A non‐volatile conjugated polymer‐based electrochemical memristor (cPECM), derived from sodium 4‐[(2,3‐dihydrothieno[3,4‐b][1,4]dioxin‐2‐yl)methoxy]butane‐2‐sulfonate (S‐EDOT), is fabricated through roll‐to‐roll printing and exhibited neuromorphic properties. The 3‐terminal device employed a “read” channel where conductivity of the water‐soluble, self‐doped S‐PEDOT is equated to synaptic weight and was electrically decoupled from the programming electrode. For the model system, a +2500 mV programming pulse of 100 ms duration resulted in a 0.136 μS resolution in conductivity change, giving over 1000 distinct conductivity states for one cycle. The minimum programming power requirements of the cPECM was 0.31 pJ mm−2and with advanced printing techniques, a 0.1 fJ requirement for a 20 μm device is achievable. The mathematical operations of addition, subtraction, multiplication, and division are demonstrated with a single cPECM, as well as the logic gates AND, OR, NAND, and NOR. This demonstration of a printed cPECM is the first step toward the implementation of a mass produced electrochemical memristor that combines information storage and processing and may allow for the realization of printable artificial neural networks.
This content will become publicly available on November 1, 2024
Electrochemical‐based memristors are highly attractive that are capable of nonvolatile analog tuning, long‐term state stability, low power consumption, device scalability, and fast switching speeds. Through the combination of film deposition techniques, i.e., vapor phase polymerization and screen printing, fabrication of a poly(4‐(6‐hexyl)‐4
- Award ID(s):
- 1655740
- NSF-PAR ID:
- 10474245
- Publisher / Repository:
- Wiley
- Date Published:
- Journal Name:
- Advanced Materials Technologies
- Volume:
- 8
- Issue:
- 21
- ISSN:
- 2365-709X
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
more » « less
By mimicking biomimetic synaptic processes, the success of artificial intelligence (AI) has been astounding with various applications such as driving automation, big data analysis, and natural-language processing.[1-4] Due to a large quantity of data transmission between the separated memory unit and the logic unit, the classical computing system with von Neumann architecture consumes excessive energy and has a significant processing delay.[5] Furthermore, the speed difference between the two units also causes extra delay, which is referred to as the memory wall.[6, 7] To keep pace with the rapid growth of AI applications, enhanced hardware systems that particularly feature an energy-efficient and high-speed hardware system need to be secured. The novel neuromorphic computing system, an in-memory architecture with low power consumption, has been suggested as an alternative to the conventional system. Memristors with analog-type resistive switching behavior are a promising candidate for implementing the neuromorphic computing system since the devices can modulate the conductance with cycles that act as synaptic weights to process input signals and store information.[8, 9]
The memristor has sparked tremendous interest due to its simple two-terminal structure, including top electrode (TE), bottom electrode (BE), and an intermediate resistive switching (RS) layer. Many oxide materials, including HfO2, Ta2O5, and IGZO, have extensively been studied as an RS layer of memristors. Silicon dioxide (SiO2) features 3D structural conformity with the conventional CMOS technology and high wafer-scale homogeneity, which has benefited modern microelectronic devices as dielectric and/or passivation layers. Therefore, the use of SiO2as a memristor RS layer for neuromorphic computing is expected to be compatible with current Si technology with minimal processing and material-related complexities.
In this work, we proposed SiO2-based memristor and investigated switching behaviors metallized with different reduction potentials by applying pure Cu and Ag, and their alloys with varied ratios. Heavily doped p-type silicon was chosen as BE in order to exclude any effects of the BE ions on the memristor performance. We previously reported that the selection of TE is crucial for achieving a high memory window and stable switching performance. According to the study which compares the roles of Cu (switching stabilizer) and Ag (large switching window performer) TEs for oxide memristors, we have selected the TE materials and their alloys to engineer the SiO2-based memristor characteristics. The Ag TE leads to a larger memory window of the SiO2memristor, but the device shows relatively large variation and less reliability. On the other hand, the Cu TE device presents uniform gradual switching behavior which is in line with our previous report that Cu can be served as a stabilizer, but with small on/off ratio.[9] These distinct performances with Cu and Ag metallization leads us to utilize a Cu/Ag alloy as the TE. Various compositions of Cu/Ag were examined for the optimization of the memristor TEs. With a Cu/Ag alloying TE with optimized ratio, our SiO2based memristor demonstrates uniform switching behavior and memory window for analog switching applications. Also, it shows ideal potentiation and depression synaptic behavior under the positive/negative spikes (pulse train).
In conclusion, the SiO2memristors with different metallization were established. To tune the property of RS layer, the sputtering conditions of RS were varied. To investigate the influence of TE selections on switching performance of memristor, we integrated Cu, Ag and Cu/Ag alloy as TEs and compared the switch characteristics. Our encouraging results clearly demonstrate that SiO2with Cu/Ag is a promising memristor device with synaptic switching behavior in neuromorphic computing applications.
Acknowledgement This work was supported by the U.S. National Science Foundation (NSF) Award No. ECCS-1931088. S.L. and H.W.S. acknowledge the support from the Improvement of Measurement Standards and Technology for Mechanical Metrology (Grant No. 22011044) by KRISS.
References [1] Young
et al. ,IEEE Computational Intelligence Magazine, vol. 13, no. 3, pp. 55-75, 2018.[2] Hadsell
et al. ,Journal of Field Robotics, vol. 26, no. 2, pp. 120-144, 2009.[3] Najafabadi
et al. ,Journal of Big Data, vol. 2, no. 1, p. 1, 2015.[4] Zhao
et al. ,Applied Physics Reviews, vol. 7, no. 1, 2020.[5] Zidan
et al. ,Nature Electronics, vol. 1, no. 1, pp. 22-29, 2018.[6] Wulf
et al., SIGARCH Comput. Archit. News, vol. 23, no. 1, pp. 20–24, 1995.[7] Wilkes,
SIGARCH Comput. Archit. News, vol. 23, no. 4, pp. 4–6, 1995.[8] Ielmini
et al., Nature Electronics, vol. 1, no. 6, pp. 333-343, 2018.[9] Chang
et al., Nano Letters, vol. 10, no. 4, pp. 1297-1301, 2010.[10] Qin
et al. , Physica Status Solidi (RRL) - Rapid Research Letters, pssr.202200075R1, In press, 2022. -
more » « less
Growths of monoclinic (Al
x Ga1−x )2O3thin films up to 99% Al contents are demonstrated via metalorganic chemical vapor deposition (MOCVD) using trimethylgallium (TMGa) as the Ga precursor. The utilization of TMGa, rather than triethylgallium, enables a significant improvement of the growth rates (>2.5 μm h−1) of β‐(Alx Ga1−x )2O3thin films on (010), (100), and (01) β‐Ga2O3substrates. By systematically tuning the precursor molar flow rates, growth of coherently strained phase pure β‐(Alx Ga1−x )2O3films is demonstrated by comprehensive material characterizations via high‐resolution X‐ray diffraction (XRD) and atomic‐resolution scanning transmission electron microscopy (STEM) imaging. Monoclinic (Alx Ga1−x )2O3films with Al contents up to 99, 29, and 16% are achieved on (100), (010), and (01) β‐Ga2O3substrates, respectively. Beyond 29% of Al incorporation, the (010) (Alx Ga1−x )2O3films exhibit β‐ to γ‐phase segregation. β‐(Alx Ga1−x )2O3films grown on (01) β‐Ga2O3show local segregation of Al along (100) plane. Record‐high Al incorporations up to 99% in monoclinic (Alx Ga1−x )2O3grown on (100) Ga2O3are confirmed from XRD, STEM, electron nanodiffraction, and X‐ray photoelectron spectroscopy measurements. These results indicate great promises of MOCVD development of β‐(Alx Ga1−x )2O3films and heterostructures with high Al content and growth rates using TMGa for next‐generation high‐power and high‐frequency electronic devices. -
more » « less
Conformal coating of nanopores with functional polymer nanolayers is the key to many emerging technologies such as miniature sensors and membranes for advanced molecular separations. While the polymer coatings are often used to introduce functional moieties, their controlled growth under nanoconfinement could serve as a new approach to manipulate the size and shape of coated nanopores, hence, enabling novel functions like molecular separation. However, precise control of coating thickness in the longitudinal direction of a nanopore is limited by the lack of a characterization method to profile coating thickness within the nanoconfined space. Here, we report an experimental approach that combines ion milling (IM) and high-resolution field emission scanning electron microscopy (FESEM) for acquiring an accurate depth profile of ultrathin (∼20 nm or less) coatings synthesized inside nanopores via initiated chemical vapor deposition (iCVD). The enhanced capability of this approach stems from the excellent x–y resolution achieved by FESEM (i.e., 4.9 nm/pixel), robust depth ( z) control enabled by IM (step size as small as 100 nm with R2 = 0.992), and the statistical power afforded by high-throughput sampling (i.e., ∼2000 individual pores). With that capability, we were able to determine with unparalleled accuracy and precision the depth profile of coating thickness and iCVD kinetics along 110-nm-diameter nanopores. That allowed us to uncover an unexpected coating depth profile featuring a maximum rate of polymerization at ∼250 nm underneath the top surface, i.e., down the pores, which we termed “necking.” The necking phenomenon deviates considerably from the conventionally assumed monotonous decrease in thickness along the longitudinal direction into a nanopore, as predicted by the diffusion-limited kinetics model of free radical polymerization. An initiator-centric collision model was then developed, which suggests that under the experimental conditions, the confinement imposed by the nanopores may lead to local amplification of the effective free radical concentration at z ≤ 100 nm and attenuation at z ≥ 500 nm, thus contributing to the observed necking phenomenon. The ion-milling-enabled depth profiling of ultrathin coatings inside nanopores, along with the initiator-mediated coating thickness control in the z-direction, may serve to enhance the performance of size-exclusion filtration membranes and even provide more flexible control of nanopore shape in the z dimension.
-
more » « less
Water vapor (H2O) is one of the brightest molecular emitters after carbon monoxide (CO) in galaxies with high infrared (IR) luminosity, allowing us to investigate the warm and dense phase of the interstellar medium (ISM) where star formation occurs. However, due to the complexity of its radiative spectrum, H2O is not frequently exploited as an ISM tracer in distant galaxies. Therefore, H2O studies of the warm and dense gas at high-
z remain largely unexplored. In this work, we present observations conducted with the Northern Extended Millimeter Array (NOEMA) toward threez > 6 IR-bright quasarsJ2310+1855 ,J1148+5251 , andJ0439+1634 targeted in their multiple para- and ortho-H2O transitions (312 − 303, 111 − 000, 220 − 211, and 422 − 413), as well as their far-IR (FIR) dust continuum. By combining our data with previous measurements from the literature, we estimated the dust masses and temperatures, continuum optical depths, IR luminosities, and star formation rates (SFR) from the FIR continuum. We modeled the H2O lines using the MOLPOP-CEP radiative transfer code, finding that water vapor lines in our quasar host galaxies are primarily excited in the warm, dense (with a gas kinetic temperature and density ofT kin = 50 K,n H2 ∼ 104.5 − 105 cm−3) molecular medium with a water vapor column density ofN H2O ∼ 2 × 1017 − 3 × 1018 cm−3. High-J H2O lines are mainly radiatively pumped by the intense optically-thin far-IR radiation field associated with a warm dust component at temperatures ofT dust ∼ 80 − 190 K that account for < 5 − 10% of the total dust mass. In the case of J2310+1855, our analysis points to a relatively high value of the continuum optical depth at 100 μm (τ 100 ∼ 1). Our results are in agreement with expectations based on the H2O spectral line energy distribution of local and high-z ultra-luminous IR galaxies and active galactic nuclei (AGN). The analysis of the Boltzmann diagrams highlights the interplay between collisions and IR pumping in populating the high H2O energy levels and it allows us to directly compare the excitation conditions in the targeted quasar host galaxies. In addition, the observations enable us to sample the high-luminosity part of the H2O–total-IR (TIR) luminosity relations (L H2O −L TIR). Overall, our results point to supralinear trends that suggest H2O–TIR relations are likely driven by IR pumping, rather than the mere co-spatiality between the FIR continuum- and line-emitting regions. The observedL H2O/L TIRratios in ourz > 6 quasars do not show any strong deviations with respect to those measured in star-forming galaxies and AGN at lower redshifts. This supports the notion that H2O can be likely used to trace the star formation activity buried deep within the dense molecular clouds.
