d3) beneath the GO surface of GO(200°C)/ITO and GO(RT)/Al layers are shown in Figure 5a, 5b, and 5c, respectively. It provides information of carbon, oxygen and aluminium contents through GO layers on both the ITO and Al substrates. The XPS spectra of C1s in both the structures at various depth positions (Figure 5a) show peak at binding energy (BE) of ~284.8 eV and is assigned to the C1s core level. In Addition, at the GO surface (d1), all the structures show humps at ~286.8 eV and 288.6 eV corresponding to C-O and C=O respectively.52-53 The intensities corresponding of the C1s peak decreases (Figure S13a), and C-O vanishes, as the digging depths increased to d2 and d3 within the GO layers. The peak at BE of ~530.1 eV in the XPS spectra for all structures is assigned to O1s core level of oxygen (Figure 5b).53 At the surface (d1) of GO(200°C)/ITO and Al/GO(200°C)/ITO (Figure S14) structures, a hump is observed at BE of ~532.3 eV indicates the presence of C-O bonding. A prominent increase in the intensity of O1s peak and diminishing C-O


13


Page 14 of 39

peak are observed with further digging from the GO surface layers (d1) towards the GO/ITO and GO/Al interfaces (d3). The peaks at BEs of ~72.2 and ~74.7 eV in XPS spectra of Al2p core levels in GO(RT)/Al (Figure 5c) are assigned to the Al and Al-oxide (AlO), respectively.54,55 The increase in intensities of Al and AlO peaks are attributed to exposure to more Al content with digging towards Al substrate and formation of AlO during the bombardment of highly energetic Ar+ ions in the digging process, respectively. The XPS profiles of the C1s and O1s spectra of GO (200°C)/ITO and GO(RT)/Al, and Al2p spectra of GO(RT)/Al are fitting using XPSPEAK v 4.1 software (Figure S15 and S16). The XPS spectra of C1s, O1s and Al2p in Al/GO(200°C)/ITO structure (Figure S13) has the combined features identical with the individual GO(200°C)/ITO and GO(RT)/Al surfaces. Discussion: Mechanism for BRS The appropriate mechanism for BRS in Al/GO(T)/ITO structures is proposed by reviewing the detailed materials, spectroscopic, and electronic properties of the M2/GO(T)/M1 structures. FTIR spectroscopy helps identifying the functional groups by locating the vibrational modes in the materials. The structural changes for GO thin films annealed at different temperatures are understood by recording the FTIR spectra and is represented in Figure S17. It is noticed that annealing of the GO films allow detachment of oxygen containing functional groups54,56,57 within the 2D carbon network, and indicates evolution of RGO with structural defects (viz., StoneWales defect58,59) with increasing the temperature. XRD patterns of GO(T)/ITO (Figure S18) thin films represent the reduction in separation between GO flakes, and results shrinking of GO layer thickness with annealing temperatures.60-63 The observation of increasing atomic percentage of carbon and decreasing oxygen content with increasing annealing temperature in the EDX spectrum of the GO(T)/ITO thin films (Figure S19) agrees well with the FTIR results.


14

Page 15 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The temperature dependent I-V characteristics of Al/GO(RT)/ITO device structure, measured by sweeping the voltage from 0 to Vmax in the temperature range of 10 - 100 °C at an increment of 10 °C (Figure S20a) exhibits Ohmic behaviour (

), suggesting metal like electrical

properties of increase in resistance (as estimated from the slope of I-V curve) with increase in temperature. The values for the resistance (R) is fitted with the equation R (T) = R 0 [1-α (T-T0)], where, R0 is the resistance at temperature T0, and α is the temperature coefficient of resistance.64 The estimated value of α = 4.2 x 10-3 /oC (Figure S20b) is identical to the standard value (α = 4.3 x 10-3 /oC) for aluminium,65 implying the presence Al-metal connection between the two electrodes. The device physics for describing the BRS process is understood from the mechanism of electronic transport in LRS and HRS separately, during the voltage sweep cycle through Al/GO(200°C)/ITO structures. The I-V characteristics in log(I) versus log(V) plot through Al/GO(200°C)/ITO structure recorded for the complete voltage sweep cycle in +ve (Segments-I and II of Figure 2) and –ve (Segments-III and IV of Figure 2) bias directions are shown in Figures 6a and Figure 6b, respectively. In both the Segments I and IV of voltage sweep, initially the device structure is in LRS and exhibit linear dependence of current on voltage ( ) as long as the structure is in LRS state (before

in Segment-I in Figure 6a and in the

whole range of Segment-IV in Figure 6b). In Segments-II and III of voltage sweep, initially the device structure is at HRS and the I-V curve exhibits three different bias regions as long as the structure is in HRS state (Segment-II and before reaching at bias region, linear dependence of current on voltage

in Segment-III). First, in low

with slope 1 in log(I) – log(V) plot,

and the slope increases gradually from 1 to 2 in the moderately high applied bias second region (below

), indicating

with 1< m

Probing the mechanism for bipolar resistive switching in annealed

Recommend Documents