Resistive Switching Transparent SnO2 Thin Film Sensitive to Light and Humidity

doi:10.21203/rs.3.rs-3227982/v1

Download PDF

Article

Resistive Switching Transparent SnO2 Thin Film Sensitive to Light and Humidity

https://doi.org/10.21203/rs.3.rs-3227982/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 16 Nov, 2023

Read the published version in Scientific Reports →

You are reading this latest preprint version

Designing and manufacturing memristor devices with simple and cheap methods is very promising for its development. Here, an Ag/SnO2 /FTO(F-SnO2) structure is used through the deposition of the SnO2 layer attained by its sol via the air-brush method on an FTO substrate. This structure was investigated in terms of the memristive characteristics. The Negative differential resistance(NDR) effect was also observed in environment humidity conditions. In this structure, we have valance change memory (VCM) and electrometalization change memory (ECM) mechanisms that cause the current peak in the NDR region by forming an OH⁻ conductive filament(CF). In addition, the photoconductivity effect has been found under light illumination and this structure shows the positive photoconductance (PPC) effect by increasing the conductivity. This effect has the highest value at wavelengths close to the absorption wavelength of SnO₂ (~ 340 nm). Also, the device was examined for up to 100 cycles and significant stability was observed. This behavior is a valuable advantage because the stability of memristors is critical for their use in neuromorphic computing. The coexistence of the NDR effect and resistive switching (RS) memory behavior is useful for achieving high-level simulations of biomimetic or neuromorphic computing. This combination can lead to the creation of artificial synapses that can mimic the behavior of biological synapses.

Physical sciences/Materials science/Condensed matter physics/Electronic properties and materials

Physical sciences/Materials science/Materials for devices/Electronic devices

Memristor

Resistive switching

SnO2

Negative differential resistance

positive photoconductance

Memristors have gained significant attention as a promising technology for the development of nonvolatile memories¹ and neuromorphic computing.² Their structure is simple and have low power consumption for application in fast reading and writing.^3–5 They show potential scalability, high density, low cost, and compatibility with complementary metal oxide semiconductor (CMOS) technology.^6–10

The memristor, also known as “memory resistor” was first proposed as the fourth fundamental circuit element by Leon O. Chua in 1971.¹¹ It is a two-terminal passive device that can remember the amount of charge that has passed through it and alter its resistance based on the changes in direction and amount of charge. The first physical implementation of the memristor was demonstrated by Strukov et al. at Hewlett-Packard (HP) labs in 2008, using a thin film of titanium dioxide.¹²

The operation of a memristor is based on the migration of ions. There are several primary mechanisms in memristors such as valance change memory (VCM), electrochemical metallization memory (ECM), phase change memory (PCM), and thermally change memory (TCM).² Among them, the common mechanisms for a simple metal/insulator/metal (MIM) structure are the ECM¹³ and VCM.¹⁴In ECM memristors, the transfer of metal cations through the insulating layer is achieved by oxidizing the active metal in the top electrode (TE) upon applying a bias voltage.^14–16

A metallic filament is formed when cations are reduced at the inert electrode; this filament may be ruptured by reversing the bias voltage. So, the switching mechanism in ECM memristors is based on the rupture or formation of the conductive filament.¹⁷ Furthermore, the VCM memristors create or broke an oxygen-deficient filament upon applying a bias or reverse bias voltage. Switching in VCM memristors is triggered by the migration of oxygen anions/vacancies at the interfaces of the memristor electrodes, resulting in the modification of the Schottky barrier.^2,18 Therefore, a comprehensive understanding of the underlying mechanisms that govern the behavior of memristive devices is vital for the development of reliable and efficient devices for usage in a wide range of technological applications.

Transition metal oxides (TMOs) are one of the most appealing materials utilized in resistive switching devices.¹⁹ Among them, a variety of metal oxides, including (e.g., TiO₂,^20,21 HfO₂,²² Ta₂O₅,²³ ZnO²⁴ ZrO₂ ,²⁵) and transparent conducting oxides, such as SnO₂,²⁶ and indium tin oxide (ITO)^26,27have also been investigated as switching layers in Resistive Random-Access Memories (RRAMs). In these TMOs, a combination of features like high thermodynamic stability, large band gap, and high dielectric coefficient made them intriguing structure for use in memristor devices.²¹

Among metal oxides, SnO₂ is a semiconducting material that has attracted attention due to its unique properties in the field of oxide electronics. One of the key features of this semiconductor is its wide bandgap of approximately 3.6 eV,²⁸ which makes it transparent to visible light. This property makes SnO₂ a desirable material for optoelectronic devices such as transparent conductive electrodes in solar cells and displays.²⁶Apart from its optical characteristics, SnO₂ exhibits remarkable electrical properties, including high electron mobility and low resistance, offering it as a suitable candidate for employment in electronic domains. It is noteworthy that SnO₂ is classified as an n-type semiconductor due to its surplus of electrons, which facilitates electric conduction via the movement of electrons. This can be attributed to the existence of imperfections, such as oxygen vacancies and metal interstitials, that function as electron donors.^27,28 SnO₂ has also been explored as a potential material for resistive switching memory applications.²⁹ The reversible changes in resistance of SnO₂ upon application of an electric field make it a suitable candidate for data storage and retrieval purposes. Hence, SnO₂ is found to be an appealing contender for upcoming non-volatile memory equipment.³⁰

Moreover, recent studies have indicated that SnO₂ is capable of demonstrating threshold-switching phenomena.³¹ The phenomenon of threshold switching refers to the rapid transition of a material from high resistance state (HRS) to a low resistance state (LRS) upon the attainment of a specific threshold voltage. This property exhibits significant potential for utilization in electronic devices that operate in high-speed electronic devices.

Various techniques exist for the synthesis of thin films. The sol-gel technique is a common chemical wet deposition method utilized for the production of thin film oxide coatings, including TiO₂ and SnO₂.^{29, 32–34} In contrast to physical vapor deposition and chemical vapor deposition, the sol-gel method is commonly perceived as a more cost-effective and uncomplicated deposition technique, as it does not necessitate intricate vacuum systems or processing at elevated temperatures. Likewise, the sol-gel technique exhibits potential for industrial applications due to its ease of scalability for large-scale production.

In order to investigate the behavior of memristivity under the influence of light and humidity, it is imperative to consider the following traits. Firstly, positive photoconductance (PPC) and negative photoconductance (NPC) refer to those increase or decrease in the conductivity of a device upon exposure to light.³⁵ As an example of the NPC effect, Sheykhifar et al. investigated how memristivity of 2D-materials like Graphene (or MoS2) can be tuned by illumination of the light and its effect on I_on/I_off ratio.³⁶

They are commonly ascribed to the photo-induced excitation of charge carriers within the memristive material, thereby inducing alterations in carrier concentration and modulating the material's conductivity. ^37,38 The resistive switching effect in memristors can be combined with PPC, leading to the emergence of innovative capabilities, such as optically programmable resistive switching. Utilization of this methodology can facilitate optical manipulation of the device's resistance, rendering it an attractive option for optical computing and neuromorphic systems.^39,40 Secondly, the negative differential resistance (NDR) effect is characterized by an atypical deviation in the “On-state” of the I-V curve, exhibiting a different behavior to that of conventional resistive components.⁴¹ The existence of NDR in these devices can be attributed to the quantum mechanical effect of electron tunneling.³⁹ Upon application of a voltage to the device, electrons can tunnel through a thin barrier in the semiconductor material, thereby facilitating the flow of electric current. Within specific voltage ranges, tunneling exhibits enhanced conductivity resulting in a reduction of resistance and a concomitant augmentation of current flow. However, at higher voltages, the tunneling effect becomes less efficient, leading to an increase in resistance and a decrease in current flow.⁴⁰ Recently the existence of both memristance and NDR effect in Gr/MoS₂ heterostructures was reported as an interface-induced effect.⁴² The coexistence of PPC and NDR effect with resistive switching (RS) memory behavior in certain materials is advantageous for attaining advanced simulations of biomimetic or neuromorphic computing.⁴³ This field of study endeavors to create computer systems capable of executing tasks similar to what happens in the human brain. The stability of synaptic weights is crucial for precise and dependable computation in a neuromorphic computing system.⁴⁴ Hence, attention is to employ materials that possess characteristics comparable to those of neural synapses, which encompass the capacity to store and process data. ⁴³

Exploring the use of such materials and devices for developing advanced neuromorphic computing systems is under debate⁴⁵ toward implementation in pattern recognition, image processing, and machine learning. The emergence of NDR memristors offers the capacity to transform the realm of electronics through the facilitation of novel applications and functionalities that are unattainable with conventional electronic devices. The present study points out that the implementation of Ag/SnO₂/FTO structure results in the manifestation of memristive behavior in SnO₂, thereby enabling its application as a switching resistor. Furthermore, this particular configuration exhibits photosensitivity and indicates a PPC phenomenon, whereby its conductivity is enhanced, particularly under ultraviolet (UV) irradiation. Furthermore, the NDR effect is demonstrated in the memristor under humidity condition. The concurrent manifestation of the NDR phenomenon and resistive switching (RS) memory characteristics are highly advantageous for applications in synaptic and neuromorphic computing, as evidenced in ref. ⁴⁴

Thin SnO₂ layers were fabricated using an airbrush spraying technique. To produce the precursor solution, ZnCl₂. 2H₂O (Merck) was dissolved in isopropanol (Extra pure) C₃H₈O and stirred for 1 hour at ambient temperature. A stable uniform colloidal solution was made. While stirring, a certain volume of deionized water was introduced drop by drop to this solution for the hydrolysis process. The molar proportions of Zinc chloride dihydrate, isopropanol, and water were 1:113:38. After 24 hours of stabilization at room temperature and in ambient air, the hydrolysis reaction was complete. This procedure yielded a transparent solution that is used as a precursor to the deposition of SnO₂ thin film.

The FTO glasses were purchased from the market and cut to a size of 1×1 centimeter. They underwent a preliminary cleaning process that consisted of acetone and ethanol in an ultrasonic bath. The FTO glass was placed on a hot plate at 240°C for several minutes, and the prepared solution was sprayed onto it from a distance of 6 cm. SnO₂/FTO glass was fabricated, and Ag paste was applied to specific areas of the sample in order to form contacts. The I–V measurements were utilized via a source meter (Keithley 2450). The schematic diagram of an Ag/SnO₂/ FTO structure is illustrated in Fig. 1 (a). The impact of light on the sample was examined by subjecting the structure to light and measurement of electrical resistance and conductivity were recorded. The study revealed that the sample exhibits light sensitivity and the PPC effect has been detected. In order to examine the impact of humidity on the NDR phenomenon, the sample was placed in a container with an open side and the moisture applied by a humidifier. The current-voltage (I-V) characteristics were obtained and subsequently evaluated for three discrete levels of humidity, specifically 95%, 70%, and 40%. The NDR phenomenon is distinguished by a discernible peak in the current-voltage (I-V) curve, as a fingerprint of this effect.

In order to characterize some features like the crystal structure, morphology of the surface, and interaction of light with the sample, the following characterizations were performed. X-Ray diffraction (XRD) was used to investigate the crystal structure of the sample (RIGAKU-Geigerflex diffractometer). The intensity of XRD was attained by changing 2θ from 10 to 80 degrees. Also, the morphology of the surface obtained by an atomic force microscope (Nanosurf Mobile-S portable) and the roughness parameters were investigated. The observation was attained in a window of 9×9 µm² from the sample. In addition, according to the UV-visible measurements (Perkin Elmer Lambda 25 UV/vis spectrometer), SnO₂’s absorption spectrum has been acquired and its band gap has been measured.

X-ray diffraction pattern of tin dioxide thin films prepared by sol-gel technique is shown in Fig. 1 (b). The diffraction pattern of SnO₂ exhibits peaks at (110), (200), (211), (310), and (301) with corresponding 2θ diffraction angles of 26.52°, 37.74°, 51.41°, 61.55°, and 65.37°, respectively. These peaks belong to cassiterite crystal phase with tetragonal rutile structure (JPDS No. 01-077-0452).

According to the UV-visible absorption spectrum of SnO₂ at a wavelength of about 340 nm, which originates from the band gap transitions and directs the existence of highly crystalline SnO₂, the maximum absorption occurred (Fig. 1 (c)). Based on Davis-Mott and Tauc,^46,47 the absorption coefficient α is measured as a function of photon energy hν, and the data is plotted as (αhν)² in terms of hν. The plot of (αhν)² in terms of hν produces a straight line with a slope proportional to the absorption coefficient and an intercept proportional to the band gap energy E_g. The equation for this line is given by (αhν)² = B(hν - E_g)^γ, where h is Planck's constant, ν is the photon's frequency, and B is a constant. Depending on the nature of the electron transition, the value of γ is either 1/2 or 2, for direct and indirect transition band gaps, respectively. To determine the band gap energy E_g, the straight lines of the spectra are extrapolated to the point where (αhν)² = 0. The intersection of these lines with the Energy-axis gives the value of hν at E_g. According to Fig. 1 (c), the band gap energy E_g=3.6 eV. The determination of the band gap energy is a crucial step in constructing a free band diagram of the structure to acquire a more comprehensive comprehension of the movements of charge carriers. Further elaboration will be provided subsequently.

AFM measurements have been performed to examine the surface of the film, giving detailed information on the nanoscale texture and film roughness. Figure 1(d), shows the non-uniform distribution of surface height of the thin film which lighter areas show the peaks and the darker areas represent the valleys. For a better understanding of this nonuniformity, a three-dimension image of the surface height distribution is illustrated in Fig. 1 (e). which shows a highly rough surface and spatial variations are clearly visible. We performed AFM line scan measurement diagonally along the resulting image (Fig. 1 (f)). The R_q is the amount of the average deviation from the mean of the surface height and is commonly used to quantify the roughness of a surface. This R_q is attained between ~ 7–10 nm in a 9×9 µm2 window of the sample.

As evidenced by Fig. 1 (e), the surface exhibits a high degree of roughness, resulting in an increased effective surface area. This, in turn, facilitates greater charge accumulation, particularly in areas of sharp curvature. The presence of roughness is a significant factor in both the formation and disruption of conductive filaments. It is noteworthy that these sharp regions serve as the primary catalyst for the creation of conductive pathways.

The current-voltage (I-V) characteristics were examined in both presence and absence of light, to study the impact of light on the sample. These tests revealed the PPC effect as an indication of increase in conductivity. Additionally, these measurements were performed while varying the available humidity exposed to the sample in order to investigate the NDR impact. The I-V curve for this effect is asymmetric, and there is an additional peak that results from an additional redox reaction emanating from available humidity. We will discuss our results in continue.

4.1. Mechanisms of photoconductance (The PPC and NPC effect)

The Ag/SnO₂/FTO structure is a light-sensitive photo-memristor. When sample is exposed to white light, the excitons produced by the light can change the conductivity in two ways known as negative photoconductance (NPC) and positive photoconductance (PPC). The NPC, or non-photoconductive state, is characterized by the inhibition of charge carrier participation in current conduction due to light emission, leading to an increase in resistance. Charge confinement occurs within the interfacial region of two adjacent layers, leading to a reduction in conductivity. PPC, or positive photoconductance effect, is a phenomenon in which the absorption of light induces the flow of charge carriers, resulting in a decrease in resistance and an increase in current conduction.

Excitons are separated and generate charge carriers at the Ag and SnO₂ interface in our structure. The existence of a trap located at the heterogeneous interface of two distinct materials leads to the separation of excitons and the birth of hot carriers, thereby augmenting conductivity. The memristivity of the sample is illustrated in Fig. 2 under two conditions: with and without the presence of light. Figure 2 (a) depicts the current-voltage (I-V) measurement in the absence of light, whereas the I-V measurement in the presence of white light and UV emission is depicted in Fig. 2 (b and c), respectively. In a dark environment, at a certain voltage (1 V) the current that passing in “on” state is five times greater than that of “off” state i.e. (\(\frac{{\text{I}}_{\text{o}\text{n}}}{{\text{I}}_{\text{o}\text{f}\text{f}}}=5.4)\) (from 0.002 to 0.00037 mA). The ratio \(\frac{{\text{I}}_{\text{o}\text{n}}}{{\text{I}}_{\text{o}\text{f}\text{f}}}=1.12\) decreases under light illumination and by applying UV light this ratio almost vanished. It is clear that white light and UV emission increased the conductivity, confirming the PPC effect.

In the absence of light, there is a relatively long Schottky barrier between Ag electrode and SnO₂. When a small positive bias voltage is applied, most of the free carriers are trapped at the Ag/SnO₂ interface and only a small number of them can cross the barrier and the loop in I-V curve hysteresis is considerably wide (Fig. 2(a)(.

By exposing the sample to the light, the movement of electrons facilitates due to the smaller Schottky barrier that electrons face between the Ag electrode and SnO₂. So, the loop in I-V curve become narrower (Fig. 2 (b)). In this case, by applying light, excitons are generated in the presence of an external electric field and subsequently separate at the interface, as illustrated in (Fig. 2(e)(. The generation of charge carriers through the absorption of light has the potential to induce a notable impact on the energy band's curvature and the Schottky barrier's height (δ), as depicted in Fig. 2 (f). In the present state, the carriers exhibit easy traversing through barrier, resulting in an increase in conductivity and a narrower hysteresis loop in the I-V curve. The UV-visible spectrum analysis (Fig. 1 (c)) has revealed that SnO₂ exhibits an absorption peak at a wavelength of 340 nm, which falls within the UV light range. It derives from the band gap transitions and indicates to the presence of crystalline SnO₂. So, it can be recognized that by applying the UV light, the PPC effect occurs and it raises up the conductivity, which is in conformity with our result represented in (Fig. 2(c)). The intensification of the PPC effect and the near disappearance of the loop in the I-V curve can be attributed to the close proximity of the wavelength of UV emission and the band gap of SnO₂.

4.2. The coexistence of the NDR effect and RS behavior

In this work, both NDR and RS properties observed in the Ag/SnO₂/FTO structure by applying a low bias voltage. These effects were elicited by exposing the sample to moisture produced by a humidifier, concurrently with the application of a low bias voltage. The observation of the NDR effect is dependent on the ratio of water molecules found in the surrounding atmosphere. This phenomenon is inherent in normal conditions and can be enhanced through the introduction of additional humidity. In the typical ambient environment and at normal room temperature, the process of adsorption occurs where gaseous water molecules adhere to the surface of the structure. Following this, the water molecules then interact with the oxygen that is present within the lattice structure, as well as any oxygen vacancies that may be present and then produces hydroxide ions to migrate to SnO₂ bulk. Valov and co-workers have pointed out that water molecules are easily absorbed in a thin layer with a nano porous structure. ^48,49 In electrochemical concepts, the peaks observed in the plot serve as indicators of the redox reaction occurring within the sample. So, by applying extra humidity one can manipulate oxidation and reduction reactions and intensify the NDR effect.⁴³

A significant NDR peak was observed when the sample was subjected to 95% humidity and the corresponding current-voltage (I-V) curve was measured simultaneously. After then, the source of the humidifier was turned off, and over time, the amount of humidity that was accessible within the container reduced. This measurement was repeated at 70% humidity after five minutes and 40% humidity after twenty.

The memristive characteristic, particularly observed in metal oxides, is influenced by the oxygen concentration within the surrounding environment. As depicted in Fig. 3, the passage of time leads to a reduction in humidity as a result of water molecule absorption by the structure, consequently leading to a decrease in atmospheric oxygen levels. The reduction of humidity from 95–40% resulted in a decrease in the NDR Peak. However, as depicted in Fig. 3(b and c), it is evident that the conductivity exhibited an increase over time. This can be attributed to the penetration of water molecules into the structure. The sharpness of the NDR peak at 95% relative humidity is roughly twofold greater than when the relative humidity is 70%. The relativity of ∆I/∆V reduced in 0.35% by decreasing of the humidity from 95 to 70% and this value reached 0.28% with 40% of humidity.

As previously discussed, there is a notable surface roughness that contributes to an increased effective surface area. This enhanced surface area facilitates the acceleration of the redox reaction and enables the preservation of conductive pathways. Although the tin dioxide developed in this study has a tetragonal structure, its polycrystalline structure and abundant grain boundaries with oxygen vacancies may provide ion localization.

The presence of oxygen facilitates the migration of metal ions, which is necessary for the formation of these filaments.⁴³ The presence of oxygen vacancies in memristivity holds significant importance and plays a pivotal role, alongside various other factors including temperature, voltage, and material composition, in determining the behavior of these devices.

In order to elucidate the phenomenon of NDR, the current-voltage (I-V) characteristic curve was partitioned into six distinct operational stages ((i) to (vi) stages at Fig. 4 relative to Fig. 4 (c-h)). Water molecules provide \(\text{O}{\text{H}}^{-}\)ions, which have the ability to influence redox reactions during the I-V measurements. The absorbed water molecules in the reaction with oxygen inside the semiconductor as well as oxygen vacancies (\(\text{Ѵ}\)_Ox) result in the formation of hydroxide ions (\(\text{O}{\text{H}}^{-}\)) as described by Equation \((1\)). Also, high effective surface area of SnO₂ and abundant grain boundaries with provision of oxygen vacancies produce \(\text{O}{\text{H}}^{-}\). In step (i), according to the less \(\text{O}{\text{H}}^{-}\) migration at low bias voltage, memristor is still in HRS mode.

\({{\text{H}}_{2}\text{O}\left(\text{g}\right)+\text{O}}_{0}^{\text{x} }+{\text{Ѵ}}_{\text{O}\text{x}} \rightleftarrows 2\text{O}{\text{H}}^{-}\) \(\left(1\right)\)

This equation is known as hydric reaction.⁴⁶ Where \({\text{O}}_{0}^{\text{x} }\) and \({\text{Ѵ}}_{\text{O}\text{x}}\) denote the oxygen in the lattice and oxygen vacancies, respectively. By changing the voltage from 0 to -1.5 V, a large volume of \(Ѵ\)_Ox migrates toward the Ag electrode, which causes the formation of a conductive filament of \(Ѵ\)_Ox between the Ag and FTO electrodes. Afterward, the memristor switches into LRS mode (SET process).

In step (ii), as the voltage changes from − 1.5 V to 0, the memristor remains in the LRS until the slope of the I-V curve experiences a significant decrease as illustrated in Fig. 4(d) and associated with a voltage of -0.6 V in Fig. 4(a). Then according to the slope of the I-V curve, the memristor is placed in a HRS. The observed rise in resistance can be attributed to a concurrent reduction in the electric field. At this stage, regardless of the restriction imposed by the negative bias voltage on the migration of \(\text{O}{\text{H}}^{-}\) ions, these ions actively contribute to the electrical conductivity and enhance the current.

In step (iii), when the voltage increases from 0 to 1.5 V, the migration of Ag + ions towards the FTO electrode is facilitated by positive bias voltages and reduced by an electron to form the Ag as stated in Eq. (2). By increasing the electric field, the Ag + ions accumulate at the interface of FTO, eventually forming a conductive filament. Additionally, the presence of a positive bias at the interface of the FTO electrode encourages the attraction of \(\text{O}{\text{H}}^{-}\) ions. Hence, this process facilitates the formation of an \(\text{O}{\text{H}}^{-}\)conductive pathway, ultimately leading to an increase in the current. At the voltage of 1.5, the simultaneous presence of two Ag and \(\text{O}{\text{H}}^{-}\) pathways are responsible for the conductivity of the sample. However, the filament of the oxygen vacancies is also weakly present.⁴⁸

\({\text{A}\text{g}\rightleftarrows \text{A}\text{g}}^{+ }+{\text{e}}^{-}\) \(\left(2\right)\)

Plus, during this particular stage, the oxygen vacancies progressively migrate back to the FTO electrode, thereby signifying a smooth relaxation of the \(Ѵ\)_Ox filament.

In step (iv), the reduction of voltage from + 1.5 V to 0 results in a decrease in positive bias, which weakens the Ag filament. Over time, the \(\text{O}{\text{H}}^{-}\) ions find an opportune moment to enhance their strength and according to Equation \((1\)) starts to stablish a conductive path of oxygen vacancies. And by some means, these filaments are in the trade-off, and regarding the competition between them the current may remain intact. Also, it may face a decrease and appear as a deep in the I-V curve. (Here the current remained unchanged until the threshold voltage “V_th” was reached Vth = 1.3 V (see Fig. 4 (a)- step iv and Fig. 4 (d)).

In stage (v), combined with the argument presented in the step prior, at the threshold voltage (V_th), alongside the presence of dwindled silver and the \(\text{O}{\text{H}}^{-}\) filament, the oxygen vacancy filaments formed with the highest quantity resulting in a peak in current despite the reduction of voltage. The presence of these three filaments is responsible for the appearance of this peak, which induces the system to switch into the LRS (see Fig. 4 (a)- step v and Fig. 4 (g)).

The Ag⁺ ions then migrate to the FTO. In this stage, oxidation and electron loss take place (see Equation \(\left(3\right)\)). These electrons also contribute to the generation of peak current. In fact, at this juncture, two constituents of the conductive filament, \(\text{O}{\text{H}}^{-}\) and Ag⁺, as well as electrons expelled by oxidation reactions, contribute to the generation of current.⁴⁹

\({2\text{H}}_{2}\text{O}\leftrightarrow {\text{O}}_{2 }+4{\text{H}}^{+}+4{e}^{-}\) \(\left(3\right)\)

In stage (vi), through the progressive reduction of voltage and decrease in the applied field, the driving force of \(\text{O}{\text{H}}^{-}\) and Ag⁺ is weakened and both \(\text{O}{\text{H}}^{-}\) and Ag⁺ filaments are reduced until both filaments are ruptured and the value of current becomes zero. While, in this stage, the filament of oxygen vacancy has the highest concentration until the available water splitting decreases due to a drop in voltage and therefore the reduction in available \(\text{O}{\text{H}}^{-}\)which is responsible for the intensification of oxygen vacancies. And the system is managed by HRS.

It worth to mention that an unstable AgOH is formed (Equation\(\left(4\right)\)) due to the reaction of \(\text{O}{\text{H}}^{-}\) and Ag⁺ and immediately decomposes to Ag₂O at room temperature. At this time, the ion migration in the body and surface of SnO₂ is limited by Ag₂O. As a limitation of this structure, it should be mentioned that there is an undesirable reaction that leads to producing silver oxide. The current limitation by Ag₂O is also weak. Rupturing and tearing the conductive filaments cause the device to return to HRS mode (Equation \(\left(5\right)\)).

\({\text{A}\text{g}}^{+}+\text{O}{\text{H}}^{-}\leftrightarrow \text{A}\text{g}\text{O}\text{H}\) \(\left(4\right)\)

\(2\text{A}\text{g}\text{O}{\text{H}}^{-}\leftrightarrow \text{A}{\text{g}}_{2}\text{O}+{\text{H}}_{2}\text{O}\) \(\left(5\right)\)

4.2.1 Fitting the current–voltage curves with the Space Charge Limited Current (SCLC) model

To describe charge transport in thin-film devices, such as memristors, SCLC model is commonly used. This model can be described by Equation\(\left(6\right)\). ^47,50

\(\text{J} \propto \frac{{\text{V}}^{\text{m}+1}}{{\text{d}}^{2\text{m}+1}}\) \(\left(6\right)\)

Where J is the current density, “V” is the bias voltage in units of Volt, “d” is the thickness of the tin dioxide in units of nm and “m” is the fitting index which is a dimensionless constant parameter.

