Surface Morphology and Memristive Characteristics: A Comparative Analysis of Polymer-Quantum Dot Composites for Next-Generation Electronics

Abstract

This article provides a comprehensive comparative analysis of how different surface morphologies in polymer-quantum dot (PQD) composites directly influence critical memristive characteristics. Tailored for researchers and scientists in materials science and nanoelectronics, it explores the foundational mechanisms of resistive switching in PQD structures, details synthesis methodologies and material selection for targeted applications, addresses key performance challenges and optimization strategies, and establishes a rigorous framework for the validation and comparative assessment of diverse PQD morphologies. The review synthesizes these aspects to guide the development of high-performance, reliable memristive devices for neuromorphic computing and flexible electronics.

Fundamental Principles and Switching Mechanisms in PQD Memristors

The conventional computing architecture, known as the Von Neumann architecture, faces substantial limitations due to the physical separation of memory and processing units. This separation creates a performance and energy efficiency constraint commonly referred to as the "Von Neumann bottleneck" [1] [2]. As computational demands intensify, particularly with the exponential growth of artificial intelligence (AI) applications, this bottleneck has become a critical challenge for next-generation computing systems. Brain-inspired neuromorphic computing has emerged as a promising alternative, drawing inspiration from the human brain's massively parallel and energy-efficient architecture [2].

Memristors, theorized by Leon Chua in 1971 as the fourth fundamental circuit element, have gained prominence as key building blocks for neuromorphic systems [1] [2]. These nanoscale electronic components possess a unique property: their instantaneous resistance depends on the history of voltage and current that has passed through them, enabling them to "remember" their past resistance states [1]. This memory characteristic, combined with their simple two-terminal metal-insulator-metal (MIM) structure, makes memristors exceptionally suited for emulating biological synapses in artificial neural networks (ANNs), thereby paving the way for non-Von Neumann computing paradigms that integrate memory and processing [3] [1].

Comparative Performance of Memristive Technologies

Various material systems and device structures have been explored for memristive applications. The performance characteristics of these technologies vary significantly based on their switching mechanisms and material compositions. The table below provides a structured comparison of key memristive technologies based on recent research findings.

Table 1: Performance Comparison of Emerging Memristive Devices

Material System	Switching Mechanism	Switching Speed	Endurance (Cycles)	Resistance Ratio (HRS/LRS)	Key Advantages
h-BN on GaN nano-cones [3]	Multiple nano-filament confinement	Not Specified	Stable consecutive operation demonstrated	Highly linear and symmetric	Ideal for synaptic weight updates, excellent analog switching
Ultra-thin h-BN [4]	Metallic filament (Ti)	120 ps	Confirmed via cycling endurance	>50x	Fastest among 2D memristors, low switching energy (2pJ)
Ta/HfO₂/Pt [5]	Ta-rich conduction channel	≤5 ns	1.2×10¹¹ (120 billion)	Programmable to 24 discrete levels	Record high endurance, reliable retention
SnO₂₋ₓ [6]	Oxygen vacancy filament	Not Specified	11×10³	10⁵ (electrical), 10⁴ (pressure)	Pressure-sensitive, suitable for in-memory computing logic
ZnO [7]	Oxygen vacancy migration	Not Specified	1,000 (single device), 20,000 (crossbar)	~1.8 (single device), ~2.5 (crossbar)	Optical transparency (~86%), stable for neuromorphic vision

Analysis of Performance Trends

The comparative data reveals distinct performance trade-offs across different material systems. Hexagonal Boron Nitride (h-BN) demonstrates exceptional potential for neuromorphic computing applications, with atomically thin layers enabling ultra-fast switching speeds as low as 120 picoseconds and low energy consumption of approximately 2pJ [4]. The unique approach of multiple nano-filament confinement in h-BN devices grown on GaN nano-cones facilitates linear and symmetric analog switching behavior, which is crucial for precise synaptic weight updates in artificial neural networks [3].

Metal oxide-based systems such as Ta/HfO₂/Pt exhibit extraordinary endurance capabilities, sustaining up to 120 billion switching cycles while maintaining the ability to be programmed to 24 discrete resistance levels [5]. This combination of reliability and multi-level capacity makes them suitable for both non-volatile memory and neuromorphic computing applications. Transparent memristors based on ZnO films offer additional functionality for specialized applications in machine vision and transparent electronics, with reasonable endurance characteristics and consistent resistance ratios [7].

Experimental Protocols and Methodologies

Device Fabrication and Material Synthesis

The performance characteristics of memristive devices are heavily influenced by their fabrication processes and material synthesis techniques.

Table 2: Key Fabrication Methods for Memristive Devices

Device Type	Fabrication Method	Key Process Parameters	Structural Characteristics
h-BN Memristor [3]	Metal-organic chemical vapor deposition (MOCVD)	N₂ or H₂ carrier gas, high-temperature growth (>1000°C)	Suspended film over GaN nano-cones (with H₂), uniform atomically flat film (with N₂)
Ultra-thin h-BN Memristor [4]	Chemical vapor deposition (CVD) on copper foil	Transfer to Si/SiO₂ substrate, 7-8 layers	Intrinsic atomic defects enabling filament formation
Ta/HfO₂/Pt Memristor [5]	Sputtering and lithography	5 nm HfO₂ switching layer, 20 nm Pt BE, 50 nm Ta TE	Cross-point structure, sub-10 nm Ta-rich channel
SnO₂₋ₓ Memristor [6]	Hydrothermal synthesis	Low-temperature (160°C), ~13 μm thick film	Oxygen-deficient composition (12.31% vacancies)
ZnO Memristor [7]	RF magnetron sputtering	75 W power, Ar atmosphere, room temperature	Nanocrystalline structure, 60 nm thickness

Electrical Characterization Protocols

Standardized electrical characterization methodologies are essential for evaluating memristive switching performance:

• Forming Process: A initial electroforming step is typically required to activate the switching capability in pristine devices. This involves applying a higher voltage sweep (e.g., ~2.02V for Ta/HfO₂/Pt devices) to generate the initial conductive pathways [5].

• DC Sweep Measurements: Quasi-static current-voltage (I-V) characteristics are measured through bipolar voltage sweeps. For bipolar devices, positive voltages set the device to low-resistance state (LRS), while negative voltages reset it to high-resistance state (HRS) [4] [5].

• Pulse Switching Endurance Testing: Devices are subjected to repeated SET/RESET operations using voltage pulses with specific amplitudes, widths, and rise/fall times. Pulse width testing determines the ultimate switching speed capabilities of the devices [4] [5].

• Retention Measurement: The stability of LRS and HRS is evaluated by monitoring resistance states over time at various temperatures, often followed by extrapolation using Arrhenius equations to predict long-term data retention [5].

• Multi-level Capability Assessment: The ability to program intermediate resistance states is tested through compliance current control during SET operations or stop voltage control during RESET operations [5].

The following diagram illustrates the typical experimental workflow for memristive device characterization:

Resistive Switching Mechanisms

The operation of memristive devices relies on various physical mechanisms that modulate the resistance of the switching layer. Understanding these mechanisms is crucial for optimizing device performance and reliability.

Filamentary Switching Mechanisms

Conductive Filament Formation: In electrochemical metallization memory (ECM) systems, the formation and dissolution of metallic filaments (e.g., Ag or Cu) between electrodes is responsible for resistance switching [5]. In valence change memory (VCM) systems, the motion of oxygen anions (or oxygen vacancies) leads to valence changes of metal cations and consequent resistance changes [5].

Multiple Nano-filament Confinement: Recent advances in h-BN memristors demonstrate that confining multiple nanoscale filaments between suspended h-BN films and the apexes of GaN nano-cones enables analog switching behavior, reducing cycle-to-cycle variation and ensuring stable operation [3]. This approach facilitates controlled formation of multiple nano-filaments within confined geometries, which is essential for linear and symmetric synaptic weight updates in neural networks.

Metallic Filament Characteristics: In ultra-thin h-BN memristors, conductive atomic force microscopy (CAFM) measurements reveal isolated regions exhibiting high current, clearly indicating the formation of conductive filaments [4]. Electron Energy Loss Spectroscopy (EELS) confirms the presence of titanium ions migrated into the h-BN layer, constituting metallic filaments responsible for resistive switching [4].

Interface-Based Switching Mechanisms

Ferroelectric Switching: In topological semimetals such as (TaSe₄)₂I, ferroelectric surface states enable memristive switching through electric-switchable barrier heights at metal-semimetal contacts [8]. This mechanism leverages surface reconstruction-induced ferroelectric polarization, which persists despite the metallic nature of the bulk material.

Schottky Barrier Modulation: In oxide-based memristors such as SnO₂₋ₓ, the switching behavior can be attributed to the tuning of Schottky barrier height between the metal electrode and insulating layer under external bias, modulating the device conductivity between HRS and LRS [6].

The following diagram illustrates the primary resistive switching mechanisms in memristive devices:

The Scientist's Toolkit: Essential Research Reagents and Materials

Successful research in memristive devices requires specific materials and characterization tools. The following table outlines essential components for experimental work in this field.

Table 3: Essential Research Materials and Tools for Memristive Device Development

Category	Specific Materials/Tools	Function/Purpose	Examples from Research
Switching Layer Materials	h-BN, HfO₂, ZnO, SnO₂₋ₓ	Forms the active region where resistive switching occurs	h-BN for fast switching [4], HfO₂ for high endurance [5]
Electrode Materials	Au, Pt, Ti, Ta, ITO, Ag	Provides electrical contact, influences filament formation	Inert Au electrodes [4], Active Ti electrodes [4], Transparent ITO [7]
Growth/Deposition Equipment	MOCVD, RF Magnetron Sputtering, Hydrothermal Synthesis	Creates thin films with controlled composition and morphology	MOCVD for h-BN [3], Sputtering for ZnO [7]
Characterization Tools	SEM, TEM, XPS, AFM, CAFM	Analyzes structural, chemical, and morphological properties	TEM for cross-sectional imaging [4], XPS for chemical states [3]
Electrical Measurement Systems	Semiconductor Parameter Analyzer, Pulse Generator	Measures I-V characteristics, switching speed, endurance	Keithley 4200-SCS [7], Custom RF test setup [4]

Memristive devices represent a transformative technology for overcoming the fundamental limitations of conventional computing architecture. The comparative analysis presented in this guide demonstrates that diverse material systems offer distinct advantages for specific applications. h-BN-based devices provide exceptional switching speed and analog behavior ideal for neural network implementations, while metal-oxide systems such as Ta/HfO₂/Pt offer unprecedented endurance for non-volatile memory applications.

The future development of memristive technology will likely focus on optimizing filament control mechanisms, enhancing device uniformity, and developing integration strategies with existing CMOS technology. As research advances, memristor-based neuromorphic systems are poised to enable next-generation AI computing with significantly improved energy efficiency and computational density, ultimately bridging the gap between biological and electronic information processing [3] [1]. The ongoing exploration of novel materials and switching mechanisms continues to expand the possibilities for brain-inspired computing and artificial perception systems.

The Role of Quantum Dots as Charge Trapping Centers in Polymer Matrices

In the evolving landscape of non-volatile memory (NVM) technologies, quantum dot (QD)-polymer nanocomposites have emerged as a transformative material system, leveraging quantum confinement effects to achieve exceptional charge trapping capabilities. These composites combine the unique optoelectronic properties of quantum dots with the processability and flexibility of polymer matrices, creating ideal platforms for advanced memory devices such as memristors. The fundamental operating principle hinges upon the role of quantum dots as discrete, tunable charge trapping centers within the insulating polymer matrix, enabling precise control over resistive switching behavior through charge trapping and de-trapping mechanisms [9] [10]. This review systematically compares the memristive characteristics of different QD-polymer composite systems, with particular focus on how QD surface morphology, composition, and concentration within the polymer host influence critical performance parameters including switching ratio, endurance, retention, and operating power.

Comparative Analysis of QD-Polymer Composite Memristive Characteristics

Performance Metrics Across Different QD-Polymer Systems

Table 1: Comparative performance of different QD-polymer composite memristors

QD Material	Polymer Matrix	ON/OFF Ratio	Endurance (Cycles)	Retention Time	Operating Voltage	Charge Trap Depth
CdS QDs [11]	Not specified	High	>300	>60,000 s	Low	Not specified
CdSe/ZnS Core/Shell CQDs [10]	PMMA	~10⁴	400	>10⁴ s	2.5-3 V (Set)	Not specified
TiO₂ Nanofillers [12]	Low-density Polyethylene (LDPE)	Not specified	Not specified	Not specified	Not specified	0.9 eV
ZnO [7]	Not applicable	~1.8-2.5	1,000-20,000	Not specified	±3 V	Not specified

Role of Quantum Dot Characteristics in Memristive Performance

The performance of QD-polymer composite memristors is governed by several fundamental characteristics of the quantum dots themselves. Size-dependent quantum confinement in QDs (typically 2-10 nm) creates discrete energy levels that facilitate precise charge trapping and de-trapping operations, a fundamental advantage over continuous floating gate materials [9]. The composition and surface morphology of QDs significantly influence charge trapping dynamics; for instance, CdSe/ZnS core-shell QDs demonstrate improved charge retention compared to pristine CdS QDs, which require a polymer matrix to exhibit memristive behavior [11] [10]. The concentration and dispersion of QDs within the polymer matrix critically affect performance, with optimal concentrations around 1 wt% for CdSe/ZnS QDs in PMMA achieving high ON/OFF ratios of ~10⁴, while higher concentrations can lead to agglomeration and degraded performance [10].

Experimental Protocols for QD-Polymer Memristor Characterization

Fabrication Methodologies

Solution Processing of QD-Polymer Composites

The fabrication of QD-polymer composite memristors typically employs solution-processing techniques that enable uniform dispersion of quantum dots within the polymer matrix. In a representative protocol for CdSe/ZnS QD-PMMA composites, PMMA powder is first dissolved in toluene at a concentration of 4 wt% with thorough stirring [10]. Colloidal CdSe/ZnS QDs with oleic acid ligands (size 6-8 nm with ZnS shell thickness of 2-3 nm) are then added at precisely controlled concentrations (1, 3, or 5 wt%), followed by sonication to achieve uniform dispersion [10]. The solution is spin-coated onto pre-cleaned ITO-coated glass substrates, followed by thermal annealing at 110°C for 20 minutes to remove residual solvent [10]. Finally, top electrodes (typically 150 nm thick aluminum) are deposited via thermal evaporation to complete the device structure [10].

Advanced Composite Integration Techniques

For perovskite QD-polymer systems, more specialized processing approaches have been developed. The CsPbBr₃ PQD-integrated covalent organic framework (COF) nanocomposite synthesis involves a two-step process: first, CsPbBr₃ PQDs are synthesized via a modified hot-injection method using lead(II) bromide and cesium bromide precursors in dimethylformamide (DMF) with oleic acid and oleylamine capping ligands [13]. Subsequently, the COF matrix is formed via Schiff-base condensation between 1,3,5-tris(4-aminophenyl)benzene (TAPB) and 2,5-dihydroxyterephthalaldehyde (DHTA), with PQD integration achieved through in-situ embedding during COF formation [13].

Electrical Characterization Methods

Standardized electrical characterization protocols are essential for comparing memristive performance across different QD-polymer systems. Current-voltage (I-V) characteristics are typically measured using semiconductor parameter analyzers (e.g., Keithley 4200-SCS or Agilent 4155C) with voltage sweep sequences applied between top and bottom electrodes while monitoring current response [7] [10]. Endurance testing involves repeated program/erase cycles (set/reset operations) to determine the number of sustainable switching cycles before performance degradation, with measurements often conducted at room temperature in air or inert atmospheres [11] [7]. Retention testing evaluates the ability of devices to maintain resistance states over extended periods (typically 10⁴ seconds or longer) under constant voltage stress or periodic reading [11] [10]. Conduction mechanism analysis employs mathematical modeling of I-V characteristics to identify dominant charge transport mechanisms (e.g., trap-free space charge-limited current vs. trap charge-limited current) and their evolution during repeated switching operations [7] [10].

Charge Trapping Mechanisms and Signaling Pathways in QD-Polymer Composites

Fundamental Charge Trapping Dynamics

Table 2: Charge trapping mechanisms in QD-polymer composites

Mechanism	Process Description	Effect on Memristive Properties
Electron Trapping/De-trapping	Electrons injected from electrodes become trapped at QD sites due to band offset with polymer matrix, modifying conductivity [10]	Controls resistive switching between HRS and LRS
Field-Enhanced Vacancy Generation	Electric field concentration at electrode edges promotes oxygen vacancy formation in metal oxide QDs [7]	Enables conductive filament formation in metal oxide-based systems
Surface Defect Passivation	Chemical treatment (e.g., Ag epoxy paste) passivates surface defects on QDs, modifying charge trapping efficiency [14]	Enhances photoluminescence quantum yield and charge retention
Interface-Mediated Trapping	Charge trapping at QD-polymer interfaces influenced by surface ligands and interfacial chemistry [12]	Determines charge trap depth and retention characteristics

Visualization of Charge Trapping Pathways

The following diagram illustrates the fundamental charge trapping and conduction mechanisms in QD-polymer composite memristors:

The charge trapping pathway begins with voltage application, prompting electron injection from electrodes into the composite layer. These electrons become trapped at quantum dot sites due to the favorable band alignment between the QDs and polymer matrix. As trapping continues, conductive pathways form through the percolation of charged QDs or field-induced structural changes, switching the device to a low-resistance state (LRS). The reverse process (de-trapping) occurs under opposite voltage polarity, resetting the device to its high-resistance state (HRS) [11] [10].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Essential research reagents for QD-polymer memristor development

Material Category	Specific Examples	Function in Research
Quantum Dots	CdS QDs, CdSe/ZnS CQDs, CsPbBr₃ PQDs, TiO₂ nanofillers [11] [12] [13]	Primary charge trapping centers with tunable bandgaps via size control
Polymer Matrices	PMMA, Polydimethylsiloxane (PDMS), Low-density Polyethylene (LDPE) [14] [12] [10]	Insulating host material providing structural framework and processability
Electrode Materials	ITO (Indium Tin Oxide), Aluminum, Silver Epoxy Paste [14] [7] [10]	Charge injection contacts with specific work functions for asymmetric operation
Solvents & Ligands	Toluene, Dimethylformamide (DMF), Oleic Acid, Oleylamine [13] [10]	Dispersion medium and surface modification agents for QD stability
Passivation Agents	Silver Epoxy Paste (AEP) [14]	Surface defect passivation for enhanced quantum yield and charge retention

QD-polymer composites represent a versatile platform for advanced memristive devices, with performance characteristics directly tunable through rational material selection and processing optimization. The comparative analysis presented herein demonstrates that key memristive parameters—including ON/OFF ratio, endurance, retention, and operating power—are strongly influenced by QD composition, surface morphology, and integration methodology within polymer matrices. Core-shell QD architectures and appropriate surface passivation strategies emerge as particularly promising approaches for enhancing charge trapping efficiency and device stability. As research progresses toward more sophisticated control of interfacial phenomena and charge trapping dynamics, QD-polymer composites are poised to enable increasingly efficient, scalable, and multifunctional memory technologies for next-generation neuromorphic computing and transparent electronics applications.

Semiconductor nanocrystals, including quantum dots (QDs) and two-dimensional (2D) materials, have emerged as pivotal building blocks for next-generation electronic and optoelectronic devices. Their unique size-dependent and composition-tunable electronic properties make them powerful candidates for advanced applications, including memristive devices, where surface morphology and interface characteristics critically determine performance. This guide provides a comparative analysis of the electronic properties of three significant material systems: the classic CdSe/ZnS core-shell QDs, newer nanocrystals (NCQs) like perovskites and cadmium-free alternatives, and tungsten disulfide (WS2) as a representative 2D material. Framed within the context of memristive characteristics research, this comparison synthesizes synthesis methodologies, key electronic parameters, and experimental data to inform material selection and device engineering for researchers and scientists.

CdSe/ZnS Core-Shell Quantum Dots

Cadmium Selenide with a Zinc Sulfide shell (CdSe/ZnS) represents one of the most mature and extensively studied quantum dot systems [15]. Its electronic structure is characterized by a type-I core-shell heterostructure, where the wider bandgap ZnS shell effectively confines charge carriers (electrons and holes) to the CdSe core [15]. This confinement is the origin of its superior optoelectronic properties: a tunable bandgap from 1.7 eV to 2.6 eV (governed by the quantum size effect), high photoluminescence quantum yield (PLQY) often exceeding 90%, and narrow emission linewidths of 20–30 nm [15]. From a memristive perspective, the ZnS shell plays a dual role: it passivates surface defects on the CdSe core, reducing charge trap densities, and acts as a controlled charge transport barrier. The organic ligand shell (e.g., oleic acid, trioctylphosphine oxide) further influences electronic coupling between dots in films, directly affecting resistivity and switching behavior [16].

Emerging Nanocrystals (NCQs)

The class of Nanocrystal Quantum Dots (NCQs) has expanded beyond CdSe to include materials like perovskite nanocrystals (e.g., CsPbBr3) and cadmium-free alternatives (e.g., ZnSeTe/CdZnSe). These materials exhibit diverse electronic band alignments. For instance, the ZnSeTe/CdZnSe-based QD is a type-II core-shell system, where the conduction band of the ZnSeTe core and the CdZnSe shell facilitate spatial separation of electrons and holes [17]. This charge separation can be advantageous for memristive switching, which relies on charge trapping and de-trapping. These NCQs can achieve high PLQY (up to 95%) with low cadmium content (<2.5%) and demonstrate tunable emission from 430 nm to 510 nm [17]. Furthermore, advanced synthesis techniques provide molecular-level insight into NC formation, enabling precision control over size, shape, and composition, which are critical for tailoring electronic properties and surface states for memristive applications [18].

Tungsten Disulfide (WS2)

Tungsten Disulfide (WS2) is a transition metal dichalcogenide (TMDC) that exists in both bulk and 2D forms. Unlike the zero-dimensional QDs, 2D WS2 exhibits a layer-dependent bandgap, transitioning from an indirect bandgap (~1.4 eV) in bulk to a direct bandgap (~2.1 eV) in monolayers. Doping is a primary method for tuning its electronic properties. For example, substitutional doping with Sn atoms introduces n-type doping behavior, modifying its charge transport characteristics [19]. Studies show Sn-doped WS2 layers can have a bandgap of 2.09 eV, a mobility of 7.84 cm²/V·s, and a resistivity of 2.81 × 10³ Ω·cm [19]. The 2D nature of WS2 provides a distinct morphology for memristive devices, offering large, atomically flat surfaces for uniform electric field distribution and unique defect types, such as sulfur vacancies, which can act as charge trapping sites.

Table 1: Comparative Electronic Properties for Memristive Applications

Property	CdSe/ZnS QDs	NCQs (e.g., ZnSeTe/CdZnSe)	WS2 (Sn-doped)
Dimensionality	0D (Core-Shell)	0D (Core-Shell)	2D (Layered)
Bandgap Type	Type-I	Type-II	Direct (Monolayer)
Bandgap Range (eV)	1.7 - 2.6 [15]	~2.7 - 2.4 (430-510 nm) [17]	2.09 (for Sn-doped) [19]
Bandgap Tunability	Core Size	Shell Composition & Core Size	Layer Number & Doping
Charge Transport	Carrier Confinement	Charge Separation	n-type (with Sn doping) [19]
Mobility (cm²/V·s)	Not Specified	Not Specified	7.84 [19]
Resistivity (Ω·cm)	Not Specified	Not Specified	2.81 × 10³ [19]
Key Memristive Feature	Defect-passivated core, stable switching	Engineered charge transfer via band alignment	Dopant/vacancy controlled trapping

Comparative Experimental Data

The following table consolidates key experimental data and synthesis parameters from recent studies, providing a direct comparison of performance metrics relevant to device integration and characterization.

Table 2: Experimental Synthesis and Performance Data

Parameter	CdSe/ZnS (Oleylamine-Modified) [20]	ZnSeTe/CdZnSe NCQ [17]	WS2 (Sn-doped) [19]
Synthesis Method	One-pot hot-injection with OLA [20]	Colloidal synthesis	Co-sputtering (RF & DC)
Emission Peak (nm)	505 - 610 [20]	430 - 510 [17]	N/A
Photoluminescence Quantum Yield (PLQY)	High (Specific value not given)	95% [17]	N/A
Full Width at Half Maximum (FWHM)	<26 nm (Inferred from similar QDs) [15]	<26 nm [17]	N/A
External Quantum Efficiency (EQE)	>20% (for red/green QLEDs) [15]	5% (Blue QLED) [17]	N/A
Thermal/Environmental Stability	Improved with ZnS/ZnO shell [20]	Not Specified	Not Specified
Key Synthesis Insight	OLA content controls QD size & shell formation [20]	Low Cd content (<2.5%) shell [17]	Sn fills W vacancies and substitutes W atoms [19]

Detailed Experimental Protocols

This protocol is central to producing high-quality, stable QDs with tunable properties.

• Materials Preparation: The precursors are prepared separately. A mixture of trioctylphosphine (TOP), sulfur (S), and selenium (Se) is prepared via ultrasonication. In a four-neck flask, cadmium oxide (CdO), zinc acetate [Zn(OAc)₂], oleic acid (OA), 1-octadecene (ODE), and variable amounts of oleylamine (OLA; e.g., 0, 3, 5, 10 mL) are combined.
• Reaction Process: The mixture in the four-neck flask is heated to 180 °C under an inert atmosphere (e.g., Argon) until a clear solution is obtained, indicating the dissolution of precursors. The temperature is then raised to 300 °C. At this point, the pre-mixed TOP-Se-S solution is swiftly injected into the reaction flask.
• Growth and Termination: The reaction temperature is immediately lowered to 280 °C and maintained for 180 seconds to allow for core formation and shell growth. The reaction is rapidly quenched by injecting ice-cold toluene.
• Purification: The crude solution is mixed with alcohol and acetone to induce precipitation. The QDs are recovered as a pellet via centrifugation at 7000 rpm for 10 minutes. The pellet can be redispersed in a non-polar solvent for further use.

Synthesis of Oleylamine-Modified QDs

This technique offers precise control over dopant distribution in 2D materials.

• Substrate Preparation: A suitable substrate (e.g., SiO2/Si) is cleaned using standard procedures to remove organic and particulate contaminants.
• Sputtering Setup: The substrate is placed in a high-vacuum sputter deposition chamber. A WS2 target is connected to a radio frequency (RF) magnetron sputtering source. A metallic Sn target is connected to a DC sputtering source.
• Deposition Process: The chamber is evacuated to a high base vacuum. An argon plasma is ignited. The RF source is activated to sputter WS2, forming a thin film on the substrate. Simultaneously, the DC source is activated to sputter Sn atoms. The power and duration of the DC source control the concentration and distribution of Sn dopants.
• Post-Processing: The as-deposited Sn-doped WS2 film may undergo post-annealing in a controlled atmosphere (e.g., inert gas or sulfur vapor) to improve crystallinity and activate the dopants.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Reagents and Materials for PQD Synthesis and Fabrication

Item Name	Function/Application	Examples / Key Details
Oleylamine (OLA)	Surfactant & Reaction Medium	High boiling point polar solvent; controls nanocrystal size, morphology, and stability during synthesis of CdSe and perovskite NCQs [20].
Trioctylphosphine (TOP)	Precursor Solvent & Ligand	Dissolves chalcogen precursors (S, Se) and acts as a surface ligand, coordinating to the NC surface to provide colloidal stability [18] [20].
Trioctylphosphine Oxide (TOPO)	Surface Capping Ligand	A common coordinating solvent and capping agent for synthesizing CdSe QDs, passivating surface states [16].
Cadmium Oxide (CdO) & Zinc Acetate [Zn(OAc)₂]	Metal Precursors	Source of Cd²⁺ and Zn²⁺ ions for the formation of the core and shell in CdSe-based QDs [20].
Selenium (Se) & Sulfur (S)	Chalcogen Precursors	React with metal precursors to form the semiconductor material (e.g., CdSe, ZnS) [20].
1-Octadecene (ODE)	Non-coordinating Solvent	High-boiling-point non-polar solvent used as a reaction medium in hot-injection syntheses [20].
Sputtering Targets (WS2, Sn)	Solid-State Precursors	High-purity solid targets used in physical vapor deposition (co-sputtering) to fabricate WS2 thin films and introduce Sn dopants [19].

The electronic properties of PQDs are intimately tied to their composition, dimensionality, and surface morphology, each offering distinct advantages for memristive research. CdSe/ZnS QDs provide a well-understood, stable platform with excellent luminescence and tunable confinement, ideal for fundamental studies on charge trapping in 0D systems. Emerging NCQs, like the type-II ZnSeTe/CdZnSe, introduce engineered charge separation and reduced toxicity, opening pathways for novel switching mechanisms based on interfacial charge transfer. In contrast, WS2 represents the 2D frontier, where layer thickness and substitutional doping provide powerful knobs for controlling resistivity and trap densities in a planar geometry. The choice of material system depends critically on the target memristive characteristics—whether priority is placed on high on/off ratios, switching speed, stability, or biocompatibility. Future research will likely focus on hybrid structures that combine the strengths of these different material classes to create devices with superior and novel memristive functionalities.

The pursuit of advanced memristive technologies has intensified the investigation into underlying resistive switching mechanisms. Two dominant models have emerged: trap-assisted Space Charge Limited Conduction (SCLC) and conductive filament formation. Understanding the distinctions is crucial for designing next-generation memory and neuromorphic computing systems. This guide provides a comparative analysis of these mechanisms, focusing on their memristive characteristics, with specific experimental data from multilayer MoS₂ memristors and Perovskite Quantum Dot (PQD) systems. The performance of devices governed by these mechanisms is objectively compared to inform material selection and device engineering for researchers and scientists.

Comparative Analysis of Switching Mechanisms

The fundamental difference between the two mechanisms lies in the nature of the current path and its formation. The table below summarizes the core characteristics.

Table 1: Fundamental Comparison of Switching Mechanisms

Characteristic	Trap-Assisted SCLC	Filament Formation
Conduction Principle	Bulk-limited conduction where current is limited by space charges and trap states [21] [22].	Electrode-limited conduction via localized, nanoscale conductive filaments [21].
Nature of Switching	Typically homogeneous and uniform across the active area [21].	Inhomogeneous and stochastic, dependent on filament formation/rupture [21].
Current Path	Area-dependent, forming across the entire electrode interface [21].	Area-independent, confined to the filamentary path [21].
Cycle-to-Cycle Uniformity	High uniformity due to bulk mechanism [21].	Often suffers from variability and stochasticity [21].
Key Material Properties	Density and energy depth of trap states; carrier mobility [23] [24].	Mobility of metal ions (e.g., Ag⁺, Cu²⁺) or density of anion vacancies (e.g., O²⁻) [21].

Case Study: Trap-Assisted SCLC in Multilayer MoS₂ Memristors

Recent research on Au/Ti/Multilayer (ML) MoS₂/Au memristors provides a clear example of a non-filamentary, trap-assisted SCLC mechanism.

Experimental Evidence and Workflow

The conduction mechanism was probed through a combination of electrical characterization and materials analysis. The workflow below illustrates the experimental process for identifying trap-assisted SCLC.

Quantitative Performance Data

The multilayer MoS₂ devices exhibiting trap-assisted SCLC demonstrated excellent performance metrics, particularly in uniformity and analog switching, which are critical for neuromorphic computing applications like reservoir computing [21].

Table 2: Performance Metrics of ML-MoS₂ Memristors with Trap-Assisted SCLC

Performance Metric	Reported Value/Characteristic	Implication
Resistance Switching	Non-volatile	Enables long-term memory retention [21].
Conductance Tuning	Analog behavior	Permits fine-grained weight updates for neural networks [21].
Cycle-to-Cycle (C2C) Variation	< 4% standard error	High reproducibility and reduced computational errors [21].
Crossbar Array Scale	16×16	Demonstrated scalability for integrated systems [21].
Application Performance	89.56% precision in spoken-digit recognition	Validates performance in real-world tasks [21].

Case Study: Mitigating Trap States in Perovskite Quantum Dot (PQD) Morphologies

While traps are central to SCLC in memristors, they are often a source of efficiency loss in photovoltaics. Research on PQD surface morphologies provides insights into controlling trap states, which is complementary to memristor studies.

Surface Engineering Strategies and Outcomes

Different approaches to PQD surface and packing morphology directly influence trap-state density and device performance. The comparative data below illustrates this relationship.

Table 3: Impact of PQD Surface and Packing Morphology on Trap States

PQD Morphology / Treatment	Key Experimental Methodology	Effect on Trap-States & Recombination	Resulting Device Performance
Binary-Disperse Mixed Packing (10 nm & 14 nm QDs) [24]	GISAXS for packing analysis; Molecular dynamics simulations; J-V characterization.	"Largely suppressed trap-assisted recombination" and "much longer carrier lifetime" due to closer face-to-face contact and enhanced tunneling [24].	PCE: 14.42%; JSC: 17.08 mA cm⁻²; VOC: 1.19 V; FF: 71.12% [24].
Alkaline-Augmented Antisolvent Rinsing (KOH + MeBz) [25]	Alkali-augmented antisolvent hydrolysis (AAAH); FTIR and XPS spectroscopy; J-V characterization.	Assembly of films with "fewer trap-states" and "suppressed trap-assisted recombination" due to more complete conductive capping [25].	PCE: 18.37% (Certified 18.30%); Improved storage and operational stability [25].
Conventional Ester Antisolvent Rinsing (e.g., MeOAc) [25]	Rinsing with neat ester antisolvents under ambient humidity; J-V characterization.	Generates "extensive surface vacancy defects to capture carriers" due to inefficient ligand exchange and incomplete surface capping [25].	Lower PCE and compromised stability compared to AAAH-treated devices [25].

Experimental Workflow for PQD Surface Treatment

The following diagram outlines the key steps in the advanced Alkali-Augmented Antisolvent Hydrolysis (AAAH) protocol used to achieve low-trap-density PQD films.

The Scientist's Toolkit: Essential Research Reagents and Materials

The following table catalogues key reagents and materials used in the cited studies for investigating and engineering charge transport mechanisms.

Table 4: Key Research Reagents and Materials for Memristor and PQD Studies

Research Reagent / Material	Function in Experimental Protocols	Relevant Study Context
Chemical Vapor Deposited (CVD) MoS₂	Active material for memristive devices; thickness (monolayer vs. multilayer) dictates volatile vs. non-volatile switching [21].	MoS₂ Memristors [21]
Methyl Benzoate (MeBz) Antisolvent	Ester-based antisolvent for rinsing PQD solid films; hydrolyzes to form conductive benzoate ligands that passivate the PQD surface [25].	PQD Surface Engineering [25]
Potassium Hydroxide (KOH)	Alkaline additive to create an alkaline environment during antisolvent rinsing, dramatically enhancing the hydrolysis rate and efficiency of ester antisolvents [25].	PQD Surface Engineering [25]
Oleic Acid (OA) / Oleylamine (OAm)	Pristine long-chain insulating ligands used in the synthesis of PQDs to control growth; replaced by shorter ligands to enhance charge transport [24] [25].	PQD Synthesis & Ligand Exchange [24] [25]
Formamidinium (FA⁺) / Guanidinium (GA⁺)	Short-chain basic cationic ligands used in post-treatment to substitute pristine OAm⁺ ligands on the A-site of PQDs, improving electronic coupling [25].	PQD Surface Engineering [25]

The choice between engineering for trap-assisted SCLC or filament formation has profound implications for memristive device performance. Trap-assisted SCLC, as demonstrated in multilayer MoS₂ memristors, offers high uniformity, analog switching, and superior reliability, making it ideal for neuromorphic computing applications. Conversely, the literature on PQDs highlights that the same trap states which enable SCLC can be detrimental in other applications, driving the development of sophisticated surface and morphological controls, such as binary-size mixing and alkaline-augmented ligand exchange, to suppress trap-assisted recombination. The decision to exploit or mitigate trap-assisted conduction must therefore be guided by the specific performance requirements of the target electronic or optoelectronic application.

The development of advanced non-volatile memory technologies, including those based on perovskite quantum dots (PQDs), relies heavily on the optimization of several key performance metrics. These parameters collectively determine the practical viability, reliability, and integration potential of memory devices for next-generation computing systems. For researchers investigating the memristive characteristics of different PQD surface morphologies, understanding the interplay between these metrics is crucial for guiding material synthesis and device engineering strategies. The performance of memory devices is primarily evaluated through four essential parameters: the ON/OFF ratio (which dictates memory density and sensing margin), retention time (determining non-volatility), endurance (indicating operational lifetime), and operating voltage (affecting power efficiency and compatibility). These metrics are deeply interconnected with the nanoscale material properties and interfacial phenomena of the active layers, making surface morphology a critical factor in performance optimization for PQD-based memory devices.

Performance Metrics Comparison

The following tables consolidate experimental data for key performance metrics across various emerging memory technologies, providing a benchmark for evaluating PQD-based memristive devices.

Table 1: Comparison of Key Performance Metrics for Emerging Memory Technologies

Device Technology	ON/OFF Ratio	Retention Time	Endurance (Cycles)	Operating Voltage
Memristive RF Switch (Ag Filament)	Up to 1012 [26]	>3.3 years (extrapolated) [26]	>1012 [26]	SET: ≤3V, RESET: -0.4V [26]
HfSexOy Memristor	>103 [27]	>15,000 s [27]	Information Missing	<3V [27]
Flexible HfO2/Ta2O5−x Memristor	Information Missing	>104 s [28]	>105 [28]	Information Missing
Cs-Doped β-Ga2O3 Memristor	101 [29]	Information Missing	Information Missing	Information Missing
HfZrO Ferroelectric Memcapacitor	Information Missing	>105 s [30]	>106 [30]	~3V [30]
CdSe QD-Based Resistive Memory	Information Missing	Information Missing	≈105 [31] [9]	Information Missing

Table 2: Performance Advantages of Quantum Dot-Based Memory over Conventional Technologies

Performance Aspect	Quantum Dot-Based Memory	Conventional Floating-Gate Memory
Scalability	Better scaling with discrete charge storage nodes [31] [9]	Limited by gate dielectric thickness and reliability issues [31] [9]
Power Consumption	Lower operating voltages due to efficient charge trapping/detrapping [31] [9]	Higher power consumption due to continuous floating gate [31] [9]
Endurance	Improved endurance with discrete, isolated charge storage nodes [31] [9]	Higher risk of charge leakage and degradation over time [31] [9]
Retention	Enhanced retention with quantized energy levels [31] [9]	Lower retention due to higher charge leakage [31] [9]

Experimental Protocols for Memristive Device Characterization

Standardized experimental protocols are essential for obtaining reliable and comparable performance metrics across different PQD-based memristive devices. This section outlines the key methodologies employed in the evaluation of memristive characteristics.

Electrical Characterization Protocol

The fundamental electrical characterization of memristive devices involves current-voltage (I-V) measurements to determine switching behavior and key performance parameters:

• DC Sweep Measurement: Apply a voltage sweep (e.g., -V_max to +V_max) to the device while measuring current response to observe hysteretic behavior and identify SET/RESET voltages [26].
• ON/OFF Ratio Calculation: Determine the ratio between current in the low-resistance state (LRS or ON state) and high-resistance state (HRS or OFF state) at a fixed read voltage (typically lower than switching voltage) [26] [27].
• Endurance Testing: Apply repetitive voltage pulses to cycle the device between ON and OFF states while monitoring resistance degradation over cycles [28] [30].
• Retention Measurement: Program the device to specific states (ON/OFF) and monitor state stability over time at various temperatures, often using Arrhenius extrapolation for long-term prediction [26].
• Multilevel Cell Operation: For devices capable of intermediate states, apply voltage pulses with varying amplitudes or durations to program multiple distinct resistance levels [30].

Material Synthesis and Device Fabrication

The performance of PQD-based memristive devices is highly dependent on synthesis methods and surface morphology control:

• Quantum Dot Synthesis: Wet-chemical approaches enable precise control over PQD size, composition, and surface properties through manipulation of precursor concentrations, reaction temperatures, and ligand environments [29].
• Surface Functionalization: For biosensing-integrated memristors, surface treatment with oxygen plasma introduces hydroxyl groups for subsequent biofunctionalization with specific antibodies [32].
• Electrode Formation: Photolithography and electron beam lithography define electrode patterns, followed by metal deposition (e.g., Ni, Au, TiN) via evaporation or sputtering [32] [30].
• Interface Engineering: Thermal annealing under controlled atmospheres forms optimal interfacial layers (e.g., NiSi Schottky contacts) and enhances crystallinity [32].
• Morphological Control: Precise tuning of doping concentrations (e.g., Cs in Ga₂O₃) significantly alters nanostructure morphology, transitioning from diamond-shaped to flower-like structures, which directly impacts electrical properties [29].

Diagram 1: Memristive device fabrication and characterization workflow illustrating the sequential process from material synthesis to performance evaluation.

The Scientist's Toolkit: Essential Research Reagents and Materials

This section details critical materials, reagents, and equipment essential for research on PQD-based memristive devices, with particular emphasis on surface morphology studies.

Table 3: Essential Research Materials for PQD Memristive Device Development

Material/Reagent	Function/Application	Research Context
Perovskite Precursors (Cs, Pb, Br, I salts)	Quantum dot synthesis with tunable bandgaps	Enables precise control of optical/electrical properties through composition engineering [31]
Surface Ligands (Oleic acid, oleylamine)	Size stabilization and surface passivation during QD synthesis	Critical for controlling inter-dot spacing and charge transport in QD films [31]
Dopants (Cs, other cations)	Modifying electrical conductivity and switching behavior	Cs doping in Ga2O3 alters morphology and resistive switching ratio [29]
Transition Metal Dichalcogenides (HfSe2)	Base material for low-power memristors	Oxidized to form HfSexOy switching medium with pA-level operation currents [27]
Ferroelectric Materials (HfZrO)	Non-volatile capacitive switching elements	Enables multibit memcapacitors with programmable capacitance states [30]
Electrode Materials (TiN, Au, Ag, Ni)	Forming Ohmic and Schottky contacts	Ag electrodes enable filament formation; Ni forms Schottky contacts in SiNW devices [26] [32]
Bio-functionalization Agents (Anti-PSA antibodies)	Surface modification for biosensing-memristive integration	Enables simultaneous memristive switching and biomolecule detection [32]

Interplay of Surface Morphology and Memristive Characteristics

The surface morphology of PQD films significantly influences memristive performance metrics through several fundamental mechanisms that researchers must consider when designing experiments.

Diagram 2: Relationship between PQD surface morphology characteristics and memristive performance metrics.

The quantum confinement effect in QDs creates discrete energy levels that significantly impact charge retention characteristics. The quantized energy states make it more difficult for trapped charges to escape, thereby enhancing retention times compared to bulk materials [31] [9]. This effect is highly dependent on QD size and size distribution, which are controlled during synthesis through precursor concentrations and reaction conditions.

For resistive switching mechanisms, the formation and rupture of conductive filaments—composed of metal cations or oxygen vacancies—are strongly influenced by film morphology and interface properties. The controlled oxidation of HfSe₂ to HfSe_xO_y creates a switching medium where filament growth is confined to "weak" spots, enabling ultralow operation currents down to 100 pA [27]. Similarly, in nanoscale memristive RF switches, the formation of continuous Ag filaments across a 35-nm air gap enables exceptional ON/OFF ratios up to 10¹² [26].

Doping strategies represent another crucial approach for modifying surface morphology and electrical properties. Cs doping in Ga₂O₃ nanostructures causes dramatic morphological changes from diamond-shaped to flower-like structures, directly impacting the resulting memristive behavior and switching ratios [29]. These morphological modifications alter charge transport pathways and trapping dynamics, ultimately determining key performance metrics.

The comprehensive comparison of key performance metrics across emerging memory technologies reveals both the current state-of-the-art and future research directions for PQD-based memristive devices. The exceptional ON/OFF ratios demonstrated by nanoscale memristive switches (up to 10¹²), combined with the long retention times (exceeding 10⁵ seconds) and excellent endurance (over 10⁶ cycles) of ferroelectric memcapacitors, set ambitious targets for PQD memory development. The ongoing research in quantum dot-based non-volatile memory highlights the significant advantages of these materials, including scalability, reduced power consumption, and enhanced charge retention capabilities [31] [9]. For researchers focusing on PQD surface morphologies, these metrics provide critical benchmarks for evaluating material innovations and device architectures. The experimental protocols and characterization methodologies outlined in this guide serve as essential tools for systematic investigation of the complex relationships between nanoscale morphology and macroscopic memristive performance, ultimately accelerating the development of advanced memory technologies tailored for specific computing applications.

Synthesis, Fabrication, and Application-Specific Design of PQD Memristors

Solution-Processing Techniques for PQD Composite Deposition

Perovskite quantum dot (PQD) composite deposition represents a frontier in the development of next-generation memristive devices, which are promising candidates for non-volatile memory and neuromorphic computing due to their high performance, low power consumption, and multilevel behavior [7] [5]. The surface morphology of the active PQD layer directly influences critical memristive characteristics such as resistive switching uniformity, endurance, and retention properties [7]. Solution-processing techniques have emerged as particularly attractive methods for depositing PQD composite films due to their exceptional potential for scalability, cost-effectiveness, and compatibility with large-area substrate coating [33] [34]. Unlike vacuum-based deposition methods that require sophisticated equipment and high-energy consumption, solution-based techniques can be implemented with relatively simple apparatus and offer superior compositional control through chemical precursor engineering [33] [35].

The performance of memristive devices is profoundly affected by the quality of the active layer, where factors such as film uniformity, thickness control, particle distribution, and interface quality determine the reliability of the resistive switching behavior [7] [5]. For PQD-based memristors, the ability to precisely control the nanoscale morphology during deposition is crucial for achieving reproducible switching characteristics, as localized conduction channels typically form at specific sites within the film [5]. This comprehensive guide objectively compares the performance of various solution-processing techniques for PQD composite deposition, providing researchers with experimental data and methodologies to inform their selection of deposition strategies for memristive device fabrication.

Comparison of Solution-Processing Techniques for PQD Deposition

Various solution-processing techniques have been adapted for PQD composite deposition, each offering distinct advantages and limitations for controlling film morphology. These techniques can be broadly categorized into spin-based, blade-coating, spray-based, and immersion methods, with each approach yielding different film characteristics that ultimately influence the memristive properties of the resulting devices [33] [34]. The table below provides a comprehensive comparison of these techniques based on critical parameters relevant to PQD memristor fabrication.

Table 1: Comparison of Solution-Processing Techniques for PQD Composite Deposition

Technique	Typical Film Thickness Range	Uniformity Control	Scalability	Material Utilization Efficiency	Compatible Substrates	Processing Speed
Spin Coating	Nanometers to microns [33]	Excellent for small, flat substrates [33]	Limited to batch processing; unsuitable for large-scale production [33]	Low (typically <5%); high solution wastage [33]	Small, flat, rigid substrates [33]	Very fast (seconds to minutes) [33]
Dip Coating	Nanometers to sub-microns; thickness dependent on withdrawal speed [33]	Good uniformity achievable with parameter control [33]	Moderate; suitable for research and development [33]	Low; requires solution volume greater than substrate volume [33]	Flat, curved, or tubed substrates [33]	Moderate (minutes) [33]
Spray Coating	Sub-microns to microns; depends on spraying passes [33]	Moderate; requires optimization of spray parameters [33]	High; compatible with roll-to-roll processing [33]	Moderate; overspray can lead to some wastage [33]	Both rigid and flexible substrates [33]	Fast (seconds to minutes, depending on area) [33]
Slot Die Coating	Tens of nanometers to microns [33]	Excellent with parameter optimization [33]	High; compatible with roll-to-roll manufacturing [33]	High; precise solution control minimizes wastage [33]	Both rigid and flexible substrates [33]	High coating speeds achievable [33]
Doctor Blade Coating	Typically >10 microns; limited minimum thickness [33]	Moderate; less precise than spin coating [33]	High; ideal for industrial scale-up [33]	High; lower solution wastage [33]	Both rigid and flexible substrates [33]	Fast; high throughput [33]

Impact on Memristive Characteristics

The choice of deposition technique directly influences PQD film morphology, which in turn affects critical memristive characteristics. Research on metal oxide memristors has demonstrated that films with uniform nanocrystalline structure and relatively smooth surfaces exhibit stable resistive switching behavior [7]. For instance, in zinc oxide-based memristors, the magnetron sputtering power during deposition significantly influences the film's structural and electrophysical properties, with optimal power levels yielding superior resistive switching characteristics [7].

Table 2: Relationship Between Deposition Parameters and Memristive Performance

Deposition Parameter	Influence on PQD Film Morphology	Impact on Memristive Characteristics	Experimental Evidence
Deposition Speed	Affects particle alignment and packing density [33]	Influences switching uniformity and operational stability [7]	Slot die coating at controlled speeds produces uniform films with stable switching for >1000 cycles [7]
Solution Viscosity	Determines film thickness and defect density [33]	Affects forming voltage and conductive filament stability [5]	Higher viscosity solutions in blade coating produce thicker films with controlled filament formation [33]
Drying/Thermal Treatment	Controls crystallinity and residual stress [33]	Impacts retention time and endurance [7]	Proper thermal treatment after deposition improves retention to >10 years at 85°C [5]
Substrate Temperature	Influences nucleation and growth kinetics [33]	Affects interfacial quality and switching voltage distribution [7]	Room temperature deposition yields films with stable resistive switching [7]
Coating Thickness	Determines quantum dot layer density and interface [33]	Influences resistance states and power consumption [5]	~60 nm ZnO films show optimal resistive switching with LRS/HRS ratio of ~1.8 [7]

Experimental Protocols for PQD Deposition and Characterization

Detailed Methodologies for PQD Film Deposition

Spin Coating Protocol for PQD Memristors

Spin coating remains the standard deposition technique for research-scale PQD memristor development due to its exceptional uniformity for small substrates [33]. The procedure begins with substrate preparation, typically involving sequential cleaning of glass/ITO substrates in ultrasonic baths of detergent, deionized water, acetone, and isopropanol (15 minutes each), followed by oxygen plasma treatment for 5-10 minutes to improve wettability. The PQD solution is prepared by dissolving perovskite precursors (e.g., PbI₂ and CH₃NH₃I) in dimethylformamide (DMF) or dimethyl sulfoxide (DMSO) at concentrations ranging from 0.5-1.5 M, with optional additives such as hydrohalic acids to control crystallization kinetics. The solution is then dispensed onto the stationary substrate (50-100 µL depending on substrate size) using a pipette, followed immediately by a two-stage spinning process: initial spreading at 500-1000 rpm for 5-10 seconds, followed by high-speed spinning at 3000-6000 rpm for 20-60 seconds to achieve the desired film thickness [33]. During the second stage, an antisolvent (typically chlorobenzene or toluene) is rapidly dripped onto the spinning substrate 10-20 seconds before completion to initiate controlled crystallization. The film is subsequently annealed on a hotplate at 90-110°C for 10-30 minutes to remove residual solvent and enhance crystallinity. This method produces uniform PQD films with thicknesses ranging from 50-500 nm, suitable for fundamental memristive characterization [33].

Slot Die Coating Protocol for Large-Area PQD Arrays

For large-area memristor crossbar arrays, slot die coating offers superior scalability and patterning capabilities [33] [7]. The process begins with the preparation of PQD ink with carefully optimized viscosity (typically 1-100 cP) and surface tension (25-40 mN/m) to ensure stable meniscus formation. The substrate (flexible or rigid) is mounted on a vacuum chuck and precisely aligned with the slot die head, maintaining a gap distance of 50-500 µm. The PQD ink is pumped through the internal manifold of the slot die head at controlled flow rates (0.1-5 mL/min) while the substrate translates at speeds of 0.1-10 m/min. The coating process is performed in a controlled environment with relative humidity below 30% to prevent premature crystallization. Immediately after deposition, the wet film undergoes in-line drying through infrared or convective heating zones (50-100°C) to remove solvents, followed by a final thermal annealing step (90-110°C) to complete PQD crystallization. This technique enables the fabrication of large-area PQD films with minimal material wastage and has been successfully employed for producing memristor crossbar arrays with 4×4 configurations and beyond [7].

Memristive Characterization Methods

The electrical characterization of PQD memristors focuses on evaluating key performance metrics including resistive switching behavior, endurance, retention, and synaptic plasticity. Current-voltage (I-V) characteristics are measured using a semiconductor parameter analyzer (e.g., Keithley 4200-SCS) with voltage sweeps typically applied in a bipolar mode ranging from -3 V to +3 V [7]. The voltage is applied to the top electrode while the bottom electrode remains grounded, using tungsten probes for electrical contact. For endurance testing, repetitive switching cycles are performed using voltage pulses (100 ns to 1 µs duration) with appropriate set and reset amplitudes, while resistance states are read at lower voltages (e.g., 0.1 V) to minimize disturbance [5]. Retention properties are evaluated by monitoring resistance state stability over time (up to 10⁴ seconds) at various temperatures (25-150°C), with data extrapolated using Arrhenius models to predict long-term behavior [5]. For synaptic applications, pulsed measurements are performed to emulate potentiation and depression characteristics, using trains of identical pulses (100 ns width) with gradually increasing or decreasing amplitudes to demonstrate analog conductance modulation [5].

Diagram 1: Experimental workflow for PQD memristor fabrication and characterization.

The Scientist's Toolkit: Essential Research Reagents and Materials

The fabrication of PQD composite memristors requires carefully selected materials and reagents to achieve optimal device performance. The table below details essential research reagents, their specific functions, and considerations for selection based on the target memristive characteristics.

Table 3: Essential Research Reagents for PQD Memristor Fabrication

Reagent/Material	Function in PQD Memristor	Specific Examples	Considerations for Memristive Applications
Perovskite Precursors	Forms the active switching layer	PbI₂, CH₃NH₃I, CsBr, FAI	Purity (>99.9%) affects defect density and switching uniformity [7]
Solvents	Dissolves precursors to form coating solution	DMF, DMSO, GBL, ACN	Boiling point and viscosity influence film morphology during deposition [33]
Antisolvents	Controls crystallization kinetics	Chlorobenzene, Toluene, EA	Timing of drip critical for quantum dot formation and size distribution [33]
Substrate Materials	Supports the PQD active layer	ITO/glass, FTO, Si/SiO₂	Surface energy and roughness affect PQD film uniformity and interface quality [7]
Electrode Materials	Forms electrical contacts	Au, Ag, ITO, Ta, Pt [7] [5]	Work function influences barrier height and switching polarity [5]
Polymers for Composites	Enhchanical stability and processability	PMMA, PS, PVP [36]	Improves switching endurance by stabilizing the PQD structure [36]

Diagram 2: Relationship between PQD film properties and memristive characteristics.

Performance Comparison and Data Analysis

Quantitative Comparison of Memristive Performance Based on Deposition Techniques

The selection of deposition technique significantly impacts the resulting memristive performance of PQD devices. Research on various metal oxide memristors provides valuable insights that can be extrapolated to PQD-based systems. For instance, zinc oxide memristors fabricated via magnetron sputtering demonstrate stable resistive switching for over 1000 cycles between high resistance states (HRS = 537.4 ± 26.7 Ω) and low resistance states (LRS = 291.4 ± 38.5 Ω), with a resistance ratio (HRS/LRS) of approximately 1.8 [7]. Similarly, Ta/HfO₂/Pt devices show exceptional endurance up to 120 billion cycles and the capability to be programmed to 24 discrete resistance levels through compliance current control [5].

Table 4: Comparative Memristive Performance Based on Deposition Approach

Deposition Technique	Endurance (Cycles)	Resistance Ratio (HRS/LRS)	Retention (Extrapolated)	Multilevel Capability
Spin Coating	10³-10⁶ [7]	~1.8-10 [7]	>10 years at 85°C [5]	8-16 levels demonstrated [5]
Slot Die Coating	>1000 demonstrated [7]	~2.5 for crossbar arrays [7]	>20,000 cycles for crossbars [7]	4×4 crossbar arrays [7]
Sputtering (Reference)	1.2×10¹¹ [5]	~1.8-2.5 [7] [5]	7×10⁴ years at 85°C [5]	24 discrete levels [5]

The data indicates that while vacuum-deposited devices currently demonstrate superior endurance and multilevel capabilities, solution-processed PQD memristors offer competitive performance for many applications, with the added advantages of low-temperature processing and compatibility with flexible substrates. The resistance window (HRS/LRS ratio) remains a critical parameter for reliable memory operation, with ratios greater than 2 generally desirable for distinguishing between states with minimal read errors [7].

Correlation Between Film Morphology and Switching Characteristics

Direct relationships have been established between film morphological characteristics and key memristive switching parameters. For instance, films with smoother surfaces and uniform nanocrystalline structure exhibit more stable resistive switching behavior with reduced cycle-to-cycle variability [7]. In transparent ITO/ZnO/ITO memristor structures, the magnetron sputtering power during deposition significantly influences both the film morphology and the resulting resistive switching characteristics, with optimal power levels (e.g., 75 W) yielding superior performance [7].

Advanced characterization techniques including scanning transmission electron microscopy (STEM) and electron energy loss spectroscopy (EELS) have enabled direct observation of conduction channels in memristive devices [5]. These studies reveal that compositionally modulated regions as small as sub-10 nm are responsible for the resistance switching behavior, highlighting the critical importance of nanoscale morphological control during PQD deposition [5]. The ability of solution-processing techniques to precisely control quantum dot size, distribution, and interfacial quality makes them particularly promising for engineering reproducible switching phenomena in PQD memristors.

Solution-processing techniques for PQD composite deposition offer diverse pathways for controlling memristive characteristics through morphological engineering. The comparative analysis presented in this guide demonstrates that each deposition method presents distinct advantages: spin coating provides exceptional uniformity for research-scale device optimization; slot die coating enables large-area patterning suitable for crossbar array fabrication; and spray coating offers versatility for complex substrate geometries. The selection of an appropriate deposition strategy must consider the target memristive application, with specific techniques optimized for metrics including endurance, retention, resistance window, and multilevel capability.

Future developments in PQD memristor technology will likely focus on hybrid approaches that combine the strengths of multiple deposition techniques, such as using spin coating for interfacial layers while employing slot die coating for the active PQD layer [34]. Additionally, advances in ink formulation and post-deposition treatments will further enhance the performance of solution-processed devices to compete with vacuum-deposited counterparts. The relationship between deposition parameters, resulting morphology, and memristive characteristics established in this guide provides a framework for researchers to strategically engineer PQD films with tailored switching behaviors for specific neuromorphic computing and non-volatile memory applications.

In the pursuit of advanced neuromorphic computing and non-volatile memory technologies, memristive devices based on quantum dot (QD)-polymer nanocomposites have emerged as a leading solution [37] [9]. The performance of these devices is not merely a function of the constituent materials but is profoundly governed by the nanoscale morphology of the composite active layer. Specifically, the quantum dot concentration and the uniformity of their dispersion within polymer hosts such as polyvinylpyrrolidone (PVP) and polymethyl methacrylate (PMMA) are critical parameters that dictate charge trapping dynamics and resistive switching behavior [38] [39]. This guide provides a comparative analysis of how these factors influence memristive characteristics, underpinned by experimental data and detailed methodologies, to inform material selection and fabrication strategies for researchers in the field.

The Critical Role of Quantum Dot Concentration

Switching Behavior Transitions

The concentration of quantum dots within the polymer matrix directly controls the nature of the resistive switching, determining whether a device exhibits digital (abrupt) or analog (gradual) switching characteristics. This transition is pivotal for tailoring devices for specific applications, from non-volatile memory to neuromorphic computing.

Table 1: Impact of QD Concentration on Memristive Switching Behavior

Polymer Host	QD Type	QD Concentration	Switching Type	Key Performance Metrics	Reference
PVP	N-doped Carbon QDs (NCQDs)	10 wt%	Digital RS (D-RS)	Abrupt HRS/LRS transition	[38]
PVP	N-doped Carbon QDs (NCQDs)	30 wt%	Digital RS (D-RS)	Lower set/forming voltage	[38]
PVP	N-doped Carbon QDs (NCQDs)	40 wt%	Analog RS (A-RS)	Gradual, continuous conductance modulation	[38]
PVP	WS₂ QDs	0.8 wt%	Non-volatile Bistable	ON/OFF ratio: ~10⁴, Set Voltage: +0.7 V	[39]

Underlying Mechanism and Experimental Evidence

The shift from digital to analog switching with increasing QD concentration, as documented in PVP/NCQD systems, can be attributed to the evolution of charge transport mechanisms [38]. At lower concentrations (e.g., 10-30 wt%), QDs act as isolated charge trapping sites. An applied electric field triggers an abrupt, collective charge de-trapping event, leading to a sharp jump in current (digital switching) [38]. At a critical concentration (e.g., 40 wt% for NCQDs), the trapping centers are sufficiently filled to promote the formation of multiple, percolative conductive paths. The conductance of these paths can be modulated gradually by the applied voltage, enabling the analog switching behavior that is essential for emulating synaptic plasticity [38]. Furthermore, optimal concentration reduces operational voltages, thereby lowering energy consumption, as seen in WS₂ QD-PVP devices with a low set voltage of +0.7 V [39].

The Imperative of Quantum Dot Dispersion

A uniform dispersion of QDs is a prerequisite for achieving reliable and high-yield device performance. Agglomeration or poor dispersion creates localized regions with high charge density and uneven current pathways, leading to performance variations and device failure.

Synthesis and Fabrication Protocols

Achieving a homogeneous QD-polymer nanocomposite requires precise control over the fabrication process. The following protocols, adapted from recent studies, have proven effective.

Detailed Experimental Protocol: Fabrication of QD-Polymer Nanocomposite Films

• Solution Processing of Nanocomposite: The foundational step involves integrating QDs into the polymer solution. For a WS₂ QD-PVP nanocomposite, synthesized WS₂ QDs in a solvent like DMF are mixed with a PVP-DMF solution. This combined colloid is then magnetically stirred for an extended period (e.g., 10 hours) at room temperature to ensure a homogenous mixture [39].
• Film Deposition via Spin-Coating: The homogeneous solution is deposited onto a substrate (e.g., ITO-coated glass or PET) using a spin coater. A typical two-step process involves spinning at 300 rpm for 20 seconds followed by 3000 rpm for 40 seconds to achieve a uniform thin film [39].
• Annealing and Electrode Deposition: The spun-cast film is annealed on a hotplate (e.g., at 100°C for 1 hour) to evaporate the residual solvent. Finally, top electrodes (e.g., 200 nm thick Al) are thermally evaporated onto the film through a shadow mask to complete the metal/active-layer/metal device structure [39].

Characterization of Dispersion and Morphology

Researchers employ several analytical techniques to verify the quality of QD dispersion and the resulting nanocomposite morphology:

• Atomic Force Microscopy (AFM): Used to confirm the surface flatness of the spin-coated film. For instance, a PVP-NCQD film showed a root mean square roughness (Rq) of 0.7 nm, indicating a very smooth and uniform surface [38].
• Transmission Electron Microscopy (TEM): Provides direct visualization of QD distribution within the polymer matrix, confirming the absence of agglomerates [39].
• Scanning Electron Microscopy (SEM): Reveals the surface morphology of the composite film. A homogeneous distribution without phase separation is a key indicator of good dispersion [40].

Comparative Performance in Memristive Applications

The synergistic effect of optimal QD concentration and excellent dispersion directly translates to superior device performance, enabling both memory and neuromorphic computing applications.

Table 2: Comparative Performance of QD-Polymer Nanocomposite Memristive Devices

Device Structure	QD Dispersion Method	Key Memristive & Synaptic Functions	Performance Metrics	Reference
Al / PVP-NCQD (40%) / ITO	Magnetic stirring	Synaptic Plasticity: STP, LTP, STDP, PPF	Analog conductance modulation, Low-power operation	[38]
Al / PVP-WS₂ QDs / ITO	Magnetic stirring + Spin-coating	Non-volatile memory	ON/OFF: 10⁴, Retention: >10⁴ s, Set Voltage: 0.7 V	[39]
Perovskite QDs / Ferrocene-β-CD Supramolecule	Host-guest interaction	Phototransistor memory	Current stability: 10⁴ s (extrap. to 10⁹ s), Low dark current	[41]

The PVP-NCQD device with 40% doping successfully emulates fundamental synaptic functions, including Short-Term Plasticity (STP), Long-Term Potentiation (LTP), and Spike-Timing-Dependent Plasticity (STDP), which are the cornerstones of neuromorphic computing that mimics the human brain [38]. This is a direct result of the analog switching behavior enabled by the optimal QD concentration and dispersion.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Materials for Fabricating QD-Polymer Memristive Devices

Material Category	Specific Example(s)	Function in the Experiment	Key Property
Polymer Host	Polyvinylpyrrolidone (PVP), Polymethyl Methacrylate (PMMA)	Forms the matrix; provides flexibility, solution processability, and charge transport medium.	High transparency, good film-forming, tunable dielectric constant [38] [42].
Quantum Dots	N-doped Carbon QDs (NCQDs), WS₂ QDs, Perovskite QDs	Act as charge trapping centers; concentration controls resistive switching type.	Strong quantum confinement, tunable bandgap, functionalizable surface [38] [39] [9].
Solvent	N,N-Dimethylformamide (DMF), Toluene	Dissolves polymer and disperses QDs to form a homogeneous solution for film casting.	High boiling point, effectively disperses nanomaterials [39] [40].
Substrate	ITO/Glass, ITO/Polyethylene Naphthalate (PEN)	Provides mechanical support and bottom electrical contact.	ITO offers transparency; PEN enables flexible devices [38] [39].
Electrodes	Aluminum (Al), Silver (Ag), ITO	Form top and bottom electrical contacts to apply bias and measure current.	Thermally evaporatable, good electrical conductivity.

This comparison guide underscores that the meticulous control of QD concentration and dispersion within polymer hosts like PVP and PMMA is not a mere fabrication detail but a fundamental determinant of memristive device performance. A critical concentration of ~40 wt% of NCQDs in PVP can trigger a vital transition from digital to analog switching, enabling synaptic plasticity. Simultaneously, synthesis protocols involving prolonged stirring and sonication are essential for achieving a uniform dispersion, which ensures consistent and reliable device operation. As the field progresses, the exploration of novel QD materials (e.g., perovskites) and advanced dispersion techniques (e.g., supramolecular chemistry [41]) will further enhance the capabilities of these nanocomposites, solidifying their role in the future of neuromorphic computing and high-density non-volatile memory.

The performance of memristive devices, which are pivotal for neuromorphic computing and next-generation data storage, is fundamentally governed by the charge trapping and de-trapping dynamics within the active layer. The strategic selection of quantum dot (QD) and polymer host materials, particularly their bandgap alignment, is critical for optimizing these dynamics. This guide provides a comparative analysis of different QD-polymer composite systems, drawing on recent experimental data to outline the material selection principles that ensure superior memristive characteristics, such as controllable resistive switching, data retention, and synaptic plasticity.

The synergy between QDs and polymers creates a composite where QDs act as primary charge trapping sites, while the polymer matrix facilitates charge transport and provides structural integrity. An optimal energy band alignment between the two components can significantly enhance charge confinement and release efficiency, leading to improved device performance and reliability. This document objectively compares the performance of several documented QD-polymer systems to guide researchers in selecting the ideal material pair for specific memristive applications.

Comparative Analysis of QD-Polymer Composite Systems

The following table summarizes key performance metrics and characteristics of three distinct QD-polymer composites reported in recent literature, providing a basis for direct comparison.

Table 1: Performance Comparison of Different QD-Polymer Composites for Memristive Applications

System Composition	QD Bandgap (eV)	Polymer Host	Resistive Switching Type	ON/OFF Ratio	Retention/Endurance	Key Memristive Characteristics
N-doped Carbon QDs (NCQDs) [38]	Information Missing	Polyvinylpyrrolidone (PVP)	Analog (A-RS)	Information Missing	Information Missing	Short-term to long-term plasticity transition, synaptic functions (EPSC, PPF, STDP) [38]
Cobalt-doped ZnS QDs [43]	~3.65 (pristine ZnS)	Polyvinyl Alcohol (PVA)	Digital (D-RS) & WORM	~10	2000 s / 2000 cycles	Write-Once-Read-Many (WORM) memory behavior; enhanced switching with higher Co doping [43]
Er-PANI:PSS Composite [44]	Information Missing	Polyaniline:Polystyrene-sulfonate (PANI:PSS)	Photoelectrochemical	Information Missing	Information Missing	Light-enhanced current; synergistic charge storage; promising for solar energy storage, not classic memristors [44]

Detailed System Breakdown and Experimental Protocols

NCQDs in PVP Matrix for Synaptic Emulation

This system demonstrates the capability for analog resistive switching, which is essential for emulating biological synapses.

• * Synthesis and Fabrication*: NCQDs were synthesized via a hydrothermal method using citric acid and urea as carbon and nitrogen sources, resulting in uniformly sized QDs of 2-5 nm [38]. The memristive synapse was fabricated by spin-coating a PVP solution embedded with 40 wt% NCQDs onto an ITO-coated substrate, followed by the thermal evaporation of Al top electrodes [38].
• Key Findings and Mechanism: The polar groups and defects on the NCQD surface serve as effective charge traps. At an optimal doping concentration of 40 wt%, the device exhibits gradual, analog-type resistive switching. This is attributed to the formation and modulation of multiple conductive paths via the trapping and de-trapping of charges at the NCQD sites [38]. The device successfully emulated fundamental synaptic functions such as Excitatory Postsynaptic Current (EPSC), Paired-Pulse Facilitation (PPF), and the transition from Short-Term Plasticity (STP) to Long-Term Plasticity (LTP) [38].

Cobalt-doped ZnS QDs in PVA for Non-Volatile Memory

This system is tailored for non-volatile memory applications, exhibiting digital switching and WORM memory behavior.

• Synthesis and Fabrication: Cobalt-doped ZnS QDs with concentrations of 10% and 15% were synthesized via a cost-effective sol-gel process [43]. The resulting QDs had a cubic zinc blende structure with crystallite sizes between 4.14 nm and 5.93 nm. The active layer for the memory device was formed by creating a composite solution of the QDs with PVA polymer and depositing it via spin coating onto an ITO substrate, followed by the deposition of Ag top electrodes [43].
• Key Findings and Mechanism: The incorporation of Co-ZnS QDs into the PVA matrix induced a pronounced resistive switching behavior. The device with 15% Co doping displayed a WORM memory effect, with an ON/OFF ratio of one order of magnitude and excellent retention for 2000 seconds [43]. The conduction mechanism was identified as Space-Charge-Limited Current (SCLC) and Poole-Frenkel emission, where the space charge region formed at the electrode-active material interface is responsible for the memory behavior [43].

Erbium-Polyaniline Composite for Photoelectrochemical Applications

While not a classic memristor, this system highlights the role of bandgap engineering and charge trapping for a related energy storage application.

• Synthesis and Fabrication: Polyaniline Quantum Dots (PANI QDs) of ~8 nm were first synthesized via an acid-assisted polymerization, stabilized with PSS [44]. The Er³⁺-PANI:PSS composite was then formed by adding a solution of Er-HCl to the pre-formed PANI:PSS suspension, localizing erbium ions at the polymer surface [44].
• Key Findings and Mechanism: UV-Vis spectroscopy indicated a blue shift in the absorption band of the composite, suggesting interaction between Er³⁺ ions and the PANI:PSS system and charge polaron localization [44]. Cyclic Voltammetry (CV) and Linear Sweep Voltammetry (LSV) measurements showed a marked improvement in electrochemical performance under light irradiation, attributed to a synergistic mechanism involving photo-assisted PANI protonation and Er³⁺-mediated charge storage [44].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Key Reagents and Materials for QD-Polymer Composite Fabrication

Material/Reagent	Function in Device Fabrication	Example System
Citric Acid & Urea	Carbon and nitrogen precursors for synthesizing N-doped Carbon QDs (NCQDs) [38].	NCQD-PVP Memristive Synapse [38]
Polyvinylpyrrolidone (PVP)	Polymer host matrix; provides film-forming capability and contains QDs [38].	NCQD-PVP Memristive Synapse [38]
Zinc Chloride & Cobalt Chloride	Metal precursors for synthesizing pristine and Cobalt-doped ZnS quantum dots [43].	Co-ZnS QD-PVA Memory Device [43]
Polyvinyl Alcohol (PVA)	Polymer host for forming composite thin films with QDs; offers flexibility and transparency [43].	Co-ZnS QD-PVA Memory Device [43]
Aniline & Formic Acid	Monomer and acid medium for the synthesis of Polyaniline (PANI) quantum dots [44].	Er-PANI:PSS Composite [44]
Polystyrene-sulfonate (PSS)	Stabilizing agent for PANI QDs, preventing aggregation [44].	Er-PANI:PSS Composite [44]

Signaling Pathways and Workflow Visualization

The following diagram illustrates the universal charge trapping and de-trapping mechanism in a QD-polymer composite memristive device, which underpins the resistive switching behavior.

Charge Trapping/De-Trapping Mechanism in QD-Polymer Memristors.

The experimental workflow for fabricating and characterizing these devices is methodical, as shown below.

Experimental Workflow for QD-Polymer Device.

Memristors have emerged as a cornerstone for next-generation non-volatile memory and neuromorphic computing, offering a promising solution to the von Neumann bottleneck through in-memory computing architectures [45]. The performance and application of these devices are profoundly influenced by their physical architecture, which governs characteristics such as integration density, power consumption, and reliability. This guide provides a comprehensive comparison of three predominant memristor device architectures: planar metal-insulator-metal (MIM), crossbar arrays, and emerging three-dimensional (3D) structures. The comparative analysis is framed within research on memristive characteristics influenced by different perovskite quantum dot (PQD) surface morphologies, providing researchers with critical insights for selecting appropriate architectures based on specific application requirements in memory storage, neuromorphic computing, and biomedical sensing.

Planar Metal-Insulator-Metal (MIM) Structures

The planar MIM structure represents the fundamental building block of memristor technology, consisting of a simple sandwich-like configuration where an insulating layer is positioned between two metal electrodes [46]. Resistive switching in this architecture occurs through the formation and rupture of conductive filaments within the insulating layer [46]. This structure typically demonstrates high endurance cycles exceeding (10^{12}), extended data retention over 10 years, and fast switching speeds below 10 ns [46]. The fabrication of MIM structures often employs physical vapor deposition methods, including direct current (DC) and radio frequency (RF) magnetron sputtering, which provide precise control over growing film parameters and high process stability [7]. For transparent memristor applications, materials such as indium tin oxide (ITO) for electrodes and zinc oxide (ZnO) for the insulating layer are deposited via magnetron sputtering to create transparent ITO/ZnO/ITO configurations [7].

Crossbar Array Architectures

Crossbar array architectures significantly enhance integration density by arranging memristor cells at the intersections of perpendicular word lines (WLs) and bit lines (BLs) [46]. This configuration achieves minimal cell sizes of approximately (4F^2), where F represents the minimum feature size, potentially reaching storage densities exceeding 100 Gb/cm² [46]. A critical challenge in crossbar arrays is the sneak path current phenomenon, where unwanted current flows through neighboring cells can corrupt stored data and lead to erroneous readouts [46]. Mitigation strategies often incorporate selector devices, such as diodes or transistors, integrated with each RRAM cell (1D1R structure) to control sneak currents and enhance array reliability [47]. Thermal crosstalk presents another significant challenge in dense memory arrays, where Joule heat generated during operation can deteriorate the retention performance of adjacent devices [47].

Emerging 3D Memristor Structures

Three-dimensional memristor architectures utilize vertical integration to dramatically increase storage capacity while reducing footprint [46]. These implementations are broadly categorized into horizontal stacked 3D RRAM (HRRAM) and vertical RRAM (VRRAM) schemes [46]. HRRAM architectures consist of multiple planar 2D RRAM layers superimposed vertically [46]. While innovative, HRRAM faces challenges including fabrication complexity, linearly increasing manufacturing costs with additional layers, and scalability constraints compared to VRRAM [46]. VRRAM extends the conventional horizontal crossbar array structure vertically, offering enhanced scalability potential [46]. Thermal management becomes increasingly critical in 3D architectures, as the time to reach thermal steady state can exceed 500 ns in a 3×3×3 block array—approximately ten times longer than for a single 1D1R element—which significantly impacts switching behavior and reliability [47].

Performance Comparison

The table below summarizes the key characteristics and performance metrics of the three memristor architectures:

Table 1: Comprehensive comparison of memristor device architectures

Architectural Feature	Planar MIM	Crossbar Array	3D Structures
Structural Layout	Single-layer metal-insulator-metal sandwich [46]	Word lines (WL) and bit lines (BL) with cells at intersections [46]	Vertical stacking of multiple memory layers [46]
Cell Size	Limited by lithography [46]	~(4F^2) [46]	Significantly reduced footprint through vertical integration [46]
Storage Density	Limited	>100 Gb/cm² [46]	Ultra-high (exceeds crossbar) [46]
Key Challenges	Lithographic limits, variability at small feature sizes [46]	Sneak path currents, thermal crosstalk [46] [47]	Fabrication complexity, thermal management, cost [46] [47]
Switching Speed	<10 ns [46]	Varies (affected by parasitic elements)	Varies (reset process dominated by transient thermal effects) [47]
Endurance	>(10^{12}) cycles [46]	Limited by sneak paths and selectors	Affected by thermal crosstalk between layers [47]
Retention	>10 years [46]	Deteriorated by thermal crosstalk [47]	Can lead to state failure (LRS to HRS) due to thermal issues [47]
Selector Requirement	Not essential	Essential (1D1R, 1T1R) [47]	Integrated in 3D (e.g., 1D1R) [47]
Thermal Behavior	Reaches steady state quickly (<5 ns for individual device) [47]	Moderate thermal crosstalk	Severe thermal crosstalk; steady state >500 ns for 3×3×3 array [47]

Experimental Protocols and Methodologies

Fabrication of Planar MIM Memristors

The experimental protocol for fabricating transparent ITO/ZnO/ITO memristor structures involves multiple stages of magnetron sputtering [7]. Initially, a 200 nm thick ITO bottom electrode is deposited on glass substrates using pulsed DC magnetron sputtering in an argon atmosphere at room temperature, with a sputtering power of 200 W and operating pressure of (2 \times 10^{-3}) mbar [7]. Subsequently, a 60 nm thick nanocrystalline ZnO film is deposited via RF magnetron sputtering of a ZnO ceramic target (99.99% purity) in argon atmosphere at room temperature, with optimal sputtering power of 75 W and working pressure of (5 \times 10^{-3}) mbar [7]. Finally, a 150 nm thick top ITO electrode is deposited using pulsed DC magnetron sputtering, completing the MIM structure [7]. Electrical characterization typically involves current-voltage (I-V) measurements using a semiconductor parameter analyzer, with the voltage applied to the top electrode while the bottom electrode is grounded [7].

Construction of Crossbar Arrays

For crossbar array fabrication, the process begins with the formation of bottom electrode lines on a substrate using photolithography and magnetron sputtering [7]. A switching material layer (e.g., ZnO, MoS₂, or perovskite) is then deposited uniformly across the substrate [7]. In some approaches, particularly for 2D materials like MoS₂, a transfer-free fabrication method may be employed where an atomically thin amorphous MoS₂ precursor is RF-sputtered directly onto pre-patterned graphene and crystallized by confined-space sulfurization at 800°C [48]. Photolithography and lift-off processes are subsequently used to pattern the top electrode lines perpendicular to the bottom electrodes, creating the cross-point structure [7]. For 1D1R arrays, diode selectors are integrated at each cross-point, often in a series configuration with the RRAM element, to mitigate sneak path currents [47].

3D Stacking Methodologies

The construction of 3D memristor architectures employs both horizontal and vertical stacking approaches. Horizontal stacked 3D RRAM (HRRAM) is created by sequentially fabricating multiple planar 2D RRAM layers separated by insulating layers [46]. Each memory layer must be individually patterned and interconnected through vias, which increases fabrication complexity [46]. Vertical RRAM (VRRAM) architectures implement memory cells along vertically oriented electrodes, significantly enhancing density [46]. These structures may incorporate selector-integrated vertical nanowires and hybrid photonic-memristive layers to enhance bandwidth and reduce energy consumption [49]. Thermal management considerations are critical during design and operation, as the increased power density in 3D structures can lead to significant thermal crosstalk, potentially causing resistance state degradation in adjacent cells [47].

Architectural Visualization

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Essential materials and reagents for memristor device fabrication

Material/Reagent	Function	Application Examples
ITO (Indium Tin Oxide)	Transparent conductive electrode material [7]	Electrodes in transparent memristors [7]
Zinc Oxide (ZnO)	Wide bandgap semiconductor switching layer [7]	Resistive switching medium in MIM structures [7]
Molybdenum Disulfide (MoS₂)	2D transition metal dichalcogenide switching layer [48]	Active layer in memristors; forms heterostructures with graphene [48]
Graphene	2D electrode material with high carrier mobility [48]	Electrodes in memristors; reduces contact resistance [48]
Halide Perovskites (CsPbI₃)	Ionic crystal for resistive switching [45] [49]	Volatile and non-volatile switching layers [45]
Hafnium Oxide (HfO₂)	High-k dielectric with ferroelectric properties [50]	Ferroelectric memory capacitors in MFM structures [50]
Elemental Sulfur	Sulfurization agent for crystallization [48]	Conversion of amorphous MoS₂ to crystalline 2H-MoS₂ [48]
PMMA (Polymethyl Methacrylate)	Polymer transfer medium [48]	Graphene transfer in 2D material-based device fabrication [48]

The selection of an appropriate memristor architecture involves critical trade-offs between integration density, performance, and fabrication complexity. Planar MIM structures provide a fundamental platform with excellent individual device characteristics but face scalability limitations. Crossbar arrays offer significantly enhanced density but require sophisticated solutions to address sneak path currents and thermal management. Emerging 3D architectures push density boundaries further while introducing additional challenges in thermal crosstalk and manufacturing complexity. For research focusing on PQD surface morphologies, planar MIM structures often serve as ideal testbeds for fundamental material characterization, while crossbar and 3D architectures become increasingly relevant for translating material properties into functional array-level demonstrations. The ongoing development of these architectures, coupled with advancements in 2D materials, lead-free perovskites, and ferroelectric HfO₂, continues to expand the application horizons for memristive technologies in neuromorphic computing, artificial intelligence, and advanced sensory systems.

The pursuit of neuromorphic computing, which aims to emulate the brain's architecture and efficiency, has catalyzed significant research into novel hardware synapses. Memristive devices, with their inherent ability to mimic synaptic plasticity, have emerged as leading candidates for this role. These devices are central to the development of artificial synapses for brain-inspired computing, flexible electronics for wearable applications, and in-memory computing architectures that circumvent the limitations of traditional von Neumann systems. This guide provides a comparative analysis of the performance and experimental methodologies underpinning several key material platforms in this field, including organic-inorganic hybrids, low-dimensional silicon, and single nanowire systems.

The operational principle of these devices often hinges on the dynamic control of conductive pathways. In biological synapses, the strength of a connection, or weight, is modulated by calcium (Ca²⁺) dynamics triggered by neuronal action potentials. In an analogous manner, the conductance state (weight) of an artificial memristive synapse is modulated by the migration of ions (e.g., Ag⁺) or the trapping/detrapping of charge in response to external electrical stimuli [51] [38]. This bio-realistic foundation enables the emulation of critical synaptic learning rules, making these devices capable of supporting complex, energy-efficient neural networks.

Comparative Performance Analysis of Memristive Technologies

Table 1: Performance Comparison of Artificial Synapse Technologies

Technology Platform	Synaptic Plasticity Demonstrated	Key Metrics (Range/Value)	Flexibility & Transparency	Switching Type & Endurance
PVP-NCQD Synapse [38]	EPSC, PPF, STDP, STP-to-LTP transition	Conductance strengthening/weakening over 100 pulses; Transmittance: 90-95% (visible spectrum)	Excellent flexibility on PET substrate; Transparent	Analog-type RS (A-RS); Stable over 50+ consecutive cycles
ZnO Nanowire Synapse [51]	STP, PPF, imitation of Ca²⁺ dynamics	Forming voltage: <5 V (tunable); Low operation voltages	Not explicitly reported; Single-crystal structure	Bipolar RS, Multilevel switching, Selector operation
Low-Dimensional Si [52]	Synaptic plasticity, Memory, Sensing	Performance dictated by bulk Si properties	Not a primary focus; Inherently rigid	Comparable to devices based on oxides, 2D materials, ferroelectrics
Paper-Based Electronics [53]	(Platform for device integration)	Resistance change rate: <2% after 500 folding cycles; Response time: 500 ms (humidity sensor)	Excellent flexibility, foldability, and lightweight	N/A (Substrate platform)

Table 2: Material Composition and Device Architecture

Technology Platform	Active Material / Structure	Electrode Configuration	Fabrication Method	Reported Neuromorphic Functions
PVP-NCQD Synapse [38]	PVP polymer matrix with N-doped Carbon Quantum Dots	Al (Top) / ITO (Bottom) on PET	Solution processing; Crossbar arrays	Potentiation/Depression, STDP, STP-to-LTP transition
ZnO Nanowire Synapse [51]	Single-crystalline ZnO Nanowire	Ag (active) / Pt (inert) asymmetric electrodes	Bottom-up Chemical Vapour Deposition (CVD)	Non-volatile memory, selector, synaptic operations
Low-Dimensional Si [52]	Si Nanowires, Quantum Dots, Nanocrystals, Nanosheets	Configuration not specified in excerpt	Leverages advanced silicon industry processes	Synaptic behaviors, memory, sensing functionalities
Paper-Based Electronics [53]	Cellulose fiber substrate with conductive layers (e.g., CNTs, metals)	Various (e.g., metal, carbon-based)	Printing, deposition, surface self-assembly	Serves as a flexible, biodegradable substrate for sensors and devices

Experimental Protocols for Key Methodologies

Fabrication and Characterization of PVP-NCQD Memristive Synapses

Device Fabrication:

• NCQD Synthesis: NCQDs are synthesized via a hydrothermal treatment process using citric acid and urea as carbon and nitrogen sources, resulting in particles with a uniform size of 2-5 nm [38].
• Nanocomposite Preparation: The NCQDs are dispersed into a polyvinylpyrrolidone (PVP) polymer matrix. An optimal doping concentration of 40 wt% is critical for achieving analog-type resistive switching (A-RS) [38].
• Device Stack Formation: The PVP-NCQD nanocomposite is deposited onto a flexible and transparent indium tin oxide (ITO)-coated poly(ethylene terephthalate) (PET) substrate. The top Al electrodes are then patterned to form a crossbar array structure, completing the Al/PVP-NCQD/ITO sandwich device [38].

Electrical Measurement Protocol:

• DC Sweep Testing: Current-voltage (I-V) characteristics are measured by applying a bias voltage to the Al top electrode with the ITO bottom electrode grounded. Voltage sweeps (e.g., 0 to +2 V for set, 0 to -2 V for reset) are used to characterize the resistive switching behavior [38].
• Pulse Testing: To emulate synaptic functions, trains of consecutive positive and negative voltage pulses are applied. The device conductance is read at a low voltage (e.g., 0.2 V) after each pulse to track the potentiation (strengthening) and depression (weakening) of the synaptic weight [38].
• Synaptic Function Emulation: Key synaptic plasticity rules are tested using specific pulse protocols:
  • Paired-Pulse Facilitation (PPF): Two closely spaced presynaptic pulses are applied, and the facilitation of the postsynaptic current response to the second pulse is measured.
  • Spike-Timing-Dependent Plasticity (STDP): The timing difference (Δt) between pre- and postsynaptic spikes is varied, and the corresponding change in synaptic weight is quantified [38].

Probing Synaptic Plasticity In Vivo with All-Optical Interrogation

This protocol details a method for measuring synaptic plasticity in identified neurons within awake, behaving animals, providing a biological benchmark for artificial synapse performance.

Experimental Workflow:

• Viral Expression: Co-inject adeno-associated viruses (AAVs) into the target brain region (e.g., cerebellar vermis). One virus expresses an optimized genetically encoded voltage indicator, JEDI-2Psub, in postsynaptic neurons (e.g., Purkinje cells). A second, Cre-dependent virus expresses the red-shifted opsin ChRmine in presynaptic neurons (e.g., granule cells) of transgenic driver mice [54].
• Animal Preparation: After 3-4 weeks for protein expression, prepare an awake, head-fixed mouse for in vivo two-photon microscopy. Sensory stimulation (e.g., air puffs to the whisker pad) can be used to activate specific native neural pathways (e.g., climbing fibers) [54].
• All-Optical Synaptic Probing:
  • Stimulation: Presynaptic granule cells are selectively activated using pulses of red light to stimulate ChRmine.
  • Recording: Simultaneously, postsynaptic voltage responses in Purkinje cell dendrites are recorded at high speed (440 Hz) via two-photon imaging of JEDI-2Psub signals [54].
• Plasticity Induction and Measurement: To induce plasticity, pair the optogenetic activation of presynaptic granule cells with sensory-evoked climbing fiber inputs. Monitor the long-term changes in the amplitude of the optically recorded inhibitory postsynaptic potentials (IPSPs) in the Purkinje cells over time [54].

Quantifying Synapse Engulfment by Glial Cells

This cell biological protocol provides a high-confidence method for visualizing and quantifying synaptic pruning, a critical biological process for neural circuit refinement, offering context for the functional environment of biological synapses.

Experimental Workflow:

• Viral Labeling: Inject an adeno-associated virus (AAV) encoding pSynDig—a presynaptic, pH-sensitive fluorescent sensor—into a specific brain region (e.g., dorsal lateral geniculate nucleus, dLGN). pSynDig is expressed in neurons and labels their presynaptic terminals in target areas (e.g., primary visual cortex, V1) [55].
• Tissue Preparation: After ~3 weeks, lightly perfusion-fix the brain and harvest the tissue. Perform immunofluorescence staining with antibodies against specific glial markers, such as NG2 for oligodendrocyte precursor cells (OPCs) [55].
• Confocal Imaging: Acquire high-resolution three-dimensional Z-stack images of the stained tissue using a standard confocal microscope at Nyquist settings to ensure optimal sampling [55].
• Image Analysis:
  • 3D Reconstruction: Use software (e.g., Imaris) to reconstruct three-dimensional surfaces of the OPCs (based on NG2 signal) and all pSynDig+ synaptic inputs.
  • Engulfment Filtering: Apply a distance-based filtering algorithm to identify and retain only the synaptic surfaces that reside entirely within the OPC volume by a set distance. These are classified as engulfed synapses.
  • Quantification: Calculate an "engulfment score" as the volume of synaptic signal within an OPC normalized to the OPC's own volume. Alternatively, measure the eGFP-to-mCherry signal ratio within the OPC to assess the degradation status of the engulfed synapses, as the eGFP signal is quenched in acidic phagolysosomes [55].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Research Reagent Solutions for Memristive and Neuromorphic Research

Reagent/Material	Function/Application	Example/Description
Nitrogen-doped Carbon Quantum Dots (NCQDs)	Charge trapping centers in organic memristive synapses; enable analog resistive switching and synaptic plasticity [38].	Hydrothermally synthesized from citric acid and urea; size range 2-5 nm; contain polar groups for charge trapping.
Polyvinylpyrrolidone (PVP)	Polymer matrix for flexible memristive devices; serves as the host material for charge trapping nanodots [38].	Used as the active switching layer in composite films; provides a flexible, solution-processable platform.
Genetically Encoded Voltage Indicator (GEVI)	Optical reporting of neuronal membrane potential dynamics in vivo [54].	JEDI-2Psub: An optimized GEVI with high sensitivity at resting membrane potentials for recording subthreshold synaptic potentials.
Red-Shifted Opsin (ChRmine)	Optogenetic activation of specific presynaptic neuron populations in vivo with minimal spectral interference from GEVIs [54].	Allows selective stimulation of presynaptic inputs (e.g., granule cells) while simultaneously imaging postsynaptic voltage responses.
pH-Sensitive Synaptic Sensor (pSynDig)	High-confidence quantification of synaptic engulfment by glial cells; differentially labels intact vs. degraded synapses [55].	AAV-encoded sensor expressing Synaptophysin fused to mCherry and pH-sensitive eGFP; eGFP quenches in acidic phagolysosomes.
Zinc Oxide Nanowires (ZnO NWs)	Single-crystalline platform for elucidating memristive mechanisms and implementing synaptic functions [51].	Grown via Chemical Vapour Deposition (CVD); serves as a solid electrolyte for Ag+ ion migration in atomic-scale memristors.

Signaling Pathways and Workflow Visualizations

Memristive Synaptic Plasticity Mechanism

The following diagram illustrates the fundamental analogy between a biological synapse and a PVP-NCQD-based memristive artificial synapse, highlighting the charge trapping mechanism.

Experimental Workflow for Flexible Synapse Characterization

This workflow outlines the key steps for fabricating and characterizing a flexible PVP-NCQD memristive synaptic device.

Performance Challenges and Optimization Strategies for Reliable PQD Devices

Addressing Cycle-to-Cycle and Device-to-Device Variability

In the pursuit of advanced neuromorphic computing and non-volatile memory technologies, memristive devices have emerged as a leading hardware candidate due to their functional resemblance to biological synapses and potential for high-density integration. However, their commercial deployment and research reproducibility are significantly hampered by inherent variability in device performance. This variability manifests primarily as cycle-to-cycle (C2C) fluctuations, where the switching parameters of a single device change between consecutive operation cycles, and device-to-device (D2D) variations, where performance differs across devices from the same fabrication batch [56] [57]. This guide objectively compares the memristive characteristics, with a focus on variability, across different material systems and device structures, particularly exploring the impact of surface and interface morphology. Understanding and mitigating these variabilities is crucial for the reliable design of memristor-based artificial neural networks and resistive random-access memory (RRAM) cells [56] [58].

Comparative Analysis of Memristive Systems and Their Variability

The performance and variability of memristive devices are profoundly influenced by the choice of active material, electrode configuration, and resulting surface and interface morphologies. The table below provides a structured comparison of different memristive systems based on recent experimental studies.

Table 1: Comparison of Memristive Device Characteristics and Variability

Device Structure / Material System	Switching Type	Key Metrics (Variability)	Synaptic Functions Demonstrated	Reported Causes of Variability
ALD-grown HfO₂/Ta₂O₅ Bilayer [57]	Bipolar Resistive Switching	C2C Variability (CV): VSET: 1.76%, VRESET: 2.14%D2D Variability (CV): VSET: 6.09%, VRESET: 3.22%	Potentiation & Depression, PPF, PPD	Process-induced interface anomalies, intrinsic stochasticity of filament formation.
Ag/ZnO Nanowire (NW) [51]	Electrochemical Metallization	Low operating voltages (<5 V), high reliability.	STP, PPF, mimicking of Ca²⁺ dynamics	Surface-confined ionic migration; active filament location can change.
PVP/N-doped Carbon Quantum Dot (NCQD) [38]	Analog-type RS (A-RS) at 40 wt% NCQD	Conductance modulation with 100 consecutive pulses.	EPSC, PPF, STDP, STP-to-LTP transition	Charge trapping/detrapping in NCQDs; doping concentration critically affects RS behavior.
Al-doped HfO₂-based RRAM [56]	Bipolar Resistive Switching	Statistically emulated C2C and D2D I-V characteristics.	--	Fluctuations in the conductive filament (CF) gap distance and initial state.
General Memristive Devices [59]	Filamentary Switching	--	--	Coexistence of multiple subfilaments; change in the active, current-carrying filament from cycle to cycle.

Experimental Protocols for Characterizing and Modeling Variability

A critical step in addressing variability is its accurate characterization and modeling, enabling the design of robust circuits. The following methodology details a statistical approach for integrating C2C and D2D variability into circuit simulations.

Statistical Parameter Extraction for Compact Models

This protocol is based on work using the Stanford model for memristive devices with an Al-doped HfO₂ switching layer [56].

• Device Fabrication: Memristive devices are fabricated with a structure of TiN/Ti/Al:HfO₂(6 nm)/TiN. The Ti layer acts as an oxygen scavenging layer to enable resistive switching.
• Electrical Characterization: DC voltage sweeps are performed. Set operations (transition to Low Resistance State, LRS) use sweeps from 0 to 1 V, and reset operations (transition to High Resistance State, HRS) use sweeps from 0 to -1 V. This is repeated for multiple cycles and across multiple devices to gather C2C and D2D data.
• Model Parameter Identification: The Stanford compact model is used, which describes resistive switching as the modulation of a conductive filament gap, g. Key model parameters influencing variability include gapini (initial gap), Ron (ON-state resistance), Roff (OFF-state resistance), and Vinit (initial voltage for switching) [56].
• Statistical Analysis and Fitting:
  • Distribution Fitting: The distributions of the identified key model parameters (e.g., Ron, Roff) from the measured data are fitted with Gaussian distributions, characterized by their mean (µ) and standard deviation (σ).
  • Monte Carlo Simulation: The extracted Gaussian parameters are used in a Monte Carlo (MC) circuit simulation. In each iteration, the MC algorithm selects a random value for each model parameter based on its distribution.
  • Validation: The standard deviation and average of the simulated I-V curves are compared to the experimental data to validate the model's accuracy in reproducing C2C and D2D dispersions [56].

Table 2: Key Research Reagent Solutions for Memristor Variability Studies

Material / Tool	Function in Research
Atomic Layer Deposition (ALD)	Enables wafer-scale, uniform deposition of high-k oxide layers (e.g., HfO₂, Ta₂O₅) for reduced D2D variability [57].
Stanford Memristor Model	A compact model for circuit simulation; its parameters can be statistically varied to emulate C2C and D2D variability [56].
Monte Carlo Simulation	A computational algorithm used in conjunction with compact models to statistically analyze the impact of parameter variations on circuit performance [56].
Spectromicroscopic Photoemission / XAS	Used to experimentally investigate the microscopic origin of variability, such as visualizing subfilamentary networks [59].
Trioctylphosphine (TOP) / L-Phenylalanine	Surface ligand modifiers for Perovskite Quantum Dots (PQDs) that passivate surface defects, improving stability and optical properties [60].

Visualization of Variability Analysis Workflow

The diagram below illustrates the integrated experimental and computational workflow for analyzing and modeling variability in memristive devices.

Mechanisms Underlying Variability and Surface Morphology Effects

The intrinsic stochasticity observed in memristive devices can be traced to nanoscale physical and electrochemical processes. A prominent mechanism in filamentary-type switches involves the formation and rupture of conductive filaments (CFs). Research has shown that multiple, competing subfilamentary networks can coexist within the switching layer. The activation and deactivation of these subfilaments, or even a complete change of the active filament path from one cycle to the next, is a primary cause of C2C variability [59]. This effect is influenced by local variations in material composition, defect density, and surface morphology.

The critical role of surface and interface engineering is evident across material systems. In single-crystal ZnO nanowire-based memristors, the entire resistive switching and synaptic plasticity are governed by the migration of Ag⁺ ions along the NW surface, rather than through the bulk. This surface-confined mechanism directly links the nanoscale surface morphology to the device's operational characteristics and reliability [51]. Similarly, in high-k oxide bilayers like HfO₂/Ta₂O₅, utilizing a single ALD system at a constant temperature minimizes interface anomalies and surface irregularities, leading to significantly improved statistical performance and lower C2C/D2D variability compared to processes using different deposition tools or temperatures [57].

Beyond inorganic systems, surface morphology and molecular engineering are equally vital in organic-inorganic hybrids. In memristive synapses based on PVP and N-doped Carbon Quantum Dots (NCQDs), the concentration of NCQDs acts as charge trapping centers. Precise control over this doping concentration is essential to transition the device from digital-type switching to the analog-type switching required for synaptic plasticity, thereby mitigating variability and enabling stable conductance modulation [38].

Addressing C2C and D2D variability is a multi-faceted challenge central to the advancement of memristive technology. As this comparison guide illustrates, the degree of variability is strongly dependent on the material system and fabrication processes. Key strategies for mitigation include the adoption of precise, uniform deposition techniques like ALD for oxide layers, surface and interface engineering to control ionic migration and defect distribution, and the use of advanced statistical models to predict and accommodate inherent stochasticity. Future research must continue to correlate macroscopic device performance with nanoscale surface and morphological characteristics to pave the way for more reliable and predictable memristive devices in neuromorphic computing and beyond.

Memristive devices, capable of retaining a history of their external stimuli, have emerged as a cornerstone for next-generation non-volatile memory and neuromorphic computing. Their operation hinges on resistive switching, a phenomenon where an external voltage modulates the device's resistance state. Despite their promising potential, large parameter variability and limited cycling endurance pose significant roadblocks to their large-scale commercialization [61]. These non-idealities are fundamentally rooted in uncontrolled ion migration during the switching process, which leads to device degradation. This degradation manifests as shifts in conduction mechanisms and the creation of electronic trap states, ultimately affecting the stability and reliability of the devices [61]. Understanding and mitigating these degradation phenomena is, therefore, critical for the advancement of memristive technology. This guide provides a comparative analysis of the degradation mechanisms in different memristive systems and outlines the experimental protocols used to characterize them, offering researchers a framework for evaluating device performance and stability.

Comparative Analysis of Memristive Degradation Mechanisms

The degradation of memristive devices is intrinsically linked to their resistive switching mechanism. The table below compares the primary memristor types, their typical conduction mechanisms, and the associated degradation origins.

Table 1: Comparative Analysis of Memristive Device Degradation

Memristor Type	Switching Mechanism	Primary Degradation Origin	Impact on Conduction & Trap States
Filamentary Cation (e.g., Ag/SiO₂)	Formation/rupture of metallic (e.g., Ag) conductive filaments (CFs) [61].	Uncontrolled growth/rupture of CFs; stochastic movement of metal cations [61].	High cycle-to-cycle variability; shift from ohmic (LRS) to tunnelling (HRS) conduction; trap generation at filament rupture points.
Filamentary Anion (e.g., TaOₓ, HfOₓ)	Formation/rupture of oxygen vacancy (Vₒ) filaments [61].	Random migration and recombination of oxygen ions/vacancies [61].	Device-to-device variability; trap-assisted tunnelling (TAT) in HRS; changes in Schottky barrier height at interfaces.
Interface-Based (e.g., Nb₂Oₓ)	Homogeneous modulation of interface barriers (e.g., Schottky) or tunnelling gaps [62].	Drift of ionic species (e.g., oxygen) within an ultra-thin layer, modifying interface properties [62].	More uniform switching; gradual resistance change; shift between tunnelling and thermionic emission currents.
Dual Ionic (e.g., Ta/HfO₂)	Competition and interaction between cation and anion migration [61].	Complex interplay and unpredictable kinetics of multiple ionic species [61].	Highly unstable switching parameters; dynamic creation/annihilation of trap states from both vacancy and metal defects.

A key differentiator is whether the device exhibits filamentary or non-filamentary (interface-based) switching. Filamentary devices suffer from inherent randomness, as the formation and rupture of nanoscale conductive paths are stochastic processes. This leads to significant cycle-to-cycle and device-to-device variability [61]. In contrast, interface-based devices, such as the double-barrier Nb₂Oₓ system, demonstrate a more homogeneous change in resistance across the entire device area, resulting in highly uniform and gradual switching that is less prone to catastrophic degradation [62].

Experimental Protocols for Degradation Analysis

A comprehensive understanding of degradation requires a synergistic approach that combines advanced in situ characterization with sophisticated device modeling. This multi-faceted methodology allows researchers to correlate real-time structural and chemical changes with electrical performance.

1In SituCharacterization Techniques

In situ techniques are indispensable for directly observing the dynamic evolution of the active region within a functioning device.

• In Situ Transmission Electron Microscopy (TEM): This technique has been pivotal in visualizing the growth and dissolution of conductive filaments in cation-based devices. For example, researchers have directly observed the step-by-step formation of Ag filaments and their subsequent retraction under a reverse bias, providing visual evidence of the kinetic processes that lead to variability [61].
• In Situ Spectroscopic and Electrical Monitoring: Combining electrical measurements with techniques like X-ray photoelectron spectroscopy (XPS) or impedance spectroscopy can identify chemical state changes and ionic mobility within the switching layer. This is particularly useful for anion and dual-ion systems, where direct observation of oxygen vacancies is challenging.

Device Modeling and State Characterization

Experimental data must be complemented with physical models to derive a quantitative understanding.

• Physics-Based Compact Models: These models use differential equations to describe the physical processes, such as ion drift and diffusion, that govern state transitions. For Self-Directed Channel (SDC) memristors, the Generalized Metastable Switch (MSS) model has been developed to describe the probabilistic nature of ionic agglomeration that constitutes the device state [63].
• Noise-Aware State Estimation: A critical challenge is identifying the device's internal state (e.g., filament geometry or interface barrier height) from noisy voltage and current measurements. Recent research introduces minimum-variance estimation techniques to accurately characterize the state of SDC memristors while quantifying the associated uncertainty, which is crucial for predicting degradation and drift [63].

The following diagram illustrates the synergistic relationship between these experimental and theoretical approaches in probing memristive degradation.

The Scientist's Toolkit: Essential Research Reagents and Materials

The following table details key materials and reagents commonly used in the fabrication and analysis of memristive devices, as referenced in the studies.

Table 2: Essential Research Reagents and Materials for Memristor Studies

Material / Reagent	Function in Research	Application Context
Niobium (Nb) / Niobium Oxide (Nb₂Oₓ)	Active memristive layer in interface-based devices [62].	Used in double-barrier Al/Al₂O₃/Nb₂Oₓ/Au structures to study homogeneous switching [62].
Silver (Ag)	Active electrode for cation migration [61].	Forms conductive filaments in ECM-type memristors (e.g., Ag/SiO₂) [61].
Hafnium Oxide (HfO₂)	Switching layer in anion-based memristors [61].	A widely studied high-κ dielectric for filamentary valence change memory (VCM) devices [61].
Tantalum Oxide (TaOₓ)	Switching layer for dual-ion mechanisms [61].	Exhibits switching involving both oxygen vacancies and metal cations (e.g., Ta/HfO₂) [61].
Aluminum Oxide (Al₂O₃)	High-quality tunnel barrier [62].	Serves as a stable, defect-free tunnel barrier in double-layer devices to confine switching to the adjacent interface [62].
Methyl Acetate (MeOAc)	Ester antisolvent for ligand exchange [25].	Used in perovskite quantum dot processing to replace long-chain insulating ligands with shorter conductive ones [25].
Oleic Acid / Oleylamine	Native capping ligands for quantum dots [64].	Provide colloidal stability to as-synthesized PQDs but act as insulating layers in solid films [64].

The path to reliable memristive devices lies in a fundamental understanding of their degradation mechanisms. As this guide has detailed, filamentary devices are plagued by inherent variability due to the stochastic nature of ionic filament formation, while interface-based devices offer a more uniform and controllable switching pathway. The relentless pursuit of this understanding through synergistic in situ characterization and device modeling is unlocking the physical origins of trap state creation and conduction mechanism shifts. Future research must continue to bridge the gap between microscopic material properties and macroscopic device performance, paving the way for the design of robust, high-endurance memristors that can fulfill their promise in next-generation computing and memory technologies.

Combating Sneak Path Currents in Crossbar Array Integration

The integration of memristive devices into high-density crossbar arrays is a cornerstone for developing next-generation non-volatile memory and neuromorphic computing systems. These architectures enable massively parallel computing capabilities, which are essential for accelerating artificial intelligence applications at the edge. However, a fundamental limitation plaguing their large-scale integration is the sneak path current phenomenon. In passive crossbar arrays without selective elements, unintended current paths emerge between adjacent cells during read/write operations, leading to misreading, crosstalk, and substantial data distortion. This issue becomes increasingly severe with array scaling, ultimately restricting the practical size and reliable operational window of memory arrays. Consequently, developing effective strategies to combat sneak path currents represents a critical research frontier in the advancement of in-memory computing architectures, directly impacting their energy efficiency, accuracy, and ultimate scalability.

Comparative Analysis of Sneak Path Mitigation Strategies

Various architectural and materials-based approaches have been developed to address the sneak path challenge, each with distinct advantages and limitations. The following table provides a systematic comparison of the primary mitigation strategies based on recent experimental demonstrations.

Table 1: Performance Comparison of Sneak Path Current Mitigation Technologies

Technology Approach	Key Structural Feature	Reported Rectification Ratio	Nonlinearity	Array Size Demonstrated	Key Advantages
Self-Rectifying Memristor (SRM) [65]	TiN/HfO~x~/Pt single layer	>10⁸	>10⁵	32×32 (simulated)	Record high RR; Simple fabrication; Eliminates external selector
Three-Terminal Memtransistor [66]	Monolayer MoS₂ with gate control	Read Margin: 10⁵	N/A	2048 devices per array	Dynamic weight tuning; Resolves inference ambiguities without retraining
Self-Rectifying Bipolar/Complementary Device [67]	BiFeO₃ (BFO) and BiFeTiO₃/BiFO₃ (BiBFO) structures	Efficient sneak path suppression demonstrated	N/A	64×64 (simulated)	Analyzed multiple reading schemes (OneWLPU, AllWLPU, FL)
MoS₂/Graphene Lateral Heterostructure [48]	Lateral Au/MoS₂/graphene architecture	ON/OFF Ratio: ≈2.1	N/A	Device level	Forming-free operation; BEOL-compatible, transfer-free fabrication

The comparative data reveals a clear performance trade-off between the achieved rectification ratio (RR) and technological maturity. The TiN/HfO~x~/Pt self-rectifying memristor demonstrates an exceptional rectification ratio exceeding 10⁸, which is crucial for suppressing leakage in large arrays [65]. In contrast, three-terminal memtransistors based on MoS₂, while offering more moderate ON/OFF ratios, provide the distinct advantage of dynamic conductance tuning via a gate terminal, enabling real-time resolution of classification ambiguities in neuromorphic applications without costly retraining [66]. The lateral MoS₂/graphene heterostructure offers a different trade-off, with a relatively low ON/OFF ratio but featuring forming-free operation and a fabrication process compatible with back-end-of-line (BEOL) temperatures [48].

Experimental Protocols and Methodologies

Fabrication of High-Performance Self-Rectifying Memristors

The record-breaking TiN/HfO~x~/Pt self-rectifying memristor was fabricated using a combination of deposition and annealing techniques [65]. The experimental workflow is as follows:

• Bottom Electrode Formation: A Pt bottom electrode is deposited and patterned.
• Active Layer Deposition: A 10 nm thick HfO~x~ switching layer is deposited via magnetron sputtering. This thickness balances adequate insulation with sufficient defect concentration for resistive switching.
• Rapid Thermal Annealing (RTA): A critical RTA step is performed at 150°C for 600 seconds in an oxygen-rich environment. X-ray photoelectron spectroscopy (XPS) confirms this step reduces oxygen vacancy (O~V~) concentration from 61.35% to 18.11%, which is essential for suppressing continuous conductive filament formation and enhancing self-rectification.
• Top Electrode Formation: A TiN top electrode is deposited and patterned to complete the metal-insulator-metal (MIM) structure.
• Electrical Characterization: DC current-voltage (I-V) sweeps are performed with the Pt electrode biased and the TiN electrode grounded. The device is operated using a 1/6 V scheme to optimize both the rectification ratio and nonlinearity.

Fabrication of MoS₂ Memtransistor Crossbar Arrays

The large-scale integration of three-terminal memtransistors involves a more complex process to form the gate, dielectric, and channel layers [66]:

• Gate Line Deposition: Dedicated Ti/Pt gate lines (word lines) are first defined on a Si/SiO₂ substrate using electron-beam evaporation.
• Dielectric Stack Growth: A charge-trapping gate stack is formed via atomic layer deposition (ALD), consisting of a 15 nm Al₂O₃ blocking layer, a 3 nm HfO₂ trapping layer, and a 7 nm Al₂O₃ tunneling layer.
• MoS₂ Channel Transfer: Large-area monolayer MoS₂, grown by metal-organic chemical vapor deposition (MOCVD), is wet-transferred onto the substrate and etched into discrete channels for individual devices.
• Source/Drain Formation: Ni/Au source and drain (bit) lines are evaporated onto the MoS₂ channel.
• Insulation and Interconnection: A final Al₂O₃ insulation layer and Ti/Au conductive bridges are deposited to complete the crosspoint connections. This process has yielded functional arrays of up to 2048 devices with a reported yield exceeding 92%.

Visualization of Mitigation Mechanisms

The following diagrams illustrate the core principles behind two dominant approaches for sneak path suppression.

Sneak Paths in Crossbar Arrays

Self-Rectifying Memristor Operation

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Key Materials and Reagents for Memristor Crossbar Research

Material/Reagent	Function in Research	Specific Example & Citation
Transition Metal Oxides	Serve as the active switching layer where conductive filaments form and rupture.	HfO~x~ [65]: Used in TiN/HfO~x~/Pt structures. Annealing controls oxygen vacancies.
2D Materials	Provide atomically thin channels for memtransistors, enabling gate-tunable conductance.	MoS₂ [66] [48]: Used as a semiconducting channel in memtransistors and lateral heterostructures.
Graphene & Derivatives	Act as ultra-thin, high-conductivity electrodes that minimize power consumption and interface effects.	Few-Layer Graphene (FLG) [48]: Serves as a low-resistance electrode in lateral MoS₂/graphene devices.
Electrode Metals	Form Ohmic or Schottky contacts to the active layer, influencing switching characteristics.	Pt, TiN, Au [65] [48]: Pt and TiN are common bottom/top electrodes. Au is used for contacts in lateral devices.
Annealing Equipment	Critical for post-deposition treatment to control defect density (e.g., oxygen vacancies) in oxide films.	Rapid Thermal Annealer (RTA) [65]: Used at 150°C to fine-tune HfO~x~ properties, drastically boosting rectification.

The relentless pursuit of higher density and energy efficiency in memory-centric computing necessitates effective solutions to the sneak path current challenge. Current research demonstrates two promising, complementary pathways: the development of high-performance self-rectifying memristors with exceptional native nonlinearity, and the implementation of three-terminal memtransistors that offer dynamic control for enhanced neuromorphic functionality. The choice between these strategies involves a fundamental trade-off between the simplicity and ultra-high rectification of two-terminal devices and the algorithmic flexibility provided by gate-tunable three-terminal devices.

Future progress will likely hinge on further material innovations, particularly in the refinement of perovskite quantum dots (PQDs) and other 2D materials where surface morphology and ligand engineering can precisely control charge transport and defect properties [60] [68]. The successful integration of these advanced materials into large-scale, high-yield crossbar arrays will be the ultimate benchmark for transitioning these promising laboratory demonstrations into commercially viable technologies for secure, low-power, and intelligent edge computing systems.

The performance of quantum dot (QD)-based memristors and other optoelectronic devices is profoundly influenced by their surface and interface properties. Defects on the QD surface and uncontrolled interfaces within device architectures can lead to charge trapping, ion migration, and non-radiative recombination, ultimately degrading device performance and stability [69] [70]. To address these challenges, researchers have developed sophisticated optimization strategies focused on the nanoscale engineering of QD surfaces and heterostructures. This guide provides a comparative analysis of three primary optimization levers: interface engineering, ligand exchange, and core/shell QD designs, with a specific focus on their impact on the memristive characteristics of perovskite quantum dots (PQDs). We present experimental data and methodologies to objectively evaluate the effectiveness of each approach in enhancing device performance for next-generation memory and neuromorphic computing applications.

Comparative Analysis of Optimization Strategies

The table below summarizes the key characteristics, performance outcomes, and associated challenges of the three main optimization strategies for quantum dots in memristive and optoelectronic applications.

Table 1: Comparison of QD Optimization Strategies for Memristive Applications

Optimization Strategy	Key Mechanism	Reported Performance Improvements	Challenges & Trade-offs
Interface Engineering	Passivation of surface defects and dangling bonds using organic/inorganic ligands [69].	Near-unity PLQY (98.56%), reduced non-radiative decay, enhanced solvent compatibility for photolithography (20.7 μm linewidth) [69].	Ligand stability under operational stress (bias, heat), potential introduction of unwanted insulating layers affecting charge transport [69] [71].
Ligand Exchange	Replacing long-chain native ligands with shorter or functionally designed molecules to improve charge transport [71].	Enables conductive QD films, improved performance in QD-based resistive random-access memory (RRAM) and photodetectors [71] [72].	Can compromise colloidal stability, may create new surface defects if not optimized, requires precise control over exchange process [71].
Core/Shell Design	Growth of a shell material with tailored band alignment around the core QD to confine charge carriers or facilitate separation [72].	In photodetectors, type-II heterostructures showed ~5 orders of magnitude higher Ilight/Idark ratio vs. type-I; responsivity of 10.64 mA/W at 0 V bias [72].	Complex synthesis; lattice mismatch between core and shell can induce strain and defects; thicker shells can hinder charge injection/extraction [72].

Experimental Protocols and Methodologies

Dual-Ligand Synergistic Passivation

Objective: To simultaneously suppress bulk and surface defects in CsPbBr3 PQDs, enhancing photoluminescence and solvent resistance for photolithography [69].

Synthesis & Passivation Protocol:

• Cs Precursor Preparation: Dissolve 1 mmol Cs2CO3 in 10 mL of octanoic acid (OTAc) with stirring at room temperature for 10 minutes [69].
• QD Synthesis with Europium Doping: Prepare a precursor solution by dissolving PbBr2 (1 mmol), tetraoctylammonium bromide (TOAB, 2 mmol), and varying amounts of europium acetylacetonate (Eu(acac)3 (e.g., 0, 0.1, 0.2, 0.3 mmol) in a suitable solvent. This introduces Eu³⁺ ions to compensate for Pb²⁺ vacancies in the lattice [69].
• Surface Ligand Exchange: Following the initial synthesis, introduce benzamide to the QD solution. The benzamide molecules, with their electron-rich amide groups, undergo ligand exchange to passivate under-coordinated Br⁻ ions on the QD surface [69].
• Purification: Precipitate the passivated QDs using a polar anti-solvent (e.g., methyl acetate) and isolate them via centrifugation. Repeat this purification step to remove excess ligands and reaction byproducts [69].

Key Characterization Techniques:

• Photoluminescence Quantum Yield (PLQY): Measure using an integrating sphere to quantify the efficiency of light emission. The dual-ligand strategy achieved a near-unity PLQY of 98.56% [69].
• Time-Resolved Photoluminescence (TRPL): Analyze fluorescence decay to understand charge carrier dynamics. The passivated QDs showed a short lifetime of 69.89 ns, indicating suppressed non-radiative recombination pathways [69].
• Density Functional Theory (DFT) Calculations: Model and confirm the strong binding energy of the Eu(acac)3 and benzamide ligands to specific defect sites (e.g., Pb²⁺ and Br⁻ vacancies) on the QD surface [69].

Core/Shell Quantum Dot Synthesis

Objective: To synthesize CdSe/CdS core/shell QDs with type-I, quasi type-II, and type-II band alignments for efficient self-powered photodetection [72].

Synthesis Protocol:

• CdSe Core Synthesis: Utilize a hot-injection method. Heat a mixture of Cadmium oxide (CdO), Oleic Acid (OA), and 1-Octadecene (ODE) to 220°C under nitrogen flow to form a clear Cd-OA complex solution. Rapidly inject a Tri-n-octylphosphine (TOP) solution containing Selenium (Se) to initiate nucleation and growth of CdSe cores. Vary the core nucleation time (1-5 minutes) to control the final core size [72].
• CdS Shell Growth: Employ the Successive Ionic Layer Adsorption and Reaction (SILAR) technique. This involves the dropwise, alternating addition of Cd and S precursor solutions to the core solution at a specific temperature (e.g., 180°C). The shell thickness is precisely controlled by the number of SILAR cycles added, fixed in this case to study the effect of core size [72].
• Band Alignment Control: By fixing the shell thickness and systematically increasing the CdSe core size, the band alignment naturally transitions from type-I (small core, e.g., CS38), to quasi-type-II (intermediate core, e.g., CS67), to type-II (large core, e.g., CS99) [72].

Key Characterization Techniques:

• Ultraviolet Photoelectron Spectroscopy (UPS): Determine the valence band maximum and work function of the core/shell QDs to construct their energy level diagrams and confirm the band alignment type [72].
• Time-Resolved Photoluminescence (TRPL): Measure carrier lifetime. Type-II structures typically exhibit longer lifetimes due to spatial separation of electrons and holes, which was observed in the champion device [72].
• Device Fabrication & I-V Testing: Fabricate photodetectors in a vertical sandwich structure (e.g., ITO/QD film/Al). Measure current-voltage (I-V) characteristics in dark and illuminated conditions to determine key metrics like light-to-dark current ratio (up to 10⁵ for type-II) and responsivity (10.64 mA/W at 0 V) [72].

Diagram 1: Core/Shell QD synthesis workflow for band engineering.

The Scientist's Toolkit: Essential Research Reagents

Table 2: Key Reagents for QD Surface and Interface Engineering

Reagent / Material	Function in Experiment	Key Outcome / Rationale
Europium Acetylacetonate (Eu(acac)₃)	Bulk defect passivator: Eu³⁺ ions compensate for positively charged Pb²⁺ vacancies in the perovskite lattice [69].	Stabilizes crystal framework, suppresses non-radiative recombination pathways from bulk defects [69].
Benzamide	Surface ligand: Short-chain ligand with amide groups that coordinate with under-coordinated Br⁻ ions on the PQD surface [69].	Passivates surface defects, reduces ligand steric hindrance, improves compatibility with polar photolithography solvents like PGMEA [69].
Polyvinylpyrrolidone (PVP)	Polymer matrix: Serves as the host material in organic-inorganic hybrid memristive devices, e.g., combined with N-doped Carbon QDs [38].	Provides a flexible, transparent, and solution-processable medium; its functional groups can interact with embedded QDs to enable analog resistive switching for synapses [38].
Propylene Glycol Monomethyl Ether Acetate (PGMEA)	Solvent: A standard polar solvent used in photolithography processes [69].	Tests the solvent resistance of ligand-engineered QDs; compatibility enables direct patterning of QD films via photolithography [69].
Cadmium Oxide (CdO) & Selenium (Se) Precursors	Core formation: Primary reactants for synthesizing CdSe quantum dot cores via the hot-injection method [72].	Allows precise control over core size, which is the primary variable for tuning band alignment in core/shell heterostructures [72].
Cadmium (Cd) & Sulfur (S) Precursors	Shell growth: Ionic sources for growing the CdS shell using the SILAR technique [72].	Enables layer-by-layer control of shell thickness, critical for optimizing charge carrier confinement and separation in heterostructures [72].

Signaling Pathways in Memristive Switching

The resistive switching in QD-based memristors, which underpins both memory and synaptic functions, is governed by ion migration and charge trapping/detrapping mechanisms. The following diagram illustrates the primary signaling pathways involved in the analog switching used for neuromorphic applications.

Diagram 2: Signaling pathways in QD-based memristive synapses.

Enhancing Flexibility and Stability for Wearable and Biomedical Applications

The convergence of flexible electronics with biomedical applications is driving a technological revolution in healthcare monitoring, drug delivery, and neuromorphic computing. Wearable and implantable devices require exceptional mechanical compliance to interface seamlessly with biological tissues while maintaining stable electrical performance under dynamic physiological conditions. This review objectively compares the performance of emerging material systems and device architectures designed to enhance both flexibility and operational stability. We focus specifically on memristive devices for neuromorphic applications and sensory materials for health monitoring, as these platforms represent the cutting edge in bio-integrated electronics. The fundamental challenge lies in reconciling the soft, dynamic nature of biological systems with the rigid, static properties of conventional electronic materials. By comparing quantitative performance data across multiple material platforms, this guide provides researchers with critical insights for selecting appropriate technologies for specific biomedical applications.

Comparative Analysis of Flexible Memristive Devices

Memristors have emerged as fundamental building blocks for neuromorphic computing and non-volatile memory in flexible systems. Their simple metal-insulator-metal (MIM) structure offers advantages for CMOS-compatible, power-efficient flexible electronics [6]. Different material systems and structural approaches have been developed to enhance both flexibility and switching stability in these devices.

Material Systems for Flexible Memristors

Table 1: Performance Comparison of Flexible Memristive Devices

Material System	Device Structure	Resistance Window	Endurance (Cycles)	Switching Speed	Mechanical Stability
Quantum Memristor	Ag/SiO₂/Pt	Quantum conductance levels (G₀, 2G₀)	Not specified	Not specified	Room temperature operation in air [73]
Tin Oxide Memristor	Ag/SnO₂₋ₓ/ITO	10⁵ Ω (electrical), 10⁴ Ω (pressure)	11×10³	Not specified	Pressure-dependent switching [6]
h-BN Memristor	Au/h-BN/GaN nano-cones	Multi-level conductance	Excellent cycle-to-cycle consistency	Not specified	Analog switching via filament confinement [3]
Ta/HfO₂ Memristor	Ta/HfO₂/Pt	24 discrete resistance levels	1.2×10¹¹ (record high)	≤5 ns	Reliable retention at 85°C [5]

Mechanisms for Enhanced Stability

Different memristive systems employ distinct physical mechanisms to achieve stable operation, particularly under mechanical stress:

Quantum Point Contacts in Nanoionic Devices: Ag/SiO₂/Pt memristive devices achieve quantum conductance levels (G₀, 2G₀) through an electrochemical polishing effect during the RESET process. This approach progressively narrows the conductive filament rather than abruptly breaking it, enabling predictable adjustment of stable quantum levels with deviations of only -3.8% and 0.6% from SI values for G₀ and 2G₀, respectively [73]. This creates an intrinsic resistance standard operable at room temperature in air.

Multiple Nano-filament Confinement in 2D Materials: Hexagonal boron nitride (h-BN) memristors with suspended structures over GaN nano-cones enable linear and symmetric analog switching through geometric confinement of multiple nano-filaments. This design prevents the formation of large-diameter conductive filaments that lead to abrupt switching behavior, instead facilitating sequential development of nano-filaments under increasing electric fields [3]. The confined structure reduces operating voltage and enhances energy efficiency while maintaining stable performance.

Oxygen Vacancy Modulation in Metal Oxides: Ta/HfO₂/Pt devices achieve record endurance (120 billion cycles) through composition modulation of a sub-10 nm Ta-rich and O-deficient conduction channel. The extracted activation energy of 1.55 eV contributes to extrapolated retention beyond 10 years at 162°C [5]. Similarly, SnO₂₋ₓ-based devices utilize oxygen vacancies (12.31% relative percentage) to enable forming-free gradual switching desirable for neuromorphic applications [6].

Flexible Substrates and Encapsulation Strategies

Achieving mechanical flexibility requires innovative substrate materials and encapsulation strategies that protect functional components while maintaining conformal contact with biological tissues.

Material Platforms for Flexible Substrates

Table 2: Performance Comparison of Flexible Substrate Materials

Material Platform	Young's Modulus	Thickness	Stretchability	Key Advantages	Applications
Ultrathin Polymers (PI, parylene)	∼2–5 GPa	Single-digit micrometers to submicrons	Limited intrinsic stretchability	Reduced bending stiffness, van der Waals adhesion	Epidermal electronics [74]
Elastomeric Composites (PDMS)	<5 MPa	Tens to hundreds of micrometers	>100% strain	Intrinsic softness, biocompatibility	Wearable thermal management [75]
Hydrogel-based Systems	10 kPa - 1 MPa	Variable	High hydration-dependent stretchability	Tissue-like mechanical properties, drug delivery	Biomedical implants, tissue engineering [76]
Porous/Mesh Architectures	Tunable	Millimeters to micrometers	>30% strain	Breathability, reduced effective resistance	Long-term wearable monitoring [74]

Thermal Management Substrates

PDMS composites containing paraffin microcapsules with silica nanoshells represent an advanced thermal management strategy for flexible electronics. The obtained PDMS composite (P10M7, 7 g paraffin microcapsules in 10 g PDMS matrix) withstands complex deformations (stretch, twist, stretch/twist) over 100% without failures. The silica nanoshells prevent paraffin leakage during solid-liquid phase transition while acting as support reinforced shells to maintain mechanical properties above the paraffin melting point [75]. This combination of thermal regulation and mechanical compliance is essential for stable operation of wearable devices in varying environmental conditions.

Experimental Protocols for Flexibility and Stability Assessment

Standardized experimental methodologies are essential for objective comparison of device performance across different material platforms.

Mechanical Cycling Tests

Procedure:

• Mount devices on custom stretching apparatus or curved substrates with controlled radius of curvature
• Subject to repeated bending cycles (typically 1,000-10,000 cycles) at specified strain rates
• Monitor electrical characteristics (resistance, capacitance, leakage current) in situ during deformation
• Perform pre- and post-test morphological characterization (SEM, AFM) to identify fracture points

Key Parameters: bending radius, cycling frequency, environmental conditions (temperature, humidity)

For ultrathin sensors (<1μm), van der Waals forces and capillary action enable adhesive-free conformal contact with skin. Testing should verify maintained adhesion under dynamic movements equivalent to joint flexion (~30% strain) [74].

Electrical Performance Validation

Memristive Device Characterization:

• Current-voltage (I-V) sweeps to determine SET/RESET voltages and resistance ratios
• Endurance testing: continuous switching between resistance states with pulse programming
• Retention tests: measure state stability over time at elevated temperatures (85°C common)
• Statistical analysis of cycle-to-cycle (C2C) and device-to-device (D2D) variability

For quantum memristors, interlaboratory comparison validates metrological consistency. Devices are driven into quantum conductance regime using electrochemical-polishing-based programming strategy, with deviations from SI values quantified [73].

Environmental Stability Assessment

Accelerated Aging Protocol:

• Expose devices to controlled environmental stressors (temperature, humidity, UV radiation)
• Perform periodic electrical and mechanical characterization
• Use Arrhenius modeling (for temperature studies) to extrapolate long-term stability

Hydrogel-based systems require additional hydration stability tests under varying humidity conditions [76].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Research Materials for Flexible Electronics Development

Material/Reagent	Function	Application Examples	Key Characteristics
Polydimethylsiloxane (PDMS)	Flexible substrate/encapsulation	Wearable sensors, stretchable electronics	Biocompatibility, transparency, tunable modulus [74] [75]
Methacrylated Gelatin (GelMA)	Hydrogel matrix for tissue integration	Drug delivery, cell carriers, tissue engineering	Photopolymerizable, contains RGD cell adhesion sites [76]
Hexagonal Boron Nitride (h-BN)	2D insulating layer	Memristive devices, neuromorphic computing	Wide bandgap (≈6 eV), excellent thermal conductivity [3]
Polycaprolactone (PCL)	3D-printable biodegradable polymer	Bone tissue engineering scaffolds, biomedical implants	Biodegradability, mechanical strength, FDA approval [77]
Hydrothermally Synthesized SnO₂₋ₓ	Oxygen-deficient switching layer	Memristive devices, gas sensors	Forming-free gradual switching, low-cost synthesis [6]
Paraffin Microcapsules with Silica Nanoshells	Phase change material for thermal regulation	Wearable thermal management	Leakage prevention, maintained mechanical properties during phase transition [75]

Structural Diagrams for Enhanced Understanding

Quantum Memristor Operation Mechanism

Diagram 1: Quantum memristor operation mechanism showing transition from unstable filament formation to stable quantum states through electrochemical polishing.

Material Selection Framework for Flexible Electronics

Diagram 2: Material selection framework for flexible electronics showing key consideration areas and their relationships.

The advancement of flexible and stable devices for wearable and biomedical applications requires careful balancing of multiple material properties and operational parameters. Quantum memristors based on electrochemical metallization mechanisms offer unprecedented stability through quantum conductance phenomena, while 2D materials like h-BN enable analog switching through nanoscale filament confinement. For substrate materials, PDMS composites and ultrathin polymers provide mechanical compliance, with specialized formulations addressing thermal management needs through phase change materials. Hydrogel systems offer exceptional biocompatibility for implantable applications but require careful handling of hydration stability.

Experimental data comparison reveals that optimal material selection depends heavily on specific application requirements: ultrathin platforms (<1μm) enable conformal contact for sensing applications, while composite elastomers provide robust mechanical protection for wearable devices. Memristive systems demonstrate trade-offs between switching speed, endurance, and mechanical flexibility, with metal oxide systems generally offering superior endurance while 2D materials provide better analog switching characteristics.

Future development should focus on enhancing the multimodal functionality of these systems while maintaining stability under real-world operating conditions, including dynamic mechanical stress, environmental fluctuations, and long-term biological exposure. The integration of self-healing capabilities and biodegradable components represents a promising direction for next-generation biomedical electronics.

Comparative Analysis and Validation Frameworks for PQD Morphologies

The development of memristive devices based on power quality disturbance (PQD) materials represents a frontier in neuromorphic computing and non-volatile memory technologies. These materials exhibit resistive switching (RS) capabilities, where their electrical resistance can be modulated between high and low states through external stimuli, enabling them to mimic neurosynaptic behaviors essential for brain-inspired computing. As research progresses, numerous material systems have emerged with distinct switching mechanisms and performance characteristics. This guide provides an objective comparison of key PQD material systems, detailing their resistive switching mechanisms, performance metrics, and material properties based on experimental data from current literature. The comparative analysis presented here serves to inform researchers and scientists in selecting appropriate material systems for specific memristive applications, from high-density memory to artificial synapses in neuromorphic hardware.

Experimental Protocols for PQD Material Characterization

Standardized Device Fabrication and Measurement

To ensure fair comparison across different PQD material systems, researchers employ standardized experimental protocols for device fabrication and electrical characterization. For two-terminal metal-insulator-metal (MIM) structures, the foundational architecture for most memristive devices, fabrication typically begins with substrate preparation involving ultrasonic cleaning in acetone, isopropanol, and deionized water followed by oxygen plasma treatment to remove organic residues [48]. Bottom electrodes are then deposited via physical vapor deposition (RF sputtering or electron-beam evaporation) using inert metals (Pt, Au) or active metals (Ag, Cu) depending on the intended switching mechanism.

The functional PQD layer is subsequently formed through various techniques: RF sputtering of amorphous precursors (for transition metal dichalcogenides), chemical vapor deposition (for 2D materials), or spin-coating (for organic/polymer materials) [48] [78]. Top electrodes are then deposited using shadow masks or lithographic patterning to define device areas. For electrical characterization, a semiconductor parameter analyzer applies voltage sweeps (typically 0 → +Vmax → 0 → -Vmax → 0) while measuring current compliance to prevent hard breakdown. Key metrics extracted include SET/RESET voltages, ON/OFF ratios, endurance cycles, and retention times at elevated temperatures (85-125°C) [48] [78].

Material-Specific Fabrication Variations

Specific material systems require tailored fabrication approaches. For MoS₂/graphene heterostructures, a transfer-free fabrication method involves direct sputtering of amorphous MoS₂ onto CVD-grown graphene followed by confined-space sulfurization at 800°C in a quartz tube furnace with elemental sulfur, producing 3-4 layer 2H-MoS₂ with 0.8-0.9 nm roughness [48]. For oxide-based RRAM, the functional layer often requires electroforming at higher voltages to initiate conductive filaments, whereas phase-change materials need tailored current pulses for reversible amorphous-crystalline transitions [79] [78]. Organic and polymer-based memristors typically employ solution-processing techniques but require controlled hydration-free environments to maintain stability [78].

The experimental workflow below illustrates the standard fabrication and characterization process for PQD-based memristive devices.

Comparative Performance of PQD Material Systems

The landscape of PQD materials for memristive applications encompasses diverse material classes, each with distinct advantages and limitations. The following table provides a systematic comparison of key performance metrics across major PQD material systems based on experimental data from current literature.

Material System	Switching Mechanism	ON/OFF Ratio	Set Voltage (V)	Endurance (Cycles)	Retention (s)	Energy Consumption (pJ/switch)
MoS₂/Graphene Heterostructure	VCM (Vacancy-driven) [48]	~2.1 [48]	~6 [48]	>10³ [48]	>10⁴ [48]	Not specified
Metal Oxide RRAM (HfO₂, TaOₓ)	ECM/VCM [79] [78]	10-10⁴ [78]	0.5-3 [78]	10⁶-10¹² [78]	>10⁷ [78]	0.1-10 [78]
Phase-Change Materials (GST)	Thermal (Phase Transition) [79]	10-10³ [79]	2-5 [79]	10⁸-10¹² [79]	>10⁹ [79]	10-100 [79]
Ferroelectric Tunnel Junctions	Polarization Switching [79]	10-100 [79]	1-3 [79]	10¹⁰-10¹² [79]	>10¹⁰ [79]	0.1-1 [79]
Organic/Polymer RRAM	Charge Transfer [78]	10²-10⁵ [78]	1-4 [78]	10³-10⁶ [78]	10⁴-10⁶ [78]	1-100 [78]
Low-Dimensional Silicon	Quantum Confinement [52]	10-10³ [52]	2-6 [52]	10⁴-10⁶ [52]	10⁵-10⁷ [52]	Not specified

Analysis of Performance Trends

The comparative data reveals significant trade-offs between different PQD material systems. Metal oxide RRAMs offer an optimal balance of performance metrics with excellent endurance (>10⁶ cycles) and moderate switching voltages (0.5-3V), making them strong candidates for commercial memory applications [78]. MoS₂/graphene heterostructures, while demonstrating relatively modest ON/OFF ratios (~2.1), provide the advantage of forming-free operation and compatibility with 2D material-based electronics, albeit with higher set voltages (~6V) [48]. Ferroelectric tunnel junctions stand out for their ultra-low energy consumption (0.1-1pJ/switch) and exceptional endurance, crucial for energy-efficient neuromorphic computing [79].

Materials such as phase-change memory demonstrate superior retention characteristics (>10⁹ seconds) but require higher switching energies (10-100pJ), limiting their density scaling due to thermal crosstalk [79]. Organic and polymer-based RRAMs achieve impressive ON/OFF ratios (10²-10⁵) through charge transfer mechanisms but face challenges with endurance variability (10³-10⁶ cycles) and environmental stability [78]. Emerging low-dimensional silicon structures leverage quantum confinement effects to achieve synaptic functionalities compatible with CMOS platforms, though comprehensive performance data remains limited compared to more mature material systems [52].

The Scientist's Toolkit: Essential Research Reagents and Materials

Successful research into PQD material systems requires specialized reagents and instrumentation. The following table catalogues essential research materials and their functions based on experimental methodologies from the literature.

Research Material	Function/Application	Example Specifications
RF Sputtering System	Deposition of uniform amorphous precursor layers	MoS₂ target (99.9%), 5×10⁻³ Torr Ar, 16W RF power [48]
CVD Reactor	Synthesis of 2D materials and heterostructures	CH₄:H₂ (24:8 sccm), 900-1000°C, 10 mbar [48]
Confined Sulfurization Setup	Crystallization of TMD precursors	Quartz tube furnace, 800°C, elemental sulfur (99.5%) [48]
Semiconductor Parameter Analyzer	I-V characterization and switching parameter extraction	Voltage sweeps (±10V), current compliance (1mA-100µA) [48] [78]
Polymer Transfer Materials	Graphene transfer to target substrates	PMMA (950K, A4), NaCl electrolyte (0.5M) [48]
Shadow Masks	Electrode patterning without lithography	Aluminum masks for discrete device zones [48]

Resistive Switching Mechanisms in PQD Materials

The operational principles of PQD-based memristive devices stem from fundamentally different physical mechanisms that govern their resistive switching behavior. Understanding these mechanisms is crucial for material selection and device optimization.

The Electrochemical Metallization (ECM) mechanism involves the formation and rupture of conductive filaments through redox processes and metal cation migration, typically in systems with active electrodes (Ag, Cu) [80] [78]. The Valence Change Mechanism (VCM), observed in transition metal oxides and 2D materials like MoS₂, relies on the migration of anion vacancies (oxygen, sulfur) that modulate Schottky barriers at metal-semiconductor interfaces [48] [78]. The Thermo-Chemical Mechanism (TCM) utilizes Joule heating to induce phase transitions between amorphous and crystalline states, particularly in phase-change materials like Ge₂Sb₂Te₅ [79]. Ferroelectric Polarization switching alters tunneling probabilities in ferroelectric tunnel junctions through polarization reversal [79]. Each mechanism presents distinct trade-offs in switching speed, energy efficiency, and retention characteristics that must be balanced for target applications.

This comparative analysis of PQD material systems reveals a diverse landscape of memristive technologies with complementary strengths and limitations. Metal oxide RRAMs currently offer the most balanced performance profile for memory applications, while 2D material heterostructures like MoS₂/graphene provide unique advantages for integration with emerging electronics platforms. Ferroelectric devices excel in energy efficiency, and phase-change materials demonstrate superior data retention. The optimal material selection depends critically on application requirements: low-power edge computing favors ferroelectric junctions, high-density memory benefits from oxide RRAM, and neuromorphic systems may leverage the analog switching capabilities of organic materials or 2D heterostructures. Future research directions should focus on improving endurance in organic systems, reducing operating voltages in 2D devices, and enhancing thermal stability across all material classes to advance PQD-based memristive technologies toward commercial viability.

Memristive devices, whose resistance depends on the history of applied voltage and current, have emerged as a cornerstone for next-generation computing architectures, including neuromorphic systems and in-memory computing [3] [81]. Their simple two-terminal structure enables high-density integration and operation at remarkably low power, sometimes consuming only atto joules of energy [57]. The core functionality of these devices hinges on the resistive switching layer, where nanoscale phenomena govern the overall electrical characteristics. Within this landscape, quantum dots (QDs) and other nanostructured materials have introduced a powerful paradigm for engineering memristive function. By controlling the size, density, and distribution of these nanoscale charge-trapping centers, researchers can directly influence key memristive parameters such as operating voltage, conductance range, switching linearity, and device stability [38]. This guide provides a comparative analysis of how different quantum dot and nanoparticle morphologies correlate with measurable memristive performance, offering a framework for the rational design of artificial synapses and non-volatile memory elements.

Comparative Analysis of Nanostructured Memristive Systems

The following table summarizes experimental data from recent studies on memristive devices incorporating various types of quantum dots, nanoparticles, and other confined structures. The data highlights the critical link between nanoscale morphology and macroscopic device function.

Table 1: Correlation between Nanostructure Morphology and Memristive Parameters in Different Material Systems

Material System & Nanostructure Type	Typical Size / Density / Distribution	Key Memristive Parameters	Synaptic Performance Metrics	Reported Variability (C2C/D2D)
NCQDs in PVP Matrix [38]	Size: 2-5 nm; Density: 40 wt% (optimal for A-RS)	Transition from Digital to Analog RS; Low Set/Forming Voltage	STP to LTP transition, PPF, STDP; Nonlinearity factor: 0.43	Stable over 100 potentiation/depression pulses
h-BN over GaN Nano-Cones [3]	Confinement at nano-cone apexes (geometry-induced)	Highly linear & symmetric analog switching; Reduced operating voltage	High accuracy in MNIST classification; Precise synaptic weight update	Excellent cycle-to-cycle consistency
HfO₂/Ta₂O₅ Bilayer [57]	Wafer-scale uniformity via single ALD system	Stable bipolar resistive switching; Requires electroforming	Paired-pulse facilitation/depression; Potentiation/Depression	CV (VSET): 6.09% D2D, 1.76% C2C
Ag in ZnO Nanowire [51]	Surface-confined Ag filament formation	Multilevel switching; Selector operation; Low switching voltages (<5 V)	Mimics Ca²⁺ dynamics; Paired-pulse facilitation	Self-limited forming process
PCMO-based Interface Switching [81]	Area-dependent switching (non-filamentary)	Analog, gradual SET/RESET; Resistance scales with device area	Spike-timing-dependent plasticity; LTP/LTD	Performance depends on tunnel barrier material

Experimental Protocols for Morphology-Function Correlation

Nanostructure Synthesis and Integration

The pathway to correlating morphology with function begins with the controlled synthesis and integration of nanostructures.

Hydrothermal Synthesis of Nitrogen-Doped Carbon Quantum Dots (NCQDs): A common protocol for creating charge-trapping centers involves synthesizing NCQDs via a one-pot hydrothermal process. In a typical procedure, citric acid serves as the carbon source and urea as the nitrogen source, dissolved in deionized water. The solution is transferred to a autoclave and heated to 160-200°C for 4-6 hours. The resulting NCQDs are then mixed with a polymer matrix like polyvinylpyrrolidone (PVP) at specific concentrations (e.g., 10-40 wt%) and spin-coated onto substrates to form the resistive switching layer [38]. The concentration directly determines the density of charge traps, with 40 wt% identified as optimal for analog resistive switching in this system.

MOCVD Growth of h-BN on Nano-Cones: For creating geometrically confined structures, a metal-organic chemical vapor deposition (MOCVD) process can be used. A GaN substrate is prepared, and at high temperatures (>1000°C), H₂ carrier gas is used to induce thermal decomposition of GaN at defect sites, forming nano-cones. Subsequently, few-layer h-BN is grown uniformly over this nano-structured surface, resulting in a suspended film where switching is confined to the nano-cone apexes [3]. This method leverages the substrate morphology rather than introducing foreign nanodots.

Electrical Characterization and Synaptic Emulation

After device fabrication, the following protocols are employed to characterize the memristive and synaptic properties.

Table 2: Key Experimental Reagents and Materials for Memristive Device Research

Research Reagent / Material	Primary Function in Experiments	Example Use-Case
Nitrogen-Doped Carbon QDs (NCQDs)	Serve as nanoscale charge-trapping centers in a polymer matrix.	Enables analog switching and synaptic plasticity in organic memristors [38].
Polyvinylpyrrolidone (PVP)	Acts as a host polymer matrix; provides flexibility and transparency.	Switching medium in flexible, transparent memristive synapses [38].
h-BN (Hexagonal Boron Nitride)	An atomically thin, excellent insulating switching medium.	Used as the resistive switching layer in memristors for neuromorphic computing [3].
HfO₂ / Ta₂O₅ Bilayer	High-k dielectric bilayer for stable, analog resistive switching.	Enables synaptic functionalities in wafer-scale, CMOS-compatible devices [57].
Pr₀.₇Ca₀.₃MnO₃ (PCMO)	A mixed-valence manganite for area-dependent, non-filamentary switching.	Used as the active layer in memristors emulating STDP and LTP/LTD [81].
ZnO Nanowires	Single-crystalline semiconductor serving as a solid electrolyte.	Platform for realizing full memristive functionalities in a nanoscale device [51].

DC I-V Sweep and Endurance Testing: A voltage sweep (e.g., 0 V → +2 V → 0 V → -2 V → 0 V) is applied to the device while measuring the current response. This identifies the fundamental switching behavior—whether it is abrupt (digital) or gradual (analog). This test is repeated for dozens to hundreds of cycles to assess endurance and cycle-to-cycle (C2C) variability [57] [38]. Parameters extracted include SET/RESET voltages, ON/OFF ratio, and the linearity of conductance change.

Pulse-Based Synaptic Emulation: To emulate biological synapses, trains of identical presynaptic voltage pulses are applied. For potentiation (synaptic strengthening), consecutive positive pulses are applied, and the conductance is read at a low read voltage after each pulse. The process is reversed for depression (synaptic weakening) using negative pulses. The number of pulses required to span the conductance range and the symmetry of the potentiation/depression curves are key metrics of synaptic quality [38] [81]. This directly tests the device's capability for multi-level storage and its suitability for neuromorphic computing.

Discussion: Linking Morphological Traits to Device Performance

The Impact of Nanostructure Density on Switching Linearity

The density of integrated nanostructures, such as QDs, is a critical factor determining the transition between digital and analog switching modes. In the case of NCQDs in a PVP matrix, a low density (10 wt%) leads to digital, abrupt resistive switching. This is because few, isolated conductive paths form and rupture catastrophically. At an optimal density (40 wt%), the filled trapping centers promote the formation of multiple, stable conductive paths, allowing for a gradual, analog change in device conductivity, which is essential for synaptic weight updates [38]. A similar principle applies to the confinement of multiple nano-filaments in the h-BN memristors, where the structured substrate encourages distributed switching events, resulting in highly linear and symmetric synaptic updates [3].

The Role of Nanostructure Distribution and Confinement in Variability

The spatial distribution of nanostructures or the confinement of the switching region directly impacts device-to-device (D2D) and cycle-to-cycle (C2C) variability—a major challenge in memristor technology. Wafer-scale deposition techniques like atomic layer deposition (ALD) for HfO₂/Ta₂O₅ bilayers ensure high uniformity, yielding low C2C variability (e.g., 1.76% for VSET) [57]. Conversely, geometric confinement, as seen in the h-BN suspended over GaN nano-cones, localizes the high electric field to specific points. This controlled environment restricts filament formation to predefined locations, drastically improving the reproducibility of switching events and reducing operational variability [3].

Nanostructure Size and Composition in Determining Switching Mechanism

The size and chemical nature of the nanostructure often dictate the fundamental switching mechanism. For instance, in ZnO nanowire-based devices, the use of an electrochemically active Ag electrode is crucial. The migration of Ag⁺ ions and their reduction on the ZnO surface leads to the formation of a conductive filament, classifying it as an Electrochemical Metallization Memory (ECM) cell [51]. In contrast, PCMO-based devices exhibit area-dependent switching, where the resistance change occurs uniformly across the entire device area rather than through a narrow filament [81]. This mechanism is inherently different from filamentary switching and results in a gradual, analog SET/RESET process that is highly desirable for synaptic emulation, with performance further tuned by the choice of tunnel barrier material (e.g., Al₂O₃, Ta₂O₅).

This comparison guide establishes a clear correlation between the morphology of integrated nanostructures—their size, density, and distribution—and the functional parameters of memristive devices. The experimental data and protocols presented provide a roadmap for researchers to systematically engineer memristive synapses. Key findings indicate that optimal nanostructure density is paramount for achieving linear analog switching, that uniform distribution or geometric confinement is crucial for low variability, and that the choice of material and switching mechanism (filamentary vs. area-dependent) dictates the applicability for specific neuromorphic functions. Future research should focus on refining the precision of nanostructure placement and exploring new hybrid material systems to further enhance the performance and reliability of brain-inspired computing hardware.

The advancement of memristor technology is fundamentally reliant on robust experimental methods to characterize and validate device performance. For researchers investigating the memristive characteristics of different perovskite quantum dot (PQD) surface morphologies, a suite of core techniques provides the critical data required to correlate material structure with device function. This guide objectively compares the experimental protocols and data output of four essential validation techniques—I-V Characterization, In-Situ Kelvin Probe Force Microscopy (KPFM), Endurance, and Retention Testing—by synthesizing current methodologies and performance metrics from recent literature. These techniques collectively enable a comprehensive analysis, from fundamental current-voltage dynamics to nanoscale electrostatic mapping and long-term reliability, providing a complete toolkit for the rigorous comparison of emerging memristive materials and architectures [68] [37].

Core Validation Techniques: Experimental Protocols and Data Comparison

The following section details the standard operating procedures for each key validation technique, providing a foundation for their application in comparative memristor studies, particularly for PQD surface morphology research.

I-V Characterization

Experimental Protocol: Current-Voltage (I-V) characterization is the foundational technique for identifying memristive behavior. The standard methodology involves applying a voltage sweep (e.g., a triangular or linear ramp waveform) across the two terminals of the memristive device while simultaneously measuring the current response [82] [83]. For volatile threshold switching devices, the sweep must capture both the forward (SET) and reverse (RESET) transitions. It is critical to use a source meter unit capable of both sourcing voltage and measuring current with high precision. To protect the nascent conductive filament in forming-free devices, a current compliance is always set in series during the SET process. The measurement should be performed in a shielded probe station to minimize noise, and multiple sweep cycles are conducted to confirm the stability and repeatability of the hysteretic I-V loop, which is the fingerprint of memristance [83].

Representative Data Output: The primary output is a plot of current (I) versus voltage (V), which exhibits a pinched hysteresis loop. Key quantitative metrics extracted include the SET voltage (( V{SET} )), RESET voltage (( V{RESET} )), ON/OFF ratio (the ratio of current in the low-resistance state (LRS) to the high-resistance state (HRS) at a read voltage), and the nonlinearity of the switching curve.

In-Situ KPFM

Experimental Protocol: Band Excitation Kelvin Probe Force Microscopy (BE-KPFM) is an advanced scanning probe technique that overcomes the limitations of traditional single-frequency KPFM. It operates by applying a multi-frequency voltage waveform (the "band excitation") to the AFM tip while it scans the surface. This allows for the simultaneous and quantitative detection of both the electrostatic force and the electrostatic force gradient between the tip and the sample [84]. A significant advancement is 3D-KPFM, or Force Volume (FV) BE-KPFM, wherein this measurement is repeated at multiple tip-sample distances at every pixel on the surface. This generates a 3D map of the electrostatic potential, which is crucial for deconvolving the true surface potential from long-range capacitive contributions from the cantilever and tip cone [84]. For in-situ studies, this can be performed while the device is under bias or in operando, providing a dynamic view of potential distribution during switching.

Representative Data Output: The technique yields a high-resolution 2D or 3D map of the surface contact potential difference (CPD). Force gradient detection typically provides superior lateral resolution compared to force detection alone. For memristors, this can visualize the formation and rupture of conductive filaments, charge trapping/detrapping at interfaces, and potential variations across different grain boundaries or surface morphologies [84].

Endurance Testing

Experimental Protocol: Endurance testing evaluates the stability of a memristor's switching over repeated operation cycles. The standard test involves applying a continuous sequence of voltage pulses to the device. A typical cycle consists of a SET pulse, a read pulse (to verify the LRS), a RESET pulse, and another read pulse (to verify the HRS) [46]. The amplitude, width, and shape (e.g., rectangular, triangular) of these pulses are carefully controlled by a pulse generator or a semiconductor parameter analyzer. The test is automated to run for hundreds of thousands to millions of cycles, during which the device's resistance states are continuously monitored. This methodology directly tests the reversibility and fatigue of the physical switching mechanism (e.g., filament formation/rupture) [46].

Representative Data Output: The primary result is a plot of the HRS and LRS resistances (often on a log scale) as a function of the cycle number. Key metrics are the number of sustained cycles before failure (e.g., when the ON/OFF ratio degrades below a threshold, typically 10) and the cycle-to-cycle variability of the resistance states.

Retention Testing

Experimental Protocol: Retention testing assesses the non-volatility of a memristor by measuring its ability to maintain a programmed resistance state over time. The experiment begins by programming the device into a specific state (either LRS or HRS) using an appropriate voltage pulse. The device is then left unbiased, often at an elevated temperature (e.g., 85°C or 125°C) to accelerate aging and identify failure mechanisms more rapidly. At regular time intervals, a small, non-destructive read voltage (e.g., 0.1-0.2 V) is applied to measure the device's resistance without disturbing the state. The test continues until the state is lost or for a duration specified by industry standards (e.g., 10 years extrapolated) [46].

Representative Data Output: The data is presented as a plot of normalized resistance versus time. The critical metric is the extrapolated retention time, which is the time for which the resistance state remains stable before significant decay. For non-volatile memory, a retention time exceeding 10 years is typically targeted.

Table 1: Comparison of Key Validation Techniques for Memristor Characterization

Technique	Primary Function	Key Measured Parameters	Spatial Resolution	Temporal Resolution	Key Advantages
I-V Characterization	Probe fundamental switching behavior	SET/RESET voltages, ON/OFF ratio, I-V hysteresis	Device-level	Medium (ms-s)	Directly confirms memristance; fast screening
In-Situ KPFM	Map nanoscale surface potential	Contact Potential Difference (CPD), charge distribution	Nanoscale (sub-10 nm) [84]	Low (minutes/hours)	Visualizes charge dynamics; identifies filament location
Endurance Testing	Evaluate switching cyclability	Cycle number, HRS/LRS stability over cycles	Device-level	Very Slow (hours/days)	Quantifies operational lifetime and reliability
Retention Testing	Assess state stability over time	Data retention time, resistance decay rate	Device-level	Very Slow (hours/days)	Quantifies non-volatility and memory retention

Table 2: Typical Performance Metrics for Memristor Devices from Literature

Performance Metric	Reported State-of-the-Art Ranges	Implication for PQD Morphology Studies
ON/OFF Ratio	(10^3) to (10^8) [46]	Indicates readability margin; influenced by filament strength/interface quality.
Endurance	(10^3) to (10^{12}) cycles [46]	Reflects switching mechanism reversibility; filament stability is key.
Retention	>10 years (extrapolated) [46]	Essential for non-volatile memory; tests charge trapping stability.
Switching Speed	<10 ns [46]	Relevant for high-speed applications; depends on ionic mobility.
SET Voltage	0.1 - 5 V	Lower voltage is desirable for energy efficiency; relates to formation barrier.

Experimental Workflow for Memristor Validation

The following diagram illustrates the logical progression and integration of the four core validation techniques in a typical research workflow for characterizing memristive devices.

The Scientist's Toolkit: Essential Research Reagents and Materials

The experimental characterization of memristors relies on a specific set of instruments, materials, and software tools. The following table details key solutions used in the featured techniques.

Table 3: Essential Research Reagent Solutions for Memristor Characterization

Item Name / Category	Function / Role in Experimentation	Representative Examples / Notes
Semiconductor Parameter Analyzer	Sources voltage/current and precisely measures the electrical response during I-V, endurance, and retention tests.	Keithley 2400/4200 Series; Keysight B1500A. Critical for setting current compliance.
Atomic Force Microscope (AFM)	Provides the platform for KPFM, enabling nanoscale topographical and surface potential imaging.	Systems from Bruker, Oxford Instruments/Cypher AFM. Must support BE-KPFM mode for quantitative measurements [84].
Band Excitation (BE) KPFM Kit	Advanced software/hardware for multi-frequency KPFM, enabling simultaneous force and force gradient detection.	Essential for 3D-KPFM to deconvolve long-range capacitive coupling [84].
Probe Station & Hot Chuck	Provides electrical and thermal control for device testing, including elevated temperature retention studies.	Micromanipulators with shielded enclosures for low-noise measurement.
Pulse Generator	Applies the precise voltage/current pulse trains required for endurance and speed testing.	Can be integrated into high-end parameter analyzers.
Memristor Emulator Circuit	Allows for circuit simulation and validation using commercial components to model memristor behavior.	Built with op-amps (e.g., UA741CN) and multipliers (e.g., AD633JN) for SPICE simulations [83].
Physical Memristor Devices	The devices under test (DUT), often in Metal-Insulator-Metal (MIM) or crossbar array structures.	Knowm Inc. "Type W" memristor [83]; devices based on 2D materials (e.g., MoSe₂) [82].

Critical Considerations for Technique Selection

When deploying these techniques to compare memristive characteristics across different PQD surface morphologies, several factors are paramount. The choice of technique should be guided by the specific research question, whether it is probing the fundamental switching mechanism (I-V, KPFM) or assessing device viability for applications (endurance, retention). A significant challenge in KPFM is the convolution of signals from the tip apex, cone, and cantilever, which can be mitigated by employing 3D-KPFM (FV BE-KPFM) to achieve quantitative deconvolution [84]. For array-level testing, the sneak path problem in crossbar structures can corrupt data during read operations; integrating a selector device (e.g., a transistor in a 1T1R configuration) is often necessary to ensure accurate endurance and retention measurements [46]. Finally, the interplay between material properties and observed performance is critical; for instance, defect engineering in 2D materials like MoSe₂ via plasma treatment can directly induce and modulate memristive behavior, which can be characterized by this suite of techniques [82].

Memristive devices have emerged as promising hardware for neuromorphic computing, capable of mimicking the neurosynaptic functions of the human brain and overcoming the limitations of traditional von Neumann architecture [85]. The core of this technology lies in resistive switching (RS), a phenomenon where a material changes its electrical resistance in response to an applied voltage. This switching can manifest as either digital (abrupt, binary changes suitable for data storage) or analog (gradual, multi-level changes ideal for neuromorphic computing) [86]. The specific properties of the active medium within the memristor dictate which type of switching dominates.

Polyvinylpyrrolidone (PVP) blended with N-doped carbon quantum dots (NCQDs) forms a sophisticated nanocomposite system for advanced memristive applications. This case study provides a objective comparison between the digital and analog resistive switching characteristics observed in NCQD-PVP composites, framing the analysis within a broader thesis on how different quantum dot surface morphologies influence memristive behavior. We summarize quantitative performance data, detail key experimental methodologies, and identify essential research reagents to equip scientists and developers with the necessary tools for further innovation in this field.

Comparative Performance Analysis of Switching Modes

The performance of NCQD-PVP composite-based memristive devices is quantified against key metrics for both digital and analog resistive switching, with comparisons to other quantum dot and composite systems.

Table 1: Digital Resistive Switching Performance Comparison

Material System	Switching Voltage (SET/RESET)	ON/OFF Ratio	Endurance (Cycles)	Retention	Key Mechanism	Reference
NCQD-PVP Composite	Information Missing	Information Missing	>1,000 (flexible)	Information Missing	Space charge trapping/detrapping [87]	[87]
CuSCN (~168 nm)	Information Missing	Information Missing	Information Missing	Information Missing	Schottky emission, charge trapping [86]	[86]
Sputtered ZnO	< ±3 V	~1.8 - 2.5	1,000 - 20,000	Information Missing	Oxygen vacancy filament [7]	[7]
MoS₂/Graphene	~ +6 V (SET)	~2.1	Information Missing	Information Missing	Vacancy-induced Schottky-barrier modulation [48]	[48]

Table 2: Analog Resistive Switching & Synaptic Performance

Material System	Synaptic Functions Demonstrated	Stimulus Required	Power Consumption	Key Plasticity	Reference
NCQD-PVP Composite	EPSC, PPF, STDP, STP to LTP transition	Electrical pulses	Information Missing	Short-term to Long-term	[87]
CuSCN (~120 nm)	Analog RS (gradual resistance change)	Electrical sweeps	Information Missing	N/A	[86]
CsPbBr₃/PDMS Nanospheres	Multi-dimensional sensing (humidity, temperature, pressure)	Environmental stimuli	Optical readout	Stimulus-specific fluorescence response [88]	[88]

Key Comparative Insights:

• NCQD-PVP for Neuromorphics: NCQD-PVP composites excel in emulating biologically plausible synaptic functions like excitatory postsynaptic current (EPSC) and spike-timing-dependent plasticity (STDP) [87]. This makes them particularly valuable for brain-inspired computing, unlike simpler binary switches.
• Switching Mechanism Divergence: The primary mechanism in NCQD-PVP devices is attributed to the trapping and detrapping of space charge [87], which is inherently more conducive to gradual analog switching. In contrast, other systems like ZnO often rely on the formation and rupture of conductive filaments [7], which can be stochastic and better suited for digital switching.
• Flexibility and Transparency: A significant advantage of the organic NCQD-PVP system is its inherent flexibility and transparency, demonstrated to withstand repeated mechanical bending [87]. This is a distinct advantage for wearable and transparent electronics applications over inorganic counterparts like sputtered ZnO [7].

Underlying Switching Mechanisms

The fundamental difference between digital and analog switching in memristive devices stems from distinct physical mechanisms, which are highly influenced by the active material's composition and structure.

Analog Switching via Space Charge Modulation

In NCQD-PVP composites, analog resistive switching is governed by the trapping and detrapping of space charge [87]. In-situ Kelvin probe force microscopy has confirmed this mechanism. The NCQDs dispersed within the PVP matrix act as charge trapping sites. Upon application of an electric field, charge carriers are injected and captured by these sites, modifying the local electric field and leading to a gradual change in the device's conductance. This continuous modulation of conductance is what enables the emulation of synaptic weight changes, making it ideal for neuromorphic computing [87].

Digital Switching via Filament Formation

In contrast, digital switching is often associated with more abrupt phenomena. In electrochemical metallization (ECM) cells, such as those using a copper electrode and a polymer electrolyte, digital switching occurs through the formation and rupture of conductive filaments [86]. A redox reaction at the active electrode (e.g., Cu) leads to the ionization of metal atoms. These metal ions migrate through the switching medium and are reduced to form a nanoscale metallic filament that shorts the two electrodes, causing a sharp drop in resistance (SET). The filament is then ruptured by a reverse voltage, resetting the device to a high resistance state (RESET) [86].

The diagram below illustrates the operational logic and contrasting mechanisms behind these two switching modes.

Experimental Protocols for Key Analyses

Fabrication of Flexible NCQD-PVP Memristive Synapses

The protocol for creating the core NCQD-PVP composite device is outlined as follows [87]:

• Solution Preparation: Prepare a PVP solution in a suitable solvent. Synthesize N-doped Carbon Quantum Dots (NCQDs) separately and dope them into the PVP solution at varying concentrations to form a homogeneous nanocomposite.
• Film Deposition: Deposit the PVP/NCQD solution onto a flexible substrate (e.g., ITO-coated PET) using a solution-processing technique like spin-coating to form a uniform thin film.
• Electrode Patterning: Deposit top electrodes onto the dried nanocomposite film through a shadow mask using thermal evaporation or sputtering, completing the metal-insulator-metal (MIM) device structure.
• Electrical Characterization: Use a semiconductor parameter analyzer (e.g., Keithley 4200-SCS) to measure current-voltage (I-V) characteristics. Apply voltage sweeps and specific pulse trains to the devices to characterize both digital switching and analog synaptic behaviors.

Probing the Memristive Mechanism

The space charge mechanism in NCQD-PVP devices was confirmed experimentally [87]:

• Technique: In situ Kelvin Probe Force Microscopy (KPFM).
• Procedure: The surface potential of the PVP/NCQD composite film is mapped under different electrical biases. Changes in the surface potential distribution directly correlate with the dynamic trapping and detrapping of charge carriers within the film, providing visual evidence of the analog switching mechanism.

Synaptic Function Emulation

Critical synaptic plasticity behaviors are tested using these protocols [87]:

• Paired-Pulse Facilitation (PPF): Two identical voltage pulses are applied in quick succession. The response to the second pulse is facilitated (i.e., a higher postsynaptic current is observed) due to the residual trapped charge from the first pulse, mimicking short-term synaptic plasticity.
• Spike-Timing-Dependent Plasticity (STDP): Pre-synaptic and post-synaptic neurons are simulated with precisely timed voltage spikes. The change in synaptic weight (device conductance) is measured as a function of the timing difference between the two spikes, replicating a fundamental Hebbian learning rule.
• Transition from STP to LTP: Repeated voltage pulse training is applied. Initially, the conductance change decays quickly (STP). With sufficient training, the conductance change becomes persistent (LTP), emulating the transition from short-term to long-term memory.

The workflow for developing and characterizing these synaptic devices is summarized below.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Essential Materials for NCQD-PVP Memristor Research

Material / Reagent	Function in the Experiment	Key Considerations
Polyvinylpyrrolidone (PVP)	Polymer matrix; provides flexibility, transparency, and host for charge trapping.	Molecular weight (e.g., 40,000 g/mol [89]) affects film formation and properties.
N-doped Carbon Quantum Dots (NCQDs)	Active nanomaterial; primary charge trapping sites that enable resistive switching.	Doping concentration and surface chemistry are critical for controlling trapping dynamics [87].
Indium Tin Oxide (ITO)	Transparent conductive electrode; enables optical transparency and electrical contact.	Coated on flexible substrates (e.g., PET) for flexible device architecture [87].
Solvent (e.g., Chloroform)	Dissolves PVP and disperses NCQDs for solution processing.	Purity and choice of solvent impact solution homogeneity and final film quality [89].
Keithley 4200-SCS	Semiconductor parameter analyzer; for measuring I-V characteristics and pulse testing.	Essential for quantifying resistive switching parameters and synaptic functions.
Kelvin Probe Force Microscope	Advanced characterization tool; maps surface potential to probe charge trapping mechanism.	Provides direct evidence of the space charge mechanism [87].

This case study demonstrates that NCQD-PVP composites are a versatile platform for memristive devices, capable of exhibiting both digital and analog resistive switching. The primary strength of this material system lies in its robust analog switching and synaptic plasticity, driven by a space charge trapping/detrapping mechanism. This makes it exceptionally suitable for neuromorphic computing applications that require emulation of biological learning processes. Furthermore, its solution processability, optical transparency, and mechanical flexibility [87] offer significant advantages for next-generation wearable and transparent electronics. When compared to other material systems like filament-based ZnO [7] or Schottky-barrier modulated MoS₂/graphene [48], the NCQD-PVP composite presents a compelling trade-off focused on low-power, brain-inspired computation rather than pure digital data storage. Future research should focus on enhancing the switching speed, improving device-to-device uniformity, and further exploring the intricate relationship between NCQD surface morphology and the fidelity of synaptic emulation.

Establishing Standardized Testing Protocols for Fair Cross-Study Comparison

The rapid advancement of memristor technologies has unveiled significant potential for revolutionizing non-volatile memory, neuromorphic computing, and reconfigurable hardware. However, the proliferation of diverse memristive technologies, each bearing distinct performance metrics across multi-bit memory capacity, low-power operation, endurance, retention, and stability, has created a critical challenge for the research community: the inability to make fair and meaningful cross-comparisons between devices reported in different studies. This challenge is particularly acute in the emerging field of perovskite quantum dot (PQD) surface morphology research, where subtle variations in material structure can significantly influence memristive characteristics yet no standardized framework exists for their evaluation. The absence of a unified testing protocol hampers the validation of published data, obscures the true performance trade-offs of different material approaches, and ultimately slows the translation of research findings into practical applications [90].

The fundamental obstacle lies in the historical lack of a holistic, technology-agnostic characterization methodology that can be universally adopted. While individual techniques for evaluating specific memristor attributes exist, they are often limited to individual performance metrics and are not integrated into a coherent workflow [90]. This review addresses this gap by presenting a comprehensive electrical characterization methodology, amalgamating several testing protocols into an appropriate sequence adapted for memristor benchmarking. The described approach is designed to extract information on all aspects of device behavior, from deciphering underlying physical mechanisms to assessing electrical performance, relying solely on standard instrumentation accessible in most electronics laboratories [90]. By establishing this common framework, researchers can objectively compare the memristive characteristics of different PQD surface morphologies, accelerating the development and optimization of next-generation memristive devices.

Core Standardized Testing Methodology

A robust characterization methodology for memristive devices must follow a logical sequence that progresses from fundamental functionality checks to advanced performance benchmarking. The illustrated workflow ensures that each stage of testing builds upon the knowledge gained from the previous one, providing a complete picture of device capabilities and limitations. The entire protocol is structured to be technology-agnostic, making it equally applicable for evaluating different PQD surface morphologies as well as other memristive systems [90].

The following diagram outlines the sequential, modular structure of the standardized characterization protocol:

Phase 1: Fundamental Functionality Testing

Switching Dynamics Analysis

The initial characterization step determines the basic switching capability of the memristive device. This is accomplished through a two-stage algorithm that first employs pulses of alternating polarity to determine the direction of resistance change for a given stimulus polarity, followed by voltage ramps to quantify the actual resistance change [90]. This analysis is crucial for establishing the operational boundaries of the device and identifying its switching threshold for pre-defined pulse widths. For PQD-based devices, this reveals whether the surface morphology influences the fundamental switching polarity (bipolar vs. unipolar) and threshold voltages.

The pulse-testing protocol should utilize pulses with fast rise and fall times (or respectively longer pulse widths) to impose a quasi-DC operating regime and minimize bandwidth limitations caused by system or device capacitance [90]. Through this methodology, researchers can classify PQD memristors as exhibiting bipolar switching (where SET and RESET occur at opposite polarities), unipolar switching (where SET and RESET occur at the same polarity), or hybrid behavior. Establishing this foundational characteristic is essential for all subsequent testing and for comparing different PQD surface morphologies.

Switching Metastability Evaluation

Once basic functionality is confirmed, the stability of the device in a given resistive state must be assessed through metastability evaluation. This involves applying non-switching pulses (with voltage amplitudes below the previously determined switching threshold) to monitor the DUT's resistive state without directly affecting it [90]. Variations in resistance under these conditions correlate with inherent device instability, potentially caused by mechanisms such as re-oxidation [90].

For PQD devices with different surface morphologies, this test can reveal crucial differences in temporal stability, which may arise from variations in surface energy, defect density, or interfacial properties. The timescale of resistance variations must be determined to inform experiment planning and application design. Devices exhibiting significant volatility over short timescales may be unsuitable for non-volatile memory applications but could potentially be exploited for neuromorphic computing functions that require short-term plasticity.

Temperature-Dependent Current-Voltage (I-V) Characterization

Temperature-dependent I-V characterization represents perhaps the most celebrated testing procedure for memristive technologies, providing insights into both operational characteristics and underlying physical mechanisms. Recording I-V characteristics across a temperature range enables researchers to: (i) verify device operation and switching type, (ii) identify volatile or stable switching capability, and (iii) gain deep insight into device physics [90].

The graphical signatures obtained through temperature-dependent I-V measurements provide critical fingerprints of the conduction mechanisms, which can be fundamentally influenced by PQD surface morphology:

Asymmetric I-V responses typically indicate transport determined by interface barriers, while symmetric curves suggest bulk-material controlled transport [90]. Unstable or non-reproducible I-V characteristics may be associated with metastability induced by movable ions, such as oxygen vacancies [90]. For PQD devices, surface morphology profoundly influences interfacial properties and defect distributions, making this characterization essential for correlating structural features with conduction mechanisms.

Phase 2: Performance Benchmarking

Memory Capacity Assessment

After establishing fundamental functionality, benchmarking routines evaluate actual performance metrics. The first such test determines the multilevel memory capacity of the device through a bespoke programming protocol that identifies the ultimate number of non-volatile resistive states the device can reliably maintain [90]. This assessment is crucial for establishing the potential of PQD devices to operate in an analogue fashion, a key requirement for neuromorphic computing applications where synaptic weights are represented by continuous resistance states.

The testing protocol must employ stimuli that progressively explore the device's resistive range while minimizing irreversible changes. For PQD-based memristors, surface morphology may significantly influence the number and stability of accessible intermediate states through effects on filament formation uniformity or interface trapping characteristics. Smooth, uniform morphologies may enable more precise and stable intermediate states compared to irregular surfaces with high roughness.

Retention and Endurance Testing

Retention and endurance represent two of the most critical performance metrics for memristive devices. Retention testing involves measuring the stability of multiple resistive states over time, typically at elevated temperatures to accelerate aging processes [90]. This evaluates the device's ability to maintain the observed memory window (dynamic range of switching) that defines its operational range. Endurance testing quantifies the number of switching cycles a device can undergo before failure, typically performed between consecutive memory states or between the extremes defining the DUT's OFF/ON ratio [90].

For PQD surface morphology studies, these tests can reveal crucial reliability differences. Morphologies that facilitate more uniform switching mechanisms may exhibit superior endurance, while those with optimized interface properties may demonstrate enhanced retention. These metrics must be evaluated under consistent conditions across different PQD samples to enable meaningful comparison.

Table 1: Key Performance Metrics for Memristive Device Benchmarking

Performance Metric	Testing Protocol	Measurement Conditions	Relevance to PQD Surface Morphology
Switching Threshold	Alternating polarity pulses followed by voltage ramps	Quasi-DC regime, pulse width 100μs [90]	Determines operating voltage and power consumption; influenced by surface energy and defect density
HRS/LRS Ratio	I-V characteristics at read voltage (typically 0.4V) [91]	DC voltage sweep, sub-threshold region	Defines memory window; affected by interfacial properties and filament completeness
Endurance	Repetitive switching between resistance states	Pulse trains with verified non-destructive read [90]	Measures cycling reliability; correlated with morphological stability during switching
Retention	Resistance monitoring over time at elevated temperature	Baking at 85-150°C, periodic reading [90]	Evaluates state stability; influenced by surface diffusion barriers and defect migration
Multilevel Capacity	Progressive programming to intermediate states	Incremental pulse amplitude/width sequences [90]	Determines analog behavior potential; affected by switching uniformity and noise characteristics
Switching Speed	Pulse width dependence of switching threshold	Nanosecond to microsecond pulses [90]	Measures operational speed; potentially influenced by surface recombination velocities

Advanced Characterization for Mechanism Elucidation

Conduction Mechanism Analysis

Beyond performance metrics, understanding the fundamental conduction and switching mechanisms is essential for correlating PQD surface morphology with device behavior. Temperature-dependent I-V characterization enables the discrimination between various conduction mechanisms through Arrhenius plots or more sophisticated signature analysis [90]. Different mechanisms exhibit distinct field- and temperature-dependencies, allowing researchers to identify whether transport is dominated by processes such as Ohmic conduction, Schottky emission, Frenkel-Poole emission, or hopping transport [90].

For PQD devices, surface morphology can dramatically influence the dominant conduction mechanism by altering barrier heights at interfaces, changing trap distributions within the material, or modifying the filament formation process. Smooth, well-defined surfaces may promote uniform interface-limited transport, while rough or highly defective surfaces may lead to bulk-limited or localized conduction paths. These differences must be characterized through careful analysis of the temperature dependence of both the high-resistance state (HRS) and low-resistance state (LRS).

Variability and Cycle-to-Cycycle Consistency

Device-to-device and cycle-to-cycle variability represents a significant challenge in memristor technology, particularly for filamentary switching devices. In organic and polymer-based memristors, this variability can arise from inconsistencies in active layer thickness, interface roughness, the stochastic nature of conductive filament formation, and environmental sensitivity of the materials [92]. Similar variability concerns apply to PQD-based devices, where surface morphology may influence the consistency of switching parameters.

Standardized testing must include statistical analysis of switching parameters (SET voltage, RESET voltage, HRS/LRS values) across multiple cycles (typically 100-1000) and across multiple devices from the same fabrication batch. The coefficient of variation (standard deviation normalized by mean) for these parameters provides a quantitative measure of device reproducibility. For PQD research, this analysis can reveal how different surface morphologies influence switching stochasticity, with more uniform morphologies potentially offering reduced variability.

The Scientist's Toolkit: Essential Research Reagents and Materials

The reliable characterization of memristive devices requires specific materials and instrumentation capable of precise electrical measurement and environmental control. The following toolkit details essential components for implementing the standardized testing protocols described in this review, with particular attention to requirements for PQD surface morphology studies.

Table 2: Essential Research Toolkit for Memristor Characterization

Tool/Reagent	Function/Application	Technical Specifications	PQD Morphology Considerations
Semiconductor Parameter Analyzer	Precise I-V characterization, pulse generation, and transient measurement	Source-Measure Units (SMUs) with pA resolution, arbitrary waveform generation [90]	Must accommodate low-current measurements for initial forming and subtle switching events
Probe Station with Thermal Chuck	Environmental control and device contacting	Temperature range -60°C to 300°C, shielded probes, micromanipulators [90]	Enables temperature-dependent studies of conduction mechanisms influenced by surface states
Shielded Enclosures/Cables	Noise reduction for sensitive measurements	Triaxial cables, electromagnetic shielding	Critical for accurate characterization of high-resistance states in ultrathin PQD films
Arbitrary Waveform Generator	Complex pulse sequences for dynamic testing	Nanosecond rise times, multiple synchronized channels	Allows testing of spike-timing-dependent plasticity (STDP) for neuromorphic applications
Material Analysis Suite	PQD surface morphology characterization	AFM, SEM, TEM, XPS for structural and chemical analysis [92] [93]	Correlates electrical performance with morphological features (roughness, grain structure)
Standardized Test Structures	Controlled device fabrication for comparison	Metal-Insulator-Metal (MIM) crossbars with defined electrode areas [91]	Enables isolation of surface morphology effects from electrode geometry contributions

Comparative Performance Analysis of Memristor Technologies

To contextualize the performance of PQD-based memristors, comparative analysis against established memristor technologies is essential. The following table summarizes key performance metrics for prominent memristor material systems, providing benchmark values for evaluating new PQD developments.

Table 3: Performance Comparison of Memristor Material Systems

Material System	Switching Mechanism	ON/OFF Ratio	Endurance (Cycles)	Retention	Power Consumption	Reference
HfO₂	Filamentary (Anion)	>10¹⁰	>10¹⁰	>10 years @ 85°C	Moderate	[92] [93]
TaOₓ	Filamentary (Anion)	10-100	>10¹²	>10 years @ 85°C	Low (~μW)	[93]
TiO₂	Filamentary (Anion)	10³-10⁵	10⁶-10⁹	>10 years @ 85°C	Moderate	[92] [93]
PCMO	Interface (Anion)	10-100	10⁶-10⁹	>10 years @ 85°C	Low	[94]
ZnO	Filamentary (Anion)	~1.8-2.5	10³-2×10⁴	Limited data	Low	[91]
2D Materials	Various	10³-10⁶	10⁴-10⁸	Varies significantly	Ultra-low	[95]
Polymers/Organics	Filamentary (Cation)	10²-10⁵	10²-10⁵	Limited stability	Low	[92]

When evaluating PQD-based memristors against these benchmarks, researchers should pay particular attention to the switching mechanism, as this fundamentally influences expected performance characteristics. Filamentary systems typically offer higher ON/OFF ratios but may suffer from greater variability, while interface-based switching often provides better uniformity but more modest resistance windows. PQD surface morphology can influence both the switching mechanism and resulting performance metrics, enabling material-by-design approaches for optimizing specific device characteristics.

Implementation Considerations for PQD Morphology Studies

Electroforming and Pre-conditioning Protocols

Many memristive devices require an initial electroforming step to activate resistive switching behavior. This process fundamentally affects the physical characteristics of the device on both structural and interfacial levels [90]. For PQD-based devices, the electroforming process may be particularly sensitive to surface morphology due to its influence on electric field distribution and ion migration pathways.

Standardized protocols must specify whether electroforming is required and document the precise conditions (voltage ramp rate, current compliance, environmental conditions) used. For comparative studies of different PQD surface morphologies, consistent electroforming parameters are essential, though optimal conditions may vary between morphologies. Some advanced memristor technologies are electroforming-free, which represents a significant advantage for device reproducibility and reliability [90].

Environmental and Aging Considerations

Memristive device performance can be significantly influenced by environmental factors, particularly oxygen and moisture exposure. Organic and polymer-based memristors are known to be prone to environmental and thermal instability [92], while oxide-based devices generally offer superior stability. For PQD devices, surface morphology may dramatically influence environmental stability through effects on surface area, defect density, and encapsulation requirements.

Standardized testing should include controlled environmental conditions (temperature, humidity, oxygen levels) and may incorporate accelerated aging studies to evaluate long-term stability. For PQD morphology studies, these tests can reveal how surface structure influences degradation mechanisms such as oxidation, ion migration, or phase separation.

The establishment of standardized testing protocols for memristive devices represents an essential step toward meaningful cross-comparison between different technologies, particularly for emerging fields like PQD surface morphology research. The methodology presented here—encompassing functionality testing, performance benchmarking, and mechanism elucidation—provides a comprehensive framework for objective device evaluation. By adopting such standardized approaches, the research community can accelerate the development of optimized memristive devices through reliable comparison of different material approaches, including the systematic evaluation of how PQD surface morphology influences key device characteristics.

As memristor technologies continue to mature toward commercial application, standardized benchmarking will become increasingly important for identifying the most promising material systems and device architectures. The methodology outlined here provides a foundation for such standardization, offering a technology-agnostic pathway for fair cross-study comparison that can advance the entire field of memristive electronics.

Conclusion

The surface morphology of polymer-quantum dot composites is a paramount factor dictating their memristive performance, directly influencing charge trapping efficiency, switching uniformity, and operational stability. This review establishes that optimal QD concentration and dispersion are critical for achieving high ON/OFF ratios, long retention, and robust endurance. While challenges in variability and degradation persist, strategic material selection and interface engineering offer clear pathways for optimization. The future of PQD-based memristors is exceptionally promising, with significant implications for the development of energy-efficient neuromorphic computing systems, flexible wearable sensors, and advanced biomedical platforms that seamlessly integrate sensing, memory, and processing. Future research should focus on developing non-toxic QD alternatives, standardizing characterization methods, and exploring 3D integration to fully realize the potential of this technology.

References

• 1. Memristor-Based Artificial Neural Networks for Hardware ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12231232/]
• 2. Emerging Memristive Devices for Brain-Inspired ... [https://www.frontiersin.org/journals/nanotechnology/articles/10.3389/fnano.2022.940825/full]
• 3. Analog Switching in Hexagonal Boron Nitride Memristors ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12366253/]
• 4. Ultra-fast switching memristors based on two-dimensional ... [https://www.nature.com/articles/s41467-024-46372-y]
• 5. Sub-10 nm Ta Channel Responsible for Superior ... [https://www.nature.com/articles/srep28525]
• 6. Experimental demonstration of in-memory computing using ... [https://pubs.rsc.org/en/content/articlehtml/2025/tc/d5tc01687h]
• 7. Transparent Zinc Oxide Memristor Structures: Magnetron ... [https://www.mdpi.com/2079-4991/14/23/1901]
• 8. Memristive switching in the surface of a charge–density– ... [https://www.sciencedirect.com/science/article/abs/pii/S209592732400344X]
• 9. Quantum dot-based non-volatile memory - PubMed Central [https://pmc.ncbi.nlm.nih.gov/articles/PMC12053827/]
• 10. Solution-Processed Memristor Devices Using a Colloidal ... [https://www.mdpi.com/2076-3417/11/11/5020]
• 11. Charge trapped CdS quantum dot embedded polymer ... [https://pubs.rsc.org/en/content/articlelanding/2023/cp/d2cp05014e]
• 12. Direct measurement of charge trap depth in polymer ... [https://ui.adsabs.harvard.edu/abs/2023JPhD...56F5301S/abstract]
• 13. Ultrasensitive dopamine detection using CsPbBr3-PQD-COF ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12136283/]
• 14. Time-/environment-dependent multilevel optical encryption ... [https://www.sciencedirect.com/science/article/pii/S2589965125000261]
• 15. Cadmium Selenide Quantum Dots (CdSe/ZnS) [https://nanographenex.com/blog/cadmium-selenide-quantum-dots-cdse-zns-properties-applications-and-research-trends/?srsltid=AfmBOort1-i3Lm4EJN-8feCe5nQ3pGOnI84E2AA1IKIU3fa6989X91JK]
• 16. Surface chemistry studies of (CdSe)ZnS quantum dots at ... [https://pubmed.ncbi.nlm.nih.gov/15924465/]
• 17. ZnSeTe/CdZnSe-based type-II core–shell quantum dots ... [https://www.sciopen.com/article/10.26599/NR.2025.94907267]
• 18. Prospects of Nanoscience with Nanocrystals: 2025 Edition [https://pmc.ncbi.nlm.nih.gov/articles/PMC12445017/]
• 19. Investigation on structural and opto-electronic properties of ... [https://www.sciencedirect.com/science/article/pii/S2238785421009017]
• 20. Improved Characteristics of CdSe/CdS/ZnS Core-Shell ... [https://www.mdpi.com/2079-4991/12/6/909]
• 21. Interfacial and bulk switching MoS 2 memristors for an all- ... [https://arxiv.org/html/2511.16557]
• 22. Trap assisted charge transport and Schottky like ... [https://www.sciencedirect.com/science/article/abs/pii/S0169433222009199]
• 23. Direct observation of trap-assisted recombination in ... [https://www.nature.com/articles/s41467-021-23870-x]
• 24. Fostering the Dense Packing of Halide Perovskite ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10604077/]
• 25. Enriching conductive capping by alkaline treatment of ... [https://www.nature.com/articles/s41467-025-63618-5]
• 26. Nanoscale memristive radiofrequency switches [https://www.nature.com/articles/ncomms8519]
• 27. Low‐Power Memristive Logic Device Enabled by ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8336485/]
• 28. High-performance self-rectifying memristor array based on ... [https://www.sciopen.com/article/10.26599/NR.2025.94907085]
• 29. Single step wet-chemical deposition of cesium doped β [https://www.sciencedirect.com/science/article/pii/S2773064622000093]
• 30. Multibit Ferroelectric Memcapacitor for Non-volatile ... [https://arxiv.org/html/2511.09838v1]
• 31. Quantum dot-based non-volatile memory - RSC Publishing [https://pubs.rsc.org/en/content/articlehtml/2025/ra/d4ra08307e]
• 32. Computational Study of Simultaneous Memristive Effect ... [https://link.springer.com/article/10.1007/s12668-025-01849-y]
• 33. Thin Film Deposition | Comparing Coating Methods [https://www.ossila.com/pages/solution-processing-techniques-comparison]
• 34. Scalable deposition techniques for large-area perovskite ... [https://www.sciencedirect.com/science/article/pii/S221128552400065X]
• 35. Thin Film Deposition Techniques: A Comprehensive Review [https://article.innovationforever.com/JMN/20240268.html]
• 36. Biomaterials Based on Organic Polymers and Layered ... [https://www.mdpi.com/1999-4923/15/2/413]
• 37. Optimization strategy of the emerging memristors [https://www.sciencedirect.com/science/article/pii/S2589004224025525]
• 38. Flexible and transparent memristive synapse based on ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9419774/]
• 39. Highly Stable and Flexible Memristive Devices Based on ... [https://www.nature.com/articles/s41598-020-62721-5]
• 40. Structural and electrical studies of PMMA and PVdF based ... [https://link.springer.com/article/10.1007/s42452-020-03868-8]
• 41. Improving the current stability of perovskite quantum dot ... [https://pubs.rsc.org/en/content/articlehtml/2025/tc/d5tc01629k]
• 42. A comparative study of PMMA/PEG polymer ... [https://www.nature.com/articles/s41598-024-63176-8]
• 43. Cobalt-doped zinc sulfide quantum dot thin films for next- ... [https://www.sciencedirect.com/science/article/abs/pii/S0038110125000917]
• 44. Harnessing light in tandem: advanced erbium-polyaniline QD ... [https://pubs.rsc.org/en/content/articlehtml/2025/nr/d5nr02298c]
• 45. Advances of Emerging Memristors for In-Memory Computing ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12508526/]
• 46. Advanced memristive architectures based on ... [https://www.frontiersin.org/journals/nanotechnology/articles/10.3389/fnano.2025.1558743/full]
• 47. Thermal crosstalk in 3-dimensional RRAM crossbar array [https://www.nature.com/articles/srep13504]
• 48. Fabrication of High-Quality MoS2/Graphene Lateral ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12388528/]
• 49. Lead-free halide perovskite memristors for scalable crossbar ... [https://nanoconvergencejournal.springeropen.com/articles/10.1186/s40580-025-00507-z]
• 50. A Full Spectrum of 3D Ferroelectric Memory Architectures ... [https://arxiv.org/html/2504.09713v1]
• 51. Self-limited single nanowire systems combining all-in-one ... [https://www.nature.com/articles/s41467-018-07330-7]
• 52. Low-dimensional Silicon for Neuromorphic Computing [https://www.sciencedirect.com/science/article/pii/S2709472325000437]
• 53. Paper-based flexible electronic devices: processing, ... [https://www.nature.com/articles/s41528-025-00446-z]
• 54. All-optical voltage interrogation for probing synaptic ... [https://www.nature.com/articles/s41467-025-63867-4]
• 55. High-confidence and high-throughput quantification of ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11805640/]
• 56. Implementation of device-to-device and cycle-to-cycle ... [https://www.sciencedirect.com/science/article/pii/S0038110122000934]
• 57. Performance and variability analysis of ALD-grown wafer ... [https://www.frontiersin.org/journals/nanotechnology/articles/10.3389/fnano.2025.1621554/full]
• 58. Resistive Switching Devices for Neuromorphic Computing [https://pmc.ncbi.nlm.nih.gov/articles/PMC10976006/]
• 59. Subfilamentary Networks Cause Cycle-to-Cycle Variability ... [https://pubmed.ncbi.nlm.nih.gov/28661649/]
• 60. Effect of Surface Ligand Modification on the Optical ... [https://www.sciencedirect.com/science/article/abs/pii/S2468023025023211]
• 61. Understanding memristive switching via in... | Nature Communications [https://www.nature.com/articles/s41467-019-11411-6?error=cookies_not_supported&code=684ca35c-bc5b-4fd2-88ff-921073ad44b3]
• 62. A double barrier memristive device | Scientific Reports [https://www.nature.com/articles/srep13753]
• 63. State Characterisation of Self-Directed Channel Memristive ... [https://arxiv.org/html/2505.15757v1]
• 64. High-Performance Perovskite Quantum Dot Solar Cells Enabled by... [https://link.springer.com/article/10.1007/s40820-022-00946-x]
• 65. Self-rectifying memristors with high rectification ratio for ... [https://www.nature.com/articles/s41467-025-60970-4]
• 66. Large-scale crossbar arrays based on three-terminal MoS2 ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12569200/]
• 67. Study on sneak path effect in self-rectifying crossbar arrays ... [https://www.frontiersin.org/journals/electronic-materials/articles/10.3389/femat.2022.988785/full]
• 68. 2D materials-memristive devices nexus: From status quo to ... [https://www.sciencedirect.com/science/article/pii/S0079642525000465]
• 69. Interface–bulk defect engineering via dual ligands for ... [https://www.sciencedirect.com/science/article/abs/pii/S1385894725094288]
• 70. Inorganic Halide Perovskite Quantum Dots for Memristors [https://link.springer.com/article/10.1007/s13391-025-00560-0]
• 71. Quantum Dots for Resistive Switching Memory and Artificial ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11478683/]
• 72. Rational synthesis of core/shell heterostructured quantum ... [https://www.sciencedirect.com/science/article/abs/pii/S2211285524007833]
• 73. A quantum resistance memristor for an intrinsically ... [https://www.nature.com/articles/s41565-025-02037-5]
• 74. Ultrathin Soft Wearable Sensor Materials and Structures [https://pmc.ncbi.nlm.nih.gov/articles/PMC12598712/]
• 75. Soft and stretchable protective substrates for wearable ... [https://www.sciencedirect.com/science/article/abs/pii/S0927775724006708]
• 76. Hydrogel Microspheres as Versatile Platforms for Biomedical ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12508632/]
• 77. hydroxyapatite/ graphene oxide scaffold: 3D printing [https://www.sciencedirect.com/science/article/pii/S2238785425012177]
• 78. Advances of RRAM Devices: Resistive Mechanisms... Switching [https://pubmed.ncbi.nlm.nih.gov/32717952/]
• 79. for... | Nature Reviews Resistive switching materials Materials [https://www.nature.com/articles/s41578-019-0159-3?error=cookies_not_supported]
• 80. : From Fundamentals of... | Semantic Scholar Resistive Switching [https://www.semanticscholar.org/paper/Resistive-Switching%3A-From-Fundamentals-of-Nanoionic-Ielmini-Waser/b5873e7f44ee698e405bbba4732bc480eb8d321a]
• 81. Exploring Area-Dependent Pr0.7Ca0.3MnO3-Based ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8282906/]
• 82. Inducing memristive behavior to MoSe2/graphene bilayer ... [https://www.nature.com/articles/s41598-025-14798-z]
• 83. Relaxation and Wien-Bridge Cases [https://www.mdpi.com/2673-8716/5/4/43]
• 84. Quantitative 3D-KPFM imaging with simultaneous ... [https://pubmed.ncbi.nlm.nih.gov/25851168/]
• 85. Quantum Dots for Resistive Switching Memory and Artificial ... [https://www.mdpi.com/2079-4991/14/19/1575]
• 86. Contrasting analog and digital resistive switching memory ... [https://pubs.rsc.org/en/content/articlelanding/2023/tc/d3tc00090g]
• 87. Flexible and transparent memristive synapse based on ... [https://www.sciencedirect.com/science/article/pii/S2516023023019603]
• 88. Neuron-inspired CsPbBr3/PDMS nanospheres for multi ... [https://www.nature.com/articles/s41377-025-01742-z]
• 89. Research on the Physical Properties and Internal Structure ... [https://www.mdpi.com/2073-4360/17/21/2939]
• 90. An Electrical Characterisation Methodology for ... [https://www.nature.com/articles/s41598-019-55322-4]
• 91. Transparent Zinc Oxide Memristor Structures: Magnetron ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11643567/]
• 92. The critical role of materials and device geometry on ... [https://www.sciencedirect.com/science/article/abs/pii/S2542529325000719]
• 93. A review of memristor: material and structure design, device ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9980037/]
• 94. Pr1–x Ca x MnO3‑Based Memristive Heterostructures [https://pmc.ncbi.nlm.nih.gov/articles/PMC12257465/]
• 95. Advancements in 2D layered material memristors [https://www.nature.com/articles/s41699-024-00522-4]