Bandwidth Density Research Articles

SSA-over-array (SSoA): A stacked DRAM architecture for near-memory computing

Aiming to enhance the bandwidth in near-memory computing, this paper proposes a SSA-over-array (SSoA) architecture. By relocating the secondary sense amplifier (SSA) from dynamic random access memory (DRAM) to the logic die and repositioning the DRAM-to-logic stacking interface closer to the DRAM core, the SSoA overcomes the layout and area limitations of SSA and master DQ (MDQ), leading to improvements in DRAM data-width density and frequency, significantly enhancing bandwidth density. The quantitative evaluation results show a 70.18 times improvement in bandwidth per unit area over the baseline, with a maximum bandwidth of 168.296 Tbps/Gb. We believe the SSoA is poised to redefine near-memory computing development strategies.

Electro-optic frequency comb generation via cascaded modulators driven at lower frequency harmonics

Electro-optical modulation of a continuous wave laser is a highly stable way to generate frequency combs, gaining popularity in telecommunication and spectroscopic applications. These combs are generated by modulating non-linear electro-optic crystals with radio frequencies, creating equally spaced side-bands centered around the single-frequency seed laser. Electro-optic frequency comb architectures often choose between optical bandwidth (cascaded GHz combs) or higher mode density (chirped RF generation). This work demonstrates an electro-optic frequency comb with > 120 GHz of bandwidth and an 80 MHz repetition rate. The comb has three cascaded electro-optic modulators driven at sequentially lower harmonics, the last megahertz modulation dictating the repetition rate. This architecture can modulate at any individual harmonic and repetition rate without changes to the components. This comb can be used in any applications where a stable and tunable repetition rate is needed.

RF crosstalk mitigation via floating shields in parallel silicon traveling-wave modulators

High-density integration of electro-optic (E-O) elements in photonic integrated circuits (PICs) offers compact, low-cost high-capacity transceiver modules. Tighter component spacing increases compactness, but leads to crosstalk that can degrade performance. We investigate and characterize the crosstalk between silicon traveling-wave Mach-Zehnder modulators (TW-MZMs) laid out in parallel in PICs. We propose the use of floating shield strips to reduce E-O crosstalk between these TW-MZMs. We show via simulation and experiment that these strips can increase shoreline bandwidth density by nearly 50%. The improvement in E-O crosstalk covers frequencies as high as 60 GHz.

16 × 112 Gbps directly modulated membrane laser array for co-packaged interconnects

A low-cost and low-power-consumption optical transmitter with a narrow shoreline is crucial for short-reach optical communication. To increase the shoreline bandwidth density (Gbps/mm) at low cost, multiple optical components, including lasers, should be integrated on a single chip. In this study, we develop a sixteen-channel membrane laser array integrated with silica-based spot-size convertors on a SiO2/Si substrate, with a footprint of 1.11 × 2.75 mm2. A thin (340 nm) membrane buried-heterostructure sandwiched between low-index silica-based materials provides high carrier and optical confinement, which contributes to reducing power consumption. We demonstrated direct modulation with 28- and 56-GBaud PAM4 signals and verified that 2-km data transmission is feasible for all sixteen channels.

Polymeric piezoelectric accelerometers with high sensitivity, broad bandwidth, and low noise density for organic electronics and wearable microsystems

Piezoelectric accelerometers excel in vibration sensing. In the emerging trend of fully organic electronic microsystems, polymeric piezoelectric accelerometers can be used as vital front-end components to capture dynamic signals, such as vocal vibrations in wearable speaking assistants for those with speaking difficulties. However, high-performance polymeric piezoelectric accelerometers suitable for such applications are rare. Piezoelectric organic compounds such as PVDF have inferior properties to their inorganic counterparts such as PZT. Consequently, most existing polymeric piezoelectric accelerometers have very unbalanced performance metrics. They often sacrifice resonance frequency and bandwidth for a flat-band sensitivity comparable to those of PZT-based accelerometers, leading to increased noise density and limited application potentials. In this study, a new polymeric piezoelectric accelerometer design to overcome the material limitations of PVDF is introduced. This new design aims to simultaneously achieve high sensitivity, broad bandwidth, and low noise. Five samples were manufactured and characterized, demonstrating an average sensitivity of 29.45 pC/g within a ± 10 g input range, a 5% flat band of 160 Hz, and an in-band noise density of 1.4 µg/Hz\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\sqrt{{Hz}}$$\\end{document}. These results surpass those of many PZT-based piezoelectric accelerometers, showing the feasibility of achieving comprehensively high performance in polymeric piezoelectric accelerometers to increase their potential in novel applications such as organic microsystems.