The SCLC model contains three regimes that describe the (I-V) characteristics of the device. (i) Low voltage Ohmic region which is described by Ohm's law, J = GV, (ii) Child's square law⁵¹ region which J ∝ V² and (iii) Current steep increase region which J ∝ V^α, where α represents the order of voltage between 3 and 4. At a relatively low bias voltage region, (m = 0), the current density is dominated by Ohm's law and it is proportional to the applied voltage. In a fairly high bias voltage region, (m = 1), the current density is dominated by Child's law. So, the current density increases with the square of the applied voltage and can be described by J ∝ V². To explain this region, double-log I–V curves in the positive voltage region are replotted in Fig. 4 (b). By obtaining the slopes of different regions in HRS mode, the fitting value of 2.13 and 2.26 imply that the Child's law (i.e., the oxygen vacancy migration) may dominate the charge behaviors in this region. In LRS mode at the low bias, the fitting value of 1.5 indicates that the Ohmic conduction⁴⁷ (i.e., the Ag metallic filaments) is dominated. In high bias voltage region (both in HRS and LRS part) the fitting value of 3.7,4.09 and 6.01 indicate the Schottky tunneling mechanism. The Schottky tunneling model ^{50, 52–54} can be described by Equation \(\left(7\right)\) and\(\left(8\right)\)

\(J \propto \text{exp}(\frac{\beta }{{k}_{B}T}{E}^{\frac{1}{2 }}-\frac{\phi }{{k}_{B}T})\) \(\left(7\right)\)

\(\beta =\sqrt{\frac{{q}^{3}}{k\pi {\epsilon }_{0}{\epsilon }_{r }}}\) \(\left(8\right)\)

That \(\phi\), \({k}_{B}\), T, E, q, \({\epsilon }_{0}\) and \({\epsilon }_{r }\)denotes the Schottky barrier, Boltzmann constant, temperature, electric field, electric charge, vacuum permittivity and relative permittivity, respectively.

4.3. Stability

As confirmed by Xiaofang Hu and his colleagues,⁴⁶ the coexistence of NDR and RS in a memristor improves storage density and reduces power consumption. Also, X. Ji and his colleagues,⁵⁵ reported that the coexistence of these two effects can change memory time without additional circuitry and cost.

One of the important characteristics of memristors is stability, which refers to their ability to maintain the resistance state over time. Also, stability has a key role in the development of reliable and efficient neuromorphic computing systems. Some studies have shown that certain types of memristors, such as those made from oxide materials, can exhibit stable resistance states over long periods of time.⁵⁶

The observed high stability of the memristor in this particular structure renders it a dependable choice for applications in memory and neuromorphic computing. Figure 5 (a) exhibits the remarkable stability of the SnO₂ memristor throughout 100 cycles. Additionally, Fig. 5 (b) showcases the endurance of the HRS and LRS over 100 voltage cycles, demonstrating exceptional repeatability.

A memristor with a structure of Ag/SnO₂/FTO exhibits coexistence of NDR and RS behavior at room temperature. We demonstrated this memristive property for the first time in this structure. This structure can be used as a photo-memristor and when it exposed to light, especially UV light, its conductivity increased, which confirms the existence of the PPC effect. Also, the NDR effect was observed due to the rupture or resurgence of metallic, oxygen vacancy, or hydroxide (\(\text{O}{\text{H}}^{-}\)) conductive filaments. These results can be beneficial for neuromorphic applications and its stability over 100 cycles enables it to maintain their synaptic weights over time which is important in strength of the connection between two neurons.

Data availability

The data supporting the findings of this study are available from the corresponding author upon request.

