اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدHybrid Piezoelectric-Magnetic Neurons: A Proposal for Energy-Efficient Machine Learning
98
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
This paper proposes a spintronic neuron structure composed of a heterostructure of magnets and a piezoelectric with a magnetic tunnel junction (MTJ). The operation of the device is simulated using SPICE models. Simulation results illustrate that the energy dissipation of the proposed neuron compared to that of other spintronic neurons exhibits 70% improvement. Compared to CMOS neurons, the proposed neuron occupies a smaller footprint area and operates using less energy. Owing to its versatility and low-energy operation, the proposed neuron is a promising candidate to be adopted in artificial neural network (ANN) systems.
قيم البحث
اقرأ أيضاً
A strong trend for quantum based technologies and applications follows the avenue of combining different platforms to exploit their complementary technological and functional advantages. Micro and nano-mechanical devices are particularly suitable for
hybrid integration due to the easiness of fabrication at multi-scales and their pervasive coupling with electrons and photons. Here, we report on a nanomechanical technological platform where a silicon chip is combined with an aluminum nitride layer. Exploiting the AlN piezoelectricity, Surface Acoustic Waves are injected in the Si layer where the material has been localy patterned and etched to form a suspended nanostring. Characterizing the nanostring vertical displacement induced by the SAW, we found an external excitation peak efficiency in excess of 500 pm/V at 1 GHz mechanical frequency. Exploiting the long term expertise in silicon photonic and electronic devices as well as the SAW robustness and versatility, our technological platform represents a strong candidate for hybrid quantum systems.
We have designed a new magnetic bed structure with desirable table-like magnetocaloric effect (MCE) by using three kinds of soft ferromagnetic Gd-Al-Co microwire arrays with different Curie temperatures ($T_C$). The $T_C$ interval of these three wire
s is ~10 K and the designed new structure named Sample A. This sample shows a smooth table-like magnetic entropy change ($Delta S_M$) at high applied field change ($mu_0 Delta H=5 T$) ranging from ~92 K to ~107 K. The maximum entropy change ($-Delta S_M^{rm max}$) and refrigerant capacity (RC) for Sample A at $mu_0 Delta H=5 T$ are calculated to be ~9.42 Jkg$^{-1}$K$^{-1}$ and ~676 Jkg$^{-1}$. The calculated curves of $-Delta S_M(T)$ and the corresponding experimental data match well with each other, suggesting that the desirable magnetocaloric properties of the microwire arrays can be designed. Simulation shows that the RC values of the designed systems increase when increasing the interval of $T_C$. The table-like MCE and the enhanced heat-transfer efficiency due to the enhanced surface areas of the microwires make this newly designed magnetic bed very promising for use in energy-efficient magnetic refrigerators.
We report on the integration of large area CVD grown single- and bilayer graphene transparent conductive electrodes (TCEs) on amorphous silicon multispectral photodetectors. The broadband transmission of graphene results in 440% enhancement of the de
tectors spectral response in the ultraviolet (UV) region at {lambda} = 320 nm compared to reference devices with conventional aluminum doped zinc oxide (ZnO:Al) electrodes. The maximum responsivity of the multispectral photodetectors can be tuned in their wavelength from 320 nm to 510 nm by an external bias voltage, allowing single pixel detection of UV to visible light. Graphene electrodes further enable fully flexible diodes on polyimide substrates. Here, an upgrade from single to bilayer graphene boosts the maximum photoresponsivity from 134 mA $W^{-1}$ to 239 mA $W^{-1}$. Interference patterns that are present in conventional TCE devices are suppressed as a result of the atomically thin graphene electrodes. The proposed detectors may be of interest in fields of UV/VIS spectroscopy or for biomedical and life science applications, where the extension to the UV range can be essential.
Many key electronic technologies (e.g., large-scale computing, machine learning, and superconducting electronics) require new memories that are fast, reliable, energy-efficient, and of low-impedance at the same time, which has remained a challenge. N
on-volatile magnetoresistive random access memories (MRAMs) driven by spin-orbit torques (SOTs) have promise to be faster and more energy-efficient than conventional semiconductor and spin-transfer-torque magnetic memories. This work reports that the spin Hall effect of low-resistivity Au0.25Pt0.75 thin films enables ultrafast antidamping-torque switching of SOT-MRAM devices for current pulse widths as short as 200 ps. If combined with industrial-quality lithography and already-demonstrated interfacial engineering, our results show that an optimized MRAM cell based on Au0.25Pt0.75 can have energy-efficient, ultrafast, and reliable switching, e.g. a write energy of < 1 fJ (< 50 fJ) for write error rate of 50% (<1e-5) for 1 ns pulses. The antidamping torque switching of the Au0.25Pt0.75 devices is 10 times faster than expected from a rigid macrospin model, most likely because of the fast micromagnetics due to the enhanced non-uniformity within the free layer. These results demonstrate the feasibility of Au0.25Pt0.75-based SOT-MRAMs as a candidate for ultrafast, reliable, energy-efficient, low-impedance, and unlimited-endurance memory.
Graphene on ferroelectric structures can be promising candidates for advanced field effect transistors, modulators and electrical transducers, providing that research of their electrotransport and electromechanical performances can be lifted up from
mostly empirical to prognostic theoretical level.Recently we have shown that alternating piezoelectric displacement of the ferroelectric domain surfaces can lead to the alternate stretching and separation of graphene areas at the steps between elongated and contracted domains, and the conductance of graphene channel can be increased essentially at room temperature, because electrons in the stretched section scatter on acoustic phonons.The piezoelectric mechanism of graphene conductance control requires systematic studies of the ambient condition impact on its manifestations. This theoretical work studies in details the temperature behavior of the graphene conductance changes induced by piezoelectric effect in a ferroelectric substrate with domain stripes.We revealed the possibility to control graphene conductance (that can change up to 100 times for PZT ferroelectric substrate) by tuning the ambient temperature from low values to the critical one for given gate voltage and channel length.Also we demonstrate the possibility to control graphene conductance changes up to one hundred of times by tuning the gate voltage from 0 to the critical value at a given temperature and channel length. Obtained results can be open the way towards graphene on ferroelectric applications in piezoresistive memories operating in a wide temperature range.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