High Performance and Stackable Trampoline Like‐Triboelectric Vibration Energy Harvester for In‐Situ Powering Sensor Node with Data Wirelessly Transmitted Over 1000‐m

AbstractAddressing the power supply challenges of wireless sensor nodes is pivotal for advancing the development of the Internet of Things (IoT). This work proposes a high performance and stackable trampoline like‐triboelectric vibration energy harvester (T‐TVEH) that can efficiently harvest ultra‐wideband vibrational energy for in‐situ powering of sensor nodes. The unique structural design and material selection enables T‐TVEH to represent a breakthrough in terms of both working bandwidth and power density compared to recent research efforts of vibration energy harvesting. Specifically, the working bandwidth and the peak power density of T‐TVEH is 192 Hz and 5.9 W m−2, which are higher than previous related studies by 156% and 59.2%, respectively. Based on the excellent performance of the T‐TVEH, a wireless sensor node for monitoring machinery condition is constructed. Temperature, humidity, and frequency information are successfully acquired and transmitted to 1000‐m through the wireless sensor node, which is nine times improved compared to related studies. Meanwhile, it achieves fully self‐powered wireless operation monitoring and abnormal alarm on a real ship's marine diesel engine. Overall, this study proposed an innovative solution for in‐situ power supply of wireless sensor nodes, which has broad application prospects in the field of the IOT.

Bringing better bandwidth and power density for wireless sensor networks

Novel mechanism promises progress for magnetic energy harvesters and the Internet of Things.

Laser isotope separation of 203Tl through pulsed laser optical pumping

Isotope separation of 203Tl through pulsed laser optical pumping into the 6s26p2P3/2∘ meta-stable state, followed by pulsed laser ionization, has been theoretically studied through density matrix formalism. The efficiency of the isotope separation process has been compared with the cw laser optical pumping followed by pulsed laser photoionization reported earlier. The effect of peak power density, bandwidth and number density of atoms on the degree of enrichment and the production rate has been systematically studied. The optimum parameters for the efficient laser isotope separation 203Tl have been derived. It has been shown that using this method, it is possible to produce about 1.5 kg/y of 203Tl isotope enriched to 90%.

High-Power Modular GaN Based Power Supply for MedAustron Scanning Magnets

MedAustron is a synchrotron-based ion therapy center in Wiener Neustadt, Austria, constantly working towards the performance improvement of cancer treatment. A major improvement opportunity comes from the scanning magnets system – a crucial element of dose delivery system at MedAustron - that is influenced by the bandwidth and power density of the magnet power supplies. Therefore, a novel highly modular power converter, based on the latest gallium nitride technology is being developed to tackle the aforementioned requirements. One sub-module of this power supply is based on two H-bridges that are operated in hard parallel and are integrated into a standardized euro-crate form factor with target output power specifications of 720 kW. This design also includes a 4th-order output Bessel filter to meet the ripple requirements for clinical operation. Furthermore, those sub-modules offer the possibility of being connected, cascaded or interleaved, in order to meet different output power requirements, providing a high modularity aspect. The status of this development in terms of requirements, global topology, and control structure is presented.

OML-PCM: optical multi-level phase change memory architecture for embedded computing systems

Unlike Dynamic Random Access Memory (DRAM), Phase Change Memory (PCM) offers higher density, longer data retention, and improved scalability because of its non-volatility and low leakage power. However, Electrically-Addressable PCM (EPCM) has a higher dynamic power and long latency than DRAM. To address these issues, scientists have developed Optically-Addressable PCM (OPCM), which uses 5-level cells instead of 2-level cells in EPCM. A silicon photonic link allows optical signals to reach OPCM cells at a high speed. Hence, OPCM can achieve a higher density while maintaining better performance at multi-level cells and consuming less power per access. However, OPCM is not suitable for general use since the photonic links do not provide an electrical interface to the processor. The aim of this paper is to present a hybrid OPCM architecture based on the use of novel multi-bank clusters with distinctive properties. Electrical-Optical-Electrical conversion (EOE) allows OPCM cells to be randomly accessed by using DRAM-like circuitry. The proposed hybrid design with multi-core processing and OPCM achieves a 2.13x speedup over previous approaches while consuming less Central Processing Unit (CPU) power. It is important to note that the proposed design offers 97 units fewer power-consistent bits than EPCM. In addition, the proposed architecture provides comparable performance and power to DDR4, as well as improved bandwidth density, space efficiency, and versatility. The Gem5 simulator was used to evaluate the design. Based on the outcomes of the analysis, the proposed architecture offers 2.08x and 2.14x better evaluations and density performance than EPCM. Furthermore, the execution time has been reduced by 2.13x, the analysis time by 1.23x, and the composition time by 4.60%.

Effective attenuation of electromagnetic waves via silane surface modified zinc oxide/polybenzoxazine nanocomposites for EMI shielding application

In recent years, there has been a significant emphasis on advancing nanomaterials with exceptional properties, including high reflection loss (RL), reduced thickness, broad bandwidth, and low density. The primary goal is to augment their effectiveness in absorbing EM waves while maintaining a streamlined manufacturing process. This study, investigated the application of novel thermosetting phenolic resin as a polymeric matrix for EM shielding. Among these resins options, benzoxazine stands out due to its exceptional characteristics, which encompass facile monomer preparation through solvent-free synthesis, the absence of by-product release during polymerization, high thermal stability and superb mechanical properties. The assessement of this resin both individually and in combination with zinc oxide nanoparticles. Silanization of the nanoparticles was confirmed via Fourier transform infrared spectroscopy (FTIR). The thermal properties were investigated through differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). The outcomes indicated that the incorporation of zinc oxide nanoparticles enhanced the material's stability. The thermomechanical attributes of the composites were evaluated using Dynamic Mechanical Analysis (DMA), revealing an increase in the storage modulus and a decrease in the glass transition temperature (Tg). In summary, the results from the electromagnetic tests underscore the remarkable capability of the developed composites in absorbing EM radiation, achieving highly impressive RL values of up to –50 dB. However, they exhibited acceptable shielding efficiency.

A Gallium Nitride Based High Bandwidth, Extended Single-Ended Primary Inductance Converter, for Fast Balancing of Ultracapacitors

Ultracapacitors (UCs) are the preferred choice for specific power requirements. Series and parallel of UCs are required to match the required voltage and power, as the available voltage range is 2.5 to 3 VDC. Passive and active balancing is inevitable here, while some demand, high bandwidth control for fast balancing and stable operations. A novel Extended-Single Ended Primary Inductance Converter (E-SEPIC) is used to actively balance the UCs. A new, simple and effective method is formulated to achieve a higher Degree of Balancing of UCs. A two-stage E-SEPIC with Silicon (Si) and Gallium Nitride (GaN) switch based are demonstrated with an input of 12 VDC to balance up to 3 VDC UCs at the output for high bandwidth and high-power density properties. An effective method of emulating impedance of UCs using Hybrid Electro-chemical Impedance Spectroscopy is evolved. The devised novel method of balancing is effective in height of balancing with GaN based converter exhibiting higher bandwidth of around 30 kHz around 18 times better, with increased efficiency of 12% with an improved power-density of 8.25 times in comparison with that of Si based. This contribution of GaN properties in electrical performance enhancements has influenced the design of Power Electronics applications.

Enhancing electromagnetic absorption in microwave of epoxy resin through incorporated surface modified aluminum oxide nanoparticles

The development of novel nanomaterials with superior reflection loss (RL), thin thickness, wide bandwidth, and low density has recently received significant attention in an effort to increase their effectiveness in electromagnetic (EM) microwave absorption while maintaining an easy manufacturing process. Naturally, using conventional materials with magnetic or dielectric loss makes it difficult to achieve the last criterion and limitations the thickness of the absorber and mass production. In this work, microwave nanocomposites samples were prepared on epoxy resin with different percentages of the nanoparticles of aluminum oxide (epoxy/Al2O3) with surfaces treated by amino propyl and silane. The surface treatment was carried out to improve the adhesion, morphology, and EM properties of the samples to enhance the microwave absorption and the Shielding Effectiveness (SE). Therefore, the treated sample can be considered as potential candidate for high frequency EM absorption applications. Then, the study is followed by structural, morphological, thermal, and electrical characterization of the prepared samples in order to enhance the performances of the latters. The EM absorption and the shielding effectiveness were done in the X band frequency (8 GHz-12 GHz). The obtained results confirmed that the nanocomposites treated samples exhibit higher EM absorption performance and better electrical characteristics. Overall, these results put forward the role played by the addition of the nanoparticles films as high-performance materials in electronic devices and EM applications.

