Complete ATmega328P reference: PDIP and QFN/MLF pinout, port registers, fuse bytes, Arduino pin mapping, errata, and lifecycle status.

ATmega328P: Pinout, Registers, Fuses, and Arduino Pin Mapping

A spectrum analyzer and signal generator are two of the fundamental equipment used in signal integrity testing

How engineers measure jitter, crosstalk, impedance, eye margin, and BER for reliable PCB designs

Signal Integrity Testing for High-Speed PCB Design

Safety is important in all designs, but it is especially important in industrial and automotive markets. While both industries have comparably rigorous standards, automotive silicon generally offers a more comprehensive feature set, including deeper diagnostic architectures, lockstep execution, greater safety software collateral, and options for coordinated supervision with a safety power management IC (PMIC). Despite those benefits, many industrial and robotics designers fear that automotive-grade silicon is too different or complex to be compliant in their own designs. Fortunately, with NXP’s long track record in both industries, industrial designers can easily harness the benefits of automotive silicon. Beyond industrial-native MCU families such as MCX E series, NXP’s S32K microcontroller family, paired with its companion PMICs, offers a direct path to cross-domain functional safety without reinventing safety architectures from scratch. <h1><img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785439356731-1785439356731.jpeg" width="624" height="329" class="fr-fic fr-dii" />Source: NXPExploring Automotive and Industrial Compatibility</h1>Automotive and industrial applications have different sets of standards and classifications, which can make the sectors seem very different on the surface. For instance:<ul><li>Automotive: Follows ISO 26262 standard and Automotive Safety Integrity Level (ASIL) classification system.</li><li>Industrial: Follows IEC 61508 standard and Safety Integrity Level (SIL) classification system.</li><li>Machine control systems: Follows ISO 13849-1 and Performance Level (PL) classification system.</li></ul>Because of these differences, many engineers don’t consider the possibility of cross-domain compatibility. But, in reality, these standards and classifications overlap more than they differ. While no direct equivalence mapping exists between an ASIL D and SIL 3, for example, engineers can still demonstrate system-level equivalence through proper analysis, documentation, and design evidence. This process may appear daunting, but with the right silicon provider, it is straightforward. If silicon vendors supply diagnostic coverage data, safety manuals, and Safety Element out of Context (SEooC) documentation aligned to both frameworks, design teams can easily prove equivalence in either direction. Otherwise, teams must prove equivalence themselves, which <a href="https://www.tandfonline.com/doi/full/10.1080/09617353.2024.2343959">can be an arduous process</a>.<h2>An Architecture for System-Level Safety</h2>Regardless of the chosen silicon, robust functional safety design should include fault detection at both the processing core and the power supply.To establish such a redundant design, engineers can implement a coordinated MCU-PMIC architecture with a continuous, bidirectional safety communication channel between the two devices over I2C or SPI. In this design, neither device relies solely on self-monitoring to detect and respond to failures or anomalous behavior. Instead, the MCU can detect power anomalies reported by the PMIC or the PMIC can detect processor failures and act independently, giving the system two independent observers.There are a number of fault-detection and response features that can add extra layers of safety to this architecture. For example, