Availability and production capabilities in diamond are rapidly expanding. To enable nextgeneration diamond-based solutions, particularly in the quantum and semiconductor realms, WD Advanced Materials (“WDAM”), in cooperation with key industry collaborators, has developed new processes for the synthesis of large-scale single-crystal diamond wafers for electronic applications. Through this talk, led by Chief Technology Officer John Ciraldo, WDAM will share an overview of recent product development breakthroughs supporting these emerging technologies, including R&D results, output from the D2 SCALE under DARPA’s LADDIS initiative, and the latest on a new U.S.-based consortium, created to supply synchrotron-grade diamond to the national market through the DOE ARDAP program. WDAM will share third-party characterization data demonstrating exceptional material quality characteristics coupled with large substrate sizes, including anonymized comparisons to other commercially available products. In furtherance of the team’s form factor expansion efforts, WDAM will also present 13mm+ wafers with full x-ray mapping, highlighting narrow rocking curves of about 30 arcseconds, ultra-high doping concentrations, and evidence of strong electronic transport properties. The discussion will conclude with a summary of new material science process and technology solutions developed at WDAM, including defect mitigation; selective patterning and deposition; heterostacks (e.g. PN/PIN junctions); and continued scale in both material size and quality.

The road to miniaturization in semiconductor microelectronic technology has led to an ongoing demand for identification of suitable approaches for device performance enhancements and raised requirements of new materials and device structures [1].

Here, we report the electronic device performance improvement opportunity of silicon nanostructures without direct channel doping by adding Al acceptor states into a thin SiO₂ layer surrounding the silicon [2]. This novel concept is analogous to modulation doping in III-V semiconductors. The manifestation of the dielectric by incorporation of trivalent Al impurities gives the possibility for electrons from silicon to tunnel to the acceptor states in SiO₂ yielding silicon with holes as majority carriers. Our method removes the disadvantages of impurity doping, namely a required thermal ionization, inelastic impurity scattering, problematic dopant activation vs. out-diffusion and statistical fluctuations of dopant density [3-4]. The SiO₂ doping also facilitates interface passivation by discharging the dangling bond defects at Si/SiO₂ interface [5]. To this end, the doping approach is demonstrated on silicon channels fabricated on a silicon-on-insulator (SOI) substrate with nickel silicide contacts. The same layers sequence is extended to a transfer length measurement (TLM) structure.

It was experimentally observed that the energetic state of Al-acceptors in SiO₂ formed by atomic layer deposition of Al₂O₃ on thin thermally grown SiO₂ induces a corresponding shift in the Fermi level of silicon towards the valence band. A variation in the density of hole carriers in the silicon channel is regulated by the variation of number of ALD Al₂O₃ cycles [6] which governs both the silicon resistivity and the specific contact resistivity. The effect of addition of a monolayer of ALD Al₂O₃ is reflected in the reduction of silicon resistivity to a value as low as 0.04-0.06 Ω●cm and a contact resistivity as low as 3.5E-7 Ω●cm² [7].

A single segment of the fabricated TLM structure can also perform equivalently as a transistor with the body contact operated as a back gate. The performance of the transistor is improved by an enhanced hole carrier transport with an I_ON/I_OFF ratio of up to 10⁶ at a drain voltage of -1V. The NiSi source/drain contacts placed directly within the SOI region also undergoes a transition from Schottky to low-resistance contact as an effect of the barrier height reduction [8]. The drawbacks of conventional impurity doped devices such as dopant scattering, fluctuating device performance due to dopant diffusion, and even requirements of lightly or heavily doped source/drain regions can be circumvented by the new method. The aforesaid impurity free doping of the channel would be ideal for a high performance, high mobility device and enables new advancements encompassing areas of low-temperature physics and quantum technology.

[1] Wong. H, and Iwai. H, Physics World 18.9 (2005): 40.
[2] König. D, et al., Scientific Reports 7.1 (2017): 46703.
[3] Norris. D. J., et al., Science 319.5871, 1776 (2008).
[4] Nagarajan. S, et al., Solid-State Electronics 208 (2023): 108739.
[5] Hiller. D, et al., J. App. Phys. 125.1 (2019): 015301.
[6] Ratschinski. I, et al., physica status solidi (a) (2023): 2300068.
[7] Nagarajan. S, et al., Advanced Materials Interfaces, Oct 2023 (Accepted).
[8] Nagarajan. S, et al., Device Research Conference. IEEE, 2023.

Electrostatic energy storage based on dielectric capacitors has garnered significant interest due to its fast charge-discharge speeds and high power density relative to electrochemical energy storage, but applications have been limited due to large leakage currents, relatively low energy storage density, and the presence of lead in materials^[1,2]. NaNbO₃ is a lead-free alternative which has received much research attention in the last decade but work has primarily focused on its bulk form, with NaNbO₃-based systems displaying both antiferroelectric and relaxor ferroelectric behaviors at room temperature with promising energy storage properties^[2,3]. Using pulsed laser deposition and selective etching, we have synthesized metal-insulator-metal heterostructures of La_0.7Sr_0.3MnO₃ and NaNbO₃ with high crystalline quality and coherent epitaxial strain across a range of thicknesses. Both strain and thickness have previously been noted to greatly affect the properties of NaNbO₃ due to the close proximity of its antiferroelectric and ferroelectric ground states^[4,5]. Through careful optimization of the growth conditions, we achieve minimal leakage current even under large external electric fields and can observe relaxor-like ferroelectric behavior in both our thinnest samples (below 50 nm of NaNbO₃) and our thickest samples (above 150 nm). This allows us to achieve improvements in both recoverable energy-storage density W_rec and energy efficiency η compared to previous NaNbO₃ thin films^[6,7]. This work demonstrates the potential of NaNbO₃ for dielectric energy storage and the advantages of epitaxial, single-crystal films for electrical characterization.

[1] H. Pan et al., Science 374, 100-104 (2021).
[2] M.-H. Zhang et al., Nat. Comm. 14, 1525 (2023).
[3] N. Luo et al., Nat. Comm. 14, 1776 (2023).
[4] T. Schneider et al., ACS Omega 8, 23587-23595 (2023).
[5] R. Xu et al., Adv. Mater. 35, 2210562 (2023).
[6] T. Shiraishi et al., J. Appl. Phys. 128, 044102 (2020).
[7] H. Dong et al., Ceram. Inter. 48, 16215-16220 (2022).

The data retention is one major issue of requirements for non-volatile memory (NVM) application, i.e. non-zero remnant polarization (P_r) of Ferroelectric random-access-memory (FeRAM) application at standby (bias-free) [1][2]. Recently, antiferroelectric HfZrO₂ (AFE-HZO) under built-in bias of fixed charge (Q_f) and/or interface dipole by Al₂O₃ interlayer leads applicable P_r [3][4]. Furthermore, AFE-HZO exhibits intrinsic capability of high-speed response and long endurance [5]. However, the validation of bilayer of AFE and dielectric (DE) is lack for further discussion. In this work, the technique of First-Order Reversal Curve (FORC) [6][7] will be employed to reveal the built-in bias based coercive field (E_c) distribution for AFE/DE.
The FORC diagram can determine the distribution of the E_C by multiple P-E loops measurement. The FORC distribution function ρ (E_r, E) shows the concentration of E_C, and the procedure can be achieved as following. The initial field of the waveform is the saturation field (E_sat), sweeping from the E_sat to the reversal field (E_r) and then back to the E_sat. Then the step is repeated and E_r range is from E_sat reversing gradually to -E_sat. The metal/ferroelectric/metal (MFM), metal/antiferroelectric/metal (MAFM) and metal/antiferroelectric/insulator/metal (MAFIM) were prepared for FORC analysis. The MFM shows typical FE-behavior and 2P_r > 60 mC/cm². For MAFM, the 2P_r almost zero due to polarization compensation at 0 MV/cm. For MAFIM, the “FE-like” characteristic is presented due to built-in bias. The origins of the built-in bias with Al₂O₃ are unbalance of surface energy for interface dipole and insufficient bonding for positive or negative fixed charge [8][9]. The single E_C distribution of MFM by FORC diagram indicates the typical FE behavior and it concentrates at ~1.2 MV/cm. For MAFM, there are two Ec distributions at ~ -0.5 MV/cm and 2.5 MV/cm, and this is attributed to the multi-peak of Ec for AFE. For MAFIM, the single Ec distribution is occurred at ~1.5-2 MV/cm. The FORC methodology is employed to validate the built-in bias existence and non-zero Pr for AFE/DE system.
The authors are grateful for the funding support by Ministry of Science and Technology (NSTC) 112-2218-E-A49-013-MBK, 112-2221-E-002-252-MY3, 111-2221-E-002-203-MY3, and process supported by TSRI & NFC, Taiwan.

References:
[1]. D. Takashima, "Overview of FeRAMs: Trends and perspectives," 2011 11th Annual Non-Volatile Memory Technology Symposium Proceeding, Shanghai, China, 2011, pp. 1-6.
[2]. J. Hur et al., "A Technology Path for Scaling Embedded FeRAM to 28nm with 2T1C Structure," 2021 IEEE International Memory Workshop (IMW), Dresden, Germany, 2021, pp. 1-4.
[3]. K.-Y. Hsiang et al., “Bilayer-based Antiferroelectric HfZrO2 Tunneling Junction with High Tunneling Electroresistance and Multilevel Nonvolatile Memory,” IEEE Electron Device Letters, vol. 42, no. 10, pp. 1464-1467, 2021.
[4]. C.-Y. Liao et al., “Experimental Insights of Reverse Switching Charge for Antiferroelectric Hf0.1Zr0.9O2,” IEEE Electron Device Letters, vol. 43, no. 9, pp. 1559-1562, 2022.
[5]. M. Pešić et al., “Comparative Study of Reliability of Ferroelectric and Anti-Ferroelectric Memories,” IEEE Transactions on Device and Materials Reliability, vol. 18, no. 2, pp. 154-162, 2018.
[6]. Tony Schenk et al.,“Complex Internal Bias Fields in Ferroelectric Hafnium Oxide,” ACS Appl. Mater. Interfaces 2015, 7, 20224−20233.
[7]. Laurentiu Stoleriu et al., “Analysis of switching properties of porous ferroelectric ceramics by means of first-order reversal curve diagrams,” Phys. Rev. B 74, 174107, 2006.
[8] D. K. Simon et al., "On the Control of the Fixed Charge Densities in Al₂O₃ - Based Silicon Surface Passivation Schemes, ” ACS Appl. Mater. Interfaces, vol. 7, no. 51, pp. 28215–28222, 2015.
[9] K. Kitaa, and A. Toriumi, “Origin of electric dipoles formed at high-k/SiO₂ interface, ” Appl. Phys. Lett., 94, 132902, 2009.

When BMO-doped REBCO superconducting thin films (BMO+REBCO films) are prepared by vapor-phase-epitaxy (VPE) such as PLD and MOCVD, BMO self-organizes into nanorods and/or nanoparticles in the REBCO matrix. On the other hand, only incoherent BMO nanoparticles are observed in the solid phase growth method such as MOD. These experimental results suggest that the kinetics of the raw material particles at the surface of the thin film crystal growth contribute significantly to the self-organization of BMO. In fact, it is known that BMO exhibits various nanostructures depending on the growth temperature (T_G), the volume fraction of BMO added (V_BMO), and the deposition rate (DR) in the VPE. Therefore, we have developed a BMO+REBCO thin film growth simulation considering the kinetics using the Monte Carlo (MC) method to investigate the effect of deposition conditions on the nanostructure and the formation mechanism of nanorods.
As a result, the following trends were obtained.
(1) When T_G is high and DR is low, nanorods are formed perpendicular and linear to the substrate surface. In addition, the diameter of the nanorods becomes thicker.
(2) At a lower T_G and higher DR than (1), the nanorods are inclined and their diameters are narrower and their number density is higher.
(3) When DR is high enough, nanoparticles consisting of short nanorods are observed.
(4) Even if DR is high enough, linear nanorods can be obtained if T_G and V_BMO are sufficiently high.
The above trends are qualitatively in good agreement with experimental results and reported cases.
The time evolution of nanostructure formation can also be observed in the MC simulations. This indicates that nanorods are formed in the following steps. First, crystal nuclei of REBCO and BMO are generated on the substrate surface, from which crystal growth proceeds. Since the volume fraction of BMO is small, the BMO islands are eventually surrounded by the REBCO layer, and BMO can only grow in the direction perpendicular to the substrate surface. This process is repeated, consequently, BMO nanorods are formed. On the other hand, at low T_{<span style="font-size:10.8333px">G</span>} and high DR, more BMO crystal nuclei are generated, which slows down the growth rate of BMO islands in the vertical direction, resulting in tilted or shortened nanorods. In other words, the competition between REBCO and BMO growth rates results in the formation of various nanostructures.
The above are mainly MC simulations in the VPE. On the other hand, the VLS growth method, which is a thin film growth method via thin liquid layer on a films surface, has the advantage that high-quality crystals can be fabricated even at high DR. However, it is difficult to control the BMO nanostructure. This is because the kinetics of the raw material particles is different from that of the VPE due to the presence of the thin liquid layer. Therefore, MC simulations concerning to the VLS growth method were developed and compared with the VPE for the formation of BMO nanostructures. MC simulations were also performed by intentionally adding screw dislocations to obtain a more detailed picture of the crystal growth environment.
In this presentation, we will discuss the growth conditions, nanostructure formation, and thin film crystal growth environment.

Acknowledgment
This work was partly supported by JST, CREST Grant Number JPMJCR2336, Japan.

The zero electrical resistivity of superconducting materials makes them ideal for long-distance electric-current transmission. In particular, high-temperature superconducting (HTS) materials require only moderate cooling power, and a few trials have successfully demonstrated their capability for power transmission. Utilizing a long extended HTS cable to generate an external magnetic field is an emerging application with unique challenges. Our study introduces a conceptual design of a flat seabed coil, proportional to a ship’s size, for the purpose of ship magnetic deperming. Magnetic deperming involves imposing a magnetic field on the ship’s hull to saturate the magnetization of its steel components. The required field is estimated to be 2,400 A/m, achievable by the flat seabed coils in shallow water. The deperming process requires an alternating magnetic field with decreasing peak intensity. We have designed the coil for the largest destroyer and submarine ships of the Japanese Maritime Self-Defense Force. The parameters required for the coil design were obtained from publicly-available information. Our design features three racetrack-shaped coils, each carrying a maximum current of 200 kA, spanning a total length of 1,200 m for a single coil, set flat at a depth of 12 m, and cooled by liquid hydrogen, using the currently available HTS materials and technologies.

Fluorescent nanodiamond (FND) has recently been regarded as a superior alternative reporter for lateral flow immunoassays (LFIA). The negatively charged nitrogen-vacancy (NV^–) center in FND is a point defect with unique magneto-optical properties, giving outstanding characteristics to detect biomarkers of diseases. Alzheimer’s disease (AD) is the most common form of dementia, characterized by decreased memory, thinking, cognition, behavior, personality, and ability to conduct everyday duties, which might result in complete brain failure and, eventually, death. The two most commonly used detection techniques to detect changes in the brain caused by AD are brain imaging (CT, MRI, PET) and lumbar puncture. However, while brain imaging is costly and time-consuming, lumbar puncture is controversial since 30% of patients get severe headaches followed by nausea and vomiting lasting more than a week. To this end, we have developed a Spin-Enhanced Lateral Flow Immuno-Assay (SELFIA) for non-invasive AD diagnostics utilizing the spin properties of the NV^– centers in FND to achieve background-free ultrasensitive detection. In addition, to enhance sensitivity and specificity, we employed a sandwich assay of SELFIA to further noncovalently conjugate FND to anti-pTau antibodies toward detecting pTau protein (a critical AD biomarker involved in AD pathology), while the capture antibody immobilized on the membrane. Note that we have a comparable result with Enzyme-Linked Immunosorbent Assay (ELISA), as our FND-based SELFIA only requires approximately 30 minutes to reach a detection limit of 7 pg/mL for pTau (the sensitivity/specificity threshold is 15 pg/mL in plasma). Hopefully, this study will contribute to developing a safe, simple, and accurate AD detection platform that can identify this illness in its earliest stages and might also be applied to other diseases in precision health.

Intracellular temperature fluctuations are thought to be closely related to higher order cellular phenomena. We have developed a method for imaging the temperature distribution in single living cells using a fluorescent polymer thermometer and fluorescence lifetime imaging microscopy to study the effects of intracellular temperature on cellular physiological activity. Previous experiments using the developed method have shown that during the G1 phase, the temperature of the nucleus is approximately 1°C higher than the temperature of the cytoplasm, as well as some mitochondria and centrosomes. The mechanisms that maintain this intracellular temperature heterogeneity and its physiological significance are not well understood. Recently, however, the mechanisms of physiological phenomena caused by fluctuations in intracellular temperature have begun to be elucidated. We hypothesized that intracellularly generated heat is not only a product of reactions, but also drives cellular responses. To test this hypothesis, we focused on cell differentiation, in which cells dramatically change their protein expression patterns. Using the neuronal model cell PC12, we investigated the relationship between neuronal differentiation and intracellular temperature. The results showed that intracellular temperature is involved in neuronal differentiation and associated neurite outgrowth. In addition to the aforementioned fluorescent polymer temperature sensor and fluorescence lifetime imaging microscopy, we also measured temperature using fluorescent nanodiamond particles containing negatively charged nitrogen vacancy centers as temperature sensors. We have previously shown that fluorescent nanodiamonds are able to measure temperature independent of external environmental influences such as pH, salt, and viscosity. To measure intracellular thermal conductivity, we also fabricated nanoparticles coated with polydopamine, which generates heat when exposed to light, around fluorescent nanodiamonds. Although the intracellular thermal conductivity has been assumed to be the same as the thermal conductivity of water, whether this is true or not was verified by actually measuring the intracellular thermal conductivity using the fabricated hybrid particles of polydopamine and fluorescent nanodiamonds. The results showed that, although the data were very uneven, on average the thermal conductivity inside the cell was several times smaller than the conductivity of water. In the present study, it is clear from microscopic observations that the hybrid particles are present inside the cell, but it is not known where in the cell they are localized. The variability of the data suggests variations in thermal conductivity depending on the location within the cell. Future work will use techniques to chemically modify the surface of the polydopamine-fluorescent nanodiamond hybrid particles, such as localizing the particles in mitochondria and nuclei, to measure thermal conductivity at different local positions in the cell to clarify what exactly produces such low thermal conductivity.

Data-driven material science has the potential to transform the way we design and develop materials. This emerging field represents a significant departure from traditional trial-and-error methods and empirical approaches that have characterized materials science for decades. Experiments in this area involve a multidimensional parameter space, making analysis challenging and time-consuming. Finding predictive empirical relations that allow for precise control over various aspects of the synthesis process has posed a challenge to human cognitive abilities alone. This becomes even more complex when combining datasets from different labs or from different scientists due to the lack of established standards for data models and methods to capture the large number of experimental details, including elaborate workflows and a large diversity of instruments for characterization. It requires a radical shift in how information is handled and research is performed. Experiment data must be complemented with its rich-metadata context, also covering ontologies and workflows according to the FAIR (findable, accessible, interoperable, reusable) principles [1].
Opening new perspectives towards finding structure, correlations, and novel information with data-analysis and AI tools, therefore, is intimately connected to the challenge of integrating this information into a FAIR infrastructure [2].

Here we present growth optimization of β-Ga₂O₃ thin films by metalorganic vapor phase epitaxy (MOVPE) on (1 0 0) β-Ga₂O₃ semi-insulating substrates which is a successful example of applying ML/AI approaches in materials synthesis. Improving the efficiency of this synthesis is highly relevant due to the wide range of electronic device applications based on Ga₂O₃. Applying AI modeling on the experimental data related to thin film growth, we effectively improved the growth rate and prediction of doping level in MOVPE-grown Si-doped β-Ga₂O₃ [3,4].

We will present the implementation of this use case in NOMAD (nomad-lab.eu) [5] and discuss the functionalities developed to digitize the entire data lifecycle in crystal growth and epitaxy, with the ultimate goal of FAIR data for materials growth. We developed and deployed Electronic Laboratory Notebooks (ELN) to document all relevant synthesis procedures. We will show how structured data opens up the potential to create tools that facilitate and automate data management, and the sustainable application of AI-based analytics, dedicated to process optimization in synthesis.

[1] Wilkinson, M., et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016; 3, 160018.
[2] Scheffler, M., et al. FAIR data enabling new horizons for materials research. Nature. 2022; 604, 635-642.
[3] Chou, T.-S. et al. Toward Precise n-Type Doping Control in MOVPE-Grown β-Ga2O3 Thin Films by Deep-Learning Approach. Crystals. 2022; 12(1), 8.
[4] Chou, T.-S. et al. Machine learning supported analysis of MOVPE grown β-Ga2O3 thin films on sapphire. Journal of Crystal Growth. 2022; 126737.
[5] Scheidgen et al. NOMAD: A distributed web-based platform for managing materials science research data. Journal of Open Source Software. 2023; 8(90), 5388.

Abstract
We have developed a pixel pitch (1.35 µm) hybrid wafer bonding technology and successfully demonstrated three layer stacked backside illuminated (BSI) image sensor fabrication with full hybrid Cu-Cu direct bonding process. We found that plasma activation condition on the bonding wafer surface is a key factor of Cu-Cu contact yield improvement for smaller size Cu contacts. Optimized pixel pitch hybrid bonding shows good electrical and reliability performances. For three layer stacking, we developed through Si via (TSV) process. The second hybrid bonding process was adapted on the first hybrid bonded wafers and it shows good electrical performances. This three layer stack technology with pixel pitch hybrid bonding is promising for many different applications.
1. INTRODUCTION
Nowadays, hybrid wafer bonding with Cu-Cu connection is a major production technology of two layer stacked backside illuminated (BSI) image sensor applied for several applications [1]. In addition, three layer stacked BSI image sensor with through Si via (TSV) connection has been introduced to high performance cameras [2]. Furthermore, three layer stacked BSI with pixel pitch Cu-Cu connection will be required in order to realize much higher performance and multi functionality image sensors. In this paper, we developed key technologies for three layer stacking. One is pixel pitch hybrid wafer bonding with minute Cu-Cu connection, the other is three layer stacking process including TSV for middle layer and the 2nd hybrid wafer bonding. Finally, three layer stacking process was demonstrated by adapting these technologies.
2. PIXEL PITCH HYBRID BONDING
Cu-Cu contact with conventional hybrid wafer bonding conditions was studied utilizing the test element group (TEG) structure of Cu-Cu connection. The Cu-Cu contact pitch is 4.00 to 1.35 µm, and Cu electrode size is 1.25 to 0.42 µm square. As the results, Cu-Cu contact yield decreased as the contact pitch and pad size became smaller. The failure was Cu-Cu open mode with small gap between upper and lower Cu electrode. Cu-N compound layer on Cu electrode surface was identified after plasma surface activation by wafer bonder. It indicates that Cu-N compound layer was formed because Nitrogen was dosed into Cu surface during plasma activation. We considered that the Cu-N compound layer formed on both upper and lower minute Cu electrode was not broken by compressive stress of Cu thermal expansion during post annealing. Therefore, the Cu-N layer acted as a barrier, and Cu-Cu metal bonding between minute electrodes was not achieved.
N/Cu ratio of Cu electrode surface was decreased with lowering N dose of plasma condition and 1.35 µm pitch Cu-Cu contact yield (438k chains TEG) was improved. The Qual test of 1.91 µm pitch Cu-Cu contact (0.60 µm sq.) namely, electro migration (EM), stress migration (SM), temperature cycling (TMCL) and I-V, were passed.
3. THREE LAYER STACKING DEMONSTRATION
Three layer stacking was demonstrated following flow; The middle layer is bonded to the top layer by the 1st hybrid bonding with 1.91 µm pitch Cu-Cu connection. After the middle layer Si thinning, the Cu-TSV and the backside BEOL are formed. Then the 2nd hybrid bonding is done in order to bond the backside of the middle layer to the bottom layer with Cu-Cu connection. Finally, top layer Si is thinned and Al-TSV through top layer Si is formed. The stack chain TEG (10 chain circuit through Al-TSV, the 1st hybrid pixel pitch Cu-Cu, Cu-TSV, back BEOL and the 2nd hybrid Cu-Cu) shows the small resistance distribution. As the results, good electrical path of three layer stacking was established.
REFERENCES
[1] Y. Kagawa et al., “Novel Stacked CMOS Image Sensor with Advanced Cu2Cu Hybrid Bonding,” Proc. of IEEE Int. Electron Devices Meeting (IEDM), 2016, pp.208-211.
[2] Y. Kagawa et al., “3D Stacking Technologies for Advanced CMOS Image Sensors,” Proc. of IEEE Int. Interconnect Technology Conf. (IITC), 2021, WS-4.

The von Neumann architecture, which physically separates the CPU and memory devices, has been dominant for a long time. However, it has limitation in computational speed due to bottlenecks and waste a large amount of energy. To address the energy and speed issue, we made a neuromorphic system based on au nanoparticle floating gate memristor (AuNp-FGM).
Our memristor, utilizing graphene as floating gate, MoS₂ as channel, have been recognized for its excellent performance [1,2]. Also, it has been recognized as one of the most promising candidates for neuromorphic system [3]. In this study, by forming au nanoparticles between floating gate and tunneling oxide, a two-terminal memristor which operates in the ±3V region is fabricated. Additionally, we made a neuromorphic array, calculated the energy used for learning simulation.
Using the change of the Fermi energy level (E_f) of graphene, (AuNp-FGM) exhibits memory characteristic. The device exhibits high on/off ratio over than 10⁶, retention more than 9 hours and robust endurance more than 80,000 times. AuNp-FGM also showed low cycle to cycle variability of C_v = 3.6% (n = 90, C_v = , = standard deviation, = mean value). Furthermore, it showed excellent linearity, indicating applicability to neuromorphic systems. For 100 level potentiation (+4V, 0.5s), non-linearity factor ranges from 0.1 to 0.6. For 100 level depression (-3V, 0.2s), it ranges from 2.3 to 4.6 (n = 15). A similar trend was shown even when the number of input pulses was changed (50, 100, 200, 300, 400 inputs).
Based on AuNp-FGM, we fabricated a neuromorphic array consisting of 2 neurons and 32 synapses. Three types of data (horizontal, vertical, diagonal) were used in 40 learning simulation. The total energy consumed was 70uJ, which confirmed to be a 97% energy reduction compared to the previous experiment [3].