Ultra‐High Bandwidth Density and Power Efficiency Chip‐To‐Chip Multimode Transmission through a Rectangular Core Few‐Mode Fiber

AbstractOptical interconnects have emerged as promising solutions for assisting electrical interconnects in short‐reach scenarios, where high bandwidth density and energy efficiency are of particular importance. Mode‐division multiplexing (MDM), capable of enhancing the bandwidth density and energy efficiency of optical systems, has drawn tremendous interest. However, the chip‐to‐chip MDM optical interconnects are impeded by the multimode chip‐fiber interfaces, where the transverse modes on the MDM chip mismatch with the linear polarization modes in the circular core few‐mode fiber (CCF). Moreover, the high differential group delays (DGDs) of a conventional CCF make the digital signal processing (DSP) computation complex thus consuming high power. To overcome these bottlenecks, a new multimode coupling solution based on a rectangular core few‐mode fiber with ultra‐low DGDs and an integrated multimode coupler is proposed. Based on this coupling scheme, a chip‐to‐chip MDM transmission experiment is performed on the TE00, TM00, TE10, and TM10 modes, and the highest bandwidth density of 19.4 Tb s mm −1 is realized at the chip facet. The DSP power consumption is ≈6% of a conventional CCF‐based system. It is believed that this work will pave the way for ultra‐compact, high density, low cost, and low computation‐complexity optical interconnects, such as data centers, cloud computing, and telecommunications.

Power equalization-based disaggregated optical network using switched gain equalization controlled amplifiers.

Optical network disaggregation is an attractive agreement to avoid vendor lock-in in choosing desirable partners based on lead time of delivery and facilitates strong negotiations. Disaggregation of conventional coherent transceivers (XCVRs) with 400G ZR+ technologies (of similar bandwidth density and form factor QSFP-DD) allows operators to enjoy significant cost reduction and substantial power savings in high-capacity IP over a WDM transmission. However, interoperability between low power and high back-to-back required optical signal-to-noise ratio (OSNR) of ZR+ and high power with relatively lower back-to-back required OSNR of conventional coherent XCVRs limits the performance of the latter, which makes disaggregation more complex. This can be avoided by equalizing the power of ZR+ and conventional XCVRs, by wavelength-selective switches (WSSs) in re-configurable optical add/drop multiplexers. However, such power equalization can lead to lower link OSNR, originating from insertion loss and penalty caused by cascaded WSSs. In this Letter, we use a dynamic gain equalizer of a switched gain equalization controlled (SGEC) amplifier at each in-line amplifier site to equalize the powers, without introducing insertion loss or filter penalty. We report a field trial of a low power 400G-ZR+ QSFP-DD-DCO as an alien channel over ten high-power coherent host channels, interoperating with open forward error correction, between two different vendors. We obtain a similar Q-margin in the absence and presence of 400-ZR+ channel, suggesting the equalization by SGEC amplifiers. We further simulate the power and OSNR evolution of channels along the field trial link to verify the equalization. Simulation results are consistent with that of the field trial.

Bandwidth Density Analysis of Coded Free-Space Optical Interconnects

The performance of free-space optical interconnects (FSOIs) system is significantly influenced by noise, similar to any wireless communication system. This noise has a notable impact on both the bandwidth density and data rate of FSOIs system. To address these challenges, this study proposes the utilization of vertical-cavity-surface-emitting laser (VCSEL) arrays on the transmitter side and photodetector arrays on the receiver side for FSOIs. The study investigates the bandwidth density of the system with and without coding while maintaining a specific bit error rate. An analysis is conducted in the presence of higher-order modes in the laser beams of the FSOIs system and a fundamental Gaussian operating mode. The presence of the higher-order modes leads to degradation in the performance of the FSOIs system in terms of bandwidth density. In addition, we examine the impact of the signal-to-noise ratio (SNR) on the system’s bandwidth density for each considered operating mode. The provided simulation results clearly demonstrate that coding significantly enhances the bandwidth density of the systems, with the extent of improvement being closely tied to the employed code rate and codeword length.

Artificial Intelligence Assisted Enhanced Energy Efficient Model for Device-to-Device Communication in 5G Networks

Device-to-device (D2D) communications promise spectral and energy efficiency, total system capacity, and excellent data rates. These improvements in network performance led to much D2D research, but it revealed significant difficulties before their full potential could be realized in 5G networks. D2D communication in 5G networks can bring about performance gains regarding spectral and energy efficiency, total system capacity, and data rate. The major challenge in the 5G network is to meet latency, bandwidth, and traffic density requirements. In addition, the next generation of cellular networks must have increased throughput, decreased power consumption, and guaranteed Quality of Service. This potential, however, is associated with substantial difficulties. To address these challenges and improve the system capabilities of D2D networks, a deep learning-based Improved D2D communication (DLID2DC) model has been proposed. The proposed model is explicitly intended for 5G networks, using the exterior public cloud to replace automation with an explainable artificial intelligence (XAI) method to analyze communication needs. The communicated needs allow a selection of methodologies to transfer machine data from the remote server to the smart devices. The model utilizes deep learning algorithms for resource allocation in D2D communication to maximize the utilization of available spectrum resources. Experimental tests prove that the DLID2DC model brings about better throughput, lower end-to-end delay, better fairness, and improved energy efficiency than traditional methods.

Improving Rainfall Fields in Data-Scarce Basins: Influence of the Kernel Bandwidth Value of Merging on Hydrometeorological Modeling

Accurate rainfall fields are important for several hydrological and meteorological applications. In poorly instrumented basins, the lack of rain gauges heavily affects the spatial rainfall estimation, yet neither remote sensed nor climate model estimates are good enough for management applications. To tackle this problem, we investigate the impacts of combining both sources of information by varying the kernel bandwidth value of the double smoothing merging algorithm and analyzing the error of the rainfall fields. We explored the correlation between rain gauge density and bandwidth and compared the results against classical geostatistical interpolation methods based solely on in-situ measurements. Propagation of rainfall error into hydrological modelling is usual, and therefore we evaluated the influence of the bandwidth in streamflow simulations implementing two hydrological models. The hydrological evaluation considered the analysis of hydrological signatures rather than just performance metrics. We found that there is a clear correlation between kernel bandwidth and monitoring network density and that the bandwidth also affects hydrological performance. Simple bilinear downscaling did not produce a significant difference in meteorological or hydrological errors, and rain gauge network configuration also impacts the error of the field. We conclude that merging outperforms the results of classical interpolation methods, in some cases by 20% or 50%, suggesting the suitability of the method for being applied in data-scarce domains.

Massively scalable Kerr comb-driven silicon photonic link

The growth of computing needs for artificial intelligence and machine learning is critically challenging data communications in today’s data-centre systems. Data movement, dominated by energy costs and limited ‘chip-escape’ bandwidth densities, is perhaps the singular factor determining the scalability of future systems. Using light to send information between compute nodes in such systems can dramatically increase the available bandwidth while simultaneously decreasing energy consumption. Through wavelength-division multiplexing with chip-based microresonator Kerr frequency combs, independent information channels can be encoded onto many distinct colours of light in the same optical fibre for massively parallel data transmission with low energy. Although previous high-bandwidth demonstrations have relied on benchtop equipment for filtering and modulating Kerr comb wavelength channels, data-centre interconnects require a compact on-chip form factor for these operations. Here we demonstrate a massively scalable chip-based silicon photonic data link using a Kerr comb source enabled by a new link architecture and experimentally show aggregate single-fibre data transmission of 512 Gb s−1 across 32 independent wavelength channels. The demonstrated architecture is fundamentally scalable to hundreds of wavelength channels, enabling massively parallel terabit-scale optical interconnects for future green hyperscale data centres.

An Active Inductor Based Transimpedance Amplifier with Two Local Feedbacks as a Fiber Optic Application

An important use of active inductor feedback is reported. It was evident that a frequency dependent impedance of an active inductor is similar to that of an ordinary spiral inductor which is a key advantage in integrated circuit design given the fact that Mosfet-based active inductor circuit occupies less volume on board a chip. In this work, a 42 dBΩ transimpedance (TIA) gain at 1 GHz is obtained with 33 𝑝𝐴⁄√𝐻𝑧 of input referred noise with a very low power consumption of 0.605 mW. The Common-Gate Common-Source input stage with an active inductor feedback is reported in this work for the first-time long side a second stage current mirror with additional local active inductor feedback. The extremely low level of power consumption is considered to be another key advantage for that matter despite the moderate levels of TIA gain, bandwidth and input-referred noise current spectral density with 1V DC supply voltage. The LTspice software was used for simulation.