Pereira, M. E., Martins, R., Fortunato, E., Barquinha, P. & Kiazadeh, A. Recent progress in optoelectronic memristors for neuromorphic and in-memory computation. Neuromorphic Computing and Engineering 3, 022002 (2023).
Waser, R., Dittmann, R., Staikov, G. & Szot, K. Redox‐Based Resistive Switching Memories – Nanoionic Mechanisms, Prospects, and Challenges. Advanced Materials 21, 2632–2663 (2009).
Xiao, Y. et al. A review of memristor: material and structure design, device performance, applications and prospects. Sci Technol Adv Mater 24, (2023).
Zhou, G. et al. Negative Photoconductance Effect: An Extension Function of the TiO _x ‐Based Memristor. Advanced Science 8, 2003765 (2021).
Jamilpanah, L., Khademi, I., Shoa e Gharehbagh, J., Aziz Mohseni, S. & Mohseni, S. M. Promising memristive behavior in MoS2–MoO2–MoO3 scalable composite thin films. J Alloys Compd 835, 155291 (2020).
An, Y. J. et al. Demonstration of high on/off ratio and linearity of SnO2-quantum-dot-based synaptic device. Mater Lett 341, 134193 (2023).
Sebastian, A., Le Gallo, M., Khaddam-Aljameh, R. & Eleftheriou, E. Memory devices and applications for in-memory computing. Nat Nanotechnol 15, 529–544 (2020).
Chen, H. et al. Time‐Tailoring van der Waals Heterostructures for Human Memory System Programming. Advanced Science 6, 1901072 (2019).
Mehonic, A. et al. Memristors—From In‐Memory Computing, Deep Learning Acceleration, and Spiking Neural Networks to the Future of Neuromorphic and Bio‐Inspired Computing. Advanced Intelligent Systems 2, 2000085 (2020).
Kim, M.-K., Kim, I.-J. & Lee, J.-S. CMOS-compatible compute-in-memory accelerators based on integrated ferroelectric synaptic arrays for convolution neural networks. Sci Adv 8, (2022).
Chua, L. Memristor-The missing circuit element. IEEE Transactions on Circuit Theory 18, 507–519 (1971).
Strukov, D. B., Snider, G. S., Stewart, D. R. & Williams, R. S. The missing memristor found. Nature 453, 80–83 (2008).
Valov, I., Waser, R., Jameson, J. R. & Kozicki, M. N. Electrochemical metallization memories—fundamentals, applications, prospects. Nanotechnology 22, 254003 (2011).
Valov, I. & Tsuruoka, T. Effects of moisture and redox reactions in VCM and ECM resistive switching memories. J Phys D Appl Phys 51, 413001 (2018).
Banerjee, W. Challenges and Applications of Emerging Nonvolatile Memory Devices. Electronics (Basel) 9, 1029 (2020).
Banerjee, W., Liu, Q. & Hwang, H. Engineering of defects in resistive random access memory devices. J Appl Phys 127, 051101 (2020).
Waser, R., Dittmann, R., Menzel, S. & Noll, T. Introduction to new memory paradigms: memristive phenomena and neuromorphic applications. Faraday Discuss 213, 11–27 (2019).
Jamilpanah, L., Khademi, I., Shoa e Gharehbagh, J., Aziz Mohseni, S. & Mohseni, S. M. Promising memristive behavior in MoS2–MoO2–MoO3 scalable composite thin films. J Alloys Compd 835, 155291 (2020).
Metal oxide memristors for neuromorphic electronics. (Nanyang Technological University, 2021). doi:10.32657/10356/155068.
Joglekar, Y. N. & Wolf, S. J. The elusive memristor: properties of basic electrical circuits. Eur J Phys 30, 661–675 (2009).
Corinto, F. & Ascoli, A. A Boundary Condition-Based Approach to the Modeling of Memristor Nanostructures. IEEE Transactions on Circuits and Systems I: Regular Papers 59, 2713–2726 (2012).
Amer, S., Sayyaparaju, S., Rose, G. S., Beckmann, K. & Cady, N. C. A practical hafnium-oxide memristor model suitable for circuit design and simulation. in 2017 IEEE International Symposium on Circuits and Systems (ISCAS) 1–4 (IEEE, 2017). doi:10.1109/ISCAS.2017.8050790.
Strachan, J. P. et al. State Dynamics and Modeling of Tantalum Oxide Memristors. IEEE Trans Electron Devices 60, 2194–2202 (2013).
Pan, F., Gao, S., Chen, C., Song, C. & Zeng, F. Recent progress in resistive random access memories: Materials, switching mechanisms, and performance. Materials Science and Engineering: R: Reports 83, 1–59 (2014).
Nauman, M. M., Zulfikre Esa, M., Zaini, J. H., Iqbal, A. & Bakar, S. A. Zirconium Oxide based memristors fabrication via Electrohydrodynamic Printing. in 2020 IEEE 11th International Conference on Mechanical and Intelligent Manufacturing Technologies (ICMIMT) 167–171 (IEEE, 2020). doi:10.1109/ICMIMT49010.2020.9041208.
Kiruthiga, G. et al. SnO2: Investigation of optical, structural, and electrical properties of transparent conductive oxide thin films prepared by nebulized spray pyrolysis for photovoltaic applications. Inorg Chem Commun 145, 109968 (2022).
Trivedi, S. J. & Joshi, U. S. Resistive Switching Properties of Highly Transparent SnO2:Fe. Journal of Nano- and Electronic Physics 9, 01025-1-01025–5 (2017).
Pan, Y. et al. Mimicking synaptic plasticity and learning behaviours in solution processed SnO2 memristor. J Alloys Compd 757, 496–503 (2018).
Komal, K., Gupta, G., Singh, M. & Singh, B. Improved resistive switching of RGO and SnO2 based resistive memory device for non-volatile memory application. J Alloys Compd 923, 166196 (2022).
Panigrahy, S. & Dhar, J. C. Non-volatile memory application of glancing angle deposition synthesized Er ₂ O ₃ capped SnO ₂ nanostructures. Semicond Sci Technol 35, 055035 (2020).
An, Y. J. et al. Demonstration of high on/off ratio and linearity of SnO2-quantum-dot-based synaptic device. Mater Lett 341, 134193 (2023).