References
[1] [1] Vu, Q., Shin, Y., Kim, Y. et al. Nat. Commun. 7, 12725 (2016).
[2] Q. A. Vu, H. Kim, V. L. Nguyen, U. Y. Won, S. Adhikari, K. Kim, Y. H. Lee, W. J. Yu, Adv. Mater. 2017, 29, 1703363.
[3] Won, U.Y., An Vu, Q., Park, S.B. et al. Nat. Commun. 14, 3070 (2023).

Additive manufacturing technology through 3D printing can efficiently produce parts with complex geometries using smaller, portable equipment. 3D printing plays a significant role in manufacturing polymer prototypes and products in various areas. One promising area of research involves incorporating metal powder into polymer filaments for 3D printing, aiming to expand the versatility of the technique and provide an alternative to conventional approaches such as Fused Filament Manufacturing (FFF). The binder polymer formulation plays a crucial role in the success of the manufacturing process. This study presents the development of a specific binding system for the 3D printing of AlSi10Mg filaments. The binder includes low-density polyethylene (LDPE) and thermoplastic starch (TPS). LDPE contributes to structural integrity during binder dissolution, reduces viscosity, and increases strength and stiffness. Meanwhile, TPS, besides being biodegradable, provides flexibility to the filaments. The study included an analysis of the shape and size distribution of metallic powder particles before their incorporation into the polymeric matrix. The composite filaments were produced using reactive extrusion (REX), being mechanically characterized and evaluated for their homogeneity. The reactive extrusion method demonstrated effectiveness in the production of homogeneous composite filaments in relation to the metallic filler incorporated into the thermoplastic polymer blend. Although extruded filaments exhibit inferior mechanical properties compared to thermoplastics generally used in FFF additive manufacturing, the method has potential for application in the manufacture of new composite filament compositions using LDPE and TPS blend matrix as a binding for application in 3D printing.

Atomic-level control of two-dimensional materials provides a novel platform for nanoscale photonics and optoelectronics. Two-dimensional silver (2D-Ag) is stabilized via silver intercalation in epitaxial graphene on SiC. Two phases of 2D-Ag, termed Ag₍₁₎ and the Ag₍₂₎, have been distinguished through 1) in-plane phonon shear modes: Ag₍₁₎ at 17 cm^-1, and Ag₍₂₎ exhibiting two modes at 66 and 92 cm^-1 and 2) linear extinction: Ag₍₁₎ with two absorption peaks centered around 460 and 650 nm, and Ag₍₂₎ with a single band centered around 570 nm. Control of graphene chemistry enables the formation of single phase 2D-Ag, allowing us to explore the inherent differences between Ag₍₁₎ and Ag₍₂₎. The symmetry breaking at the SiC/2D-Ag interface creates an out-of-plane axial dipole, resulting in strong nonlinear optical (NLO) responses. Second harmonic generation microscopy is implemented to investigate the symmetry and second order susceptibility χ⁽²⁾ of Ag₍₁₎ and Ag₍₂₎. Both phases exhibit point group symmetry 3m, verifying the epitaxial growth of the Ag layer on the modified SiC surface. The χ⁽²⁾ of Ag₍₂₎ is orders of magnitude greater than that of Ag₍₁₎. The carrier dynamics of 2D-Ag is studied with near-degenerate transient absorption (TA) spectroscopy. Both phases are excited near the resonance frequencies of 2D-Ag, setting pump wavelengths at 575 and 650 nm. Rapid decays (~2 ps) following the initial bleach are observed in the time-resolved TA spectra for both phases. The difference in coherent oscillations between Ag₍₁₎ and Ag₍₂₎ is attributed to electron-phonon coupling corresponding to the in-plane shear mode of each phase as confirmed with correlated Raman spectroscopy. The results suggest that we are able to tailor the NLO properties and carrier dynamics of 2D-Ag through phase engineering at the atomic level.