Boratto, M. H., Ramos, R. A., Congiu, M., Graeff, C. F. O. & Scalvi, L. V. A. Memristive behavior of the SnO2/TiO2 interface deposited by sol–gel. Appl Surf Sci 410, 278–281 (2017).
Ryu, H. & Kim, S. Synaptic Characteristics from Homogeneous Resistive Switching in Pt/Al2O3/TiN Stack. Nanomaterials (Basel) 10, (2020).
Xu, X., Zhuang, J. & Wang, X. SnO ₂ Quantum Dots and Quantum Wires: Controllable Synthesis, Self-Assembled 2D Architectures, and Gas-Sensing Properties. J Am Chem Soc 130, 12527–12535 (2008).
Zhang, Z. et al. In-sensor reservoir computing system for latent fingerprint recognition with deep ultraviolet photo-synapses and memristor array. Nat Commun 13, 6590 (2022).
Sheykhifar, Z. & Mohseni, S. M. Highly light-tunable memristors in solution-processed 2D materials/metal composites. Sci Rep 12, 18771 (2022).
Chen, K.-Y. et al. The Effect of Oxygen Vacancy Concentration on Indium Gallium Oxide Solar Blind Photodetector. IEEE Trans Electron Devices 65, 1817–1822 (2018).
Cui, S., Mei, Z., Zhang, Y., Liang, H. & Du, X. Room-Temperature Fabricated Amorphous Ga ₂ O ₃ High-Response-Speed Solar-Blind Photodetector on Rigid and Flexible Substrates. Adv Opt Mater 5, 1700454 (2017).
Messaris, I. et al. A Simplified Model for a NbO ₂ Mott Memristor Physical Realization. in 2020 IEEE International Symposium on Circuits and Systems (ISCAS) 1–5 (IEEE, 2020). doi:10.1109/ISCAS45731.2020.9181036.
Zhang, J. et al. A generalized model of TiO _x-based memristive devices and its application for image processing. Chinese Physics B 26, 090502 (2017).
Liu, X. et al. Understanding composite negative differential resistance in niobium oxide memristors. J Phys D Appl Phys 55, 105106 (2022).
Bastani, P., Mohseni, S. M., Jamilpanah, L., Azizi, B. & Gharehbagh, J. S. e. Interface-induced negative differential resistance and memristive behavior in Gr/MoSe2 heterostructure. Journal of Materials Science: Materials in Electronics 33, 6403–6410 (2022).
Zhou, G. et al. Coexistence of Negative Differential Resistance and Resistive Switching Memory at Room Temperature in TiO _x Modulated by Moisture. Adv Electron Mater 4, 1700567 (2018).
Lu, Q. et al. Low-Dimensional-Materials-Based Flexible Artificial Synapse: Materials, Devices, and Systems. Nanomaterials 13, 373 (2023).
Sun, B., Zhou, G., Guo, T., Zhou, Y. N. & Wu, Y. A. Biomemristors as the next generation bioelectronics. Nano Energy 75, 104938 (2020).
Hu, X. et al. Refining the Negative Differential Resistance Effect in a TiO _x-Based Memristor. J Phys Chem Lett 12, 5377–5383 (2021).
Zhou, G. et al. Hydrogen-peroxide-modified egg albumen for transparent and flexible resistive switching memory. Nanotechnology 28, 425202 (2017).
Valov, I. et al. Nanobatteries in redox-based resistive switches require extension of memristor theory. Nat Commun 4, 1771 (2013).
Tappertzhofen, S. et al. Generic Relevance of Counter Charges for Cation-Based Nanoscale Resistive Switching Memories. ACS Nano 7, 6396–6402 (2013).
Zhang, H. et al. Research on Pt/NiO _x/WO _{3– x} :Ti/W Multijunction Memristors with Synaptic Learning and Memory Functions. J Phys Chem Lett 12, 3600–3606 (2021).
Chakrabarti, S. et al. Evolution of resistive switching mechanism through H 2 O 2 sensing by using TaO x -based material in W/Al 2 O 3 /TaO x /TiN structure. Appl Surf Sci 433, 51–59 (2018).
Pradhan, N. R. et al. An Optoelectronic Switch Based on Intrinsic Dual Schottky Diodes in Ambipolar MoSe ₂ Field‐Effect Transistors. Adv Electron Mater 1, 1500215 (2015).
Sahu, M. C. et al. Effect of Charge Injection on the Conducting Filament of Valence Change Anatase TiO ₂ Resistive Random Access Memory Device. J Phys Chem Lett 12, 1876–1884 (2021).
Solanki, A., Guerrero, A., Zhang, Q., Bisquert, J. & Sum, T. C. Interfacial Mechanism for Efficient Resistive Switching in Ruddlesden–Popper Perovskites for Non-volatile Memories. J Phys Chem Lett 11, 463–470 (2020).
Ji, X., Dong, Z., Lai, C. S., Zhou, G. & Qi, D. A physics-oriented memristor model with the coexistence of NDR effect and RS memory behavior for bio-inspired computing. Mater Today Adv 16, 100293 (2022).
Gao, Q. et al. Stability and Repeatability of a Karst-like Hierarchical Porous Silicon Oxide-Based Memristor. ACS Appl Mater Interfaces 11, 21734–21740 (2019).

No competing interests reported.

Download PDF

Journal Publication

published 16 Nov, 2023

Read the published version in Scientific Reports →

Editorial decision: Major revision
31 Aug, 2023
Reviews received at journal
16 Aug, 2023
Reviewers agreed at journal
14 Aug, 2023
Reviewers agreed at journal
06 Aug, 2023
Reviewers invited by journal
06 Aug, 2023
Editor assigned by journal
06 Aug, 2023
Editor invited by journal
05 Aug, 2023
Submission checks completed at journal
05 Aug, 2023
First submitted to journal
02 Aug, 2023

You are reading this latest preprint version

Resistive Switching Transparent SnO2 Thin Film Sensitive to Light and Humidity

Status:

Journal Publication

Version 1

Abstract

Figures

1. INTRODUCTION

2. EXPERIMENT

3. CHARACTERIZATION

4. RESULT AND DISCUSSION

4.1. Mechanisms of photoconductance (The PPC and NPC effect)

4.2. The coexistence of the NDR effect and RS behavior

4.2.1 Fitting the current–voltage curves with the Space Charge Limited Current (SCLC) model

4.3. Stability

4. CONCLUSION

Declarations

Data availability

References

Additional Declarations

Status:

Journal Publication

Version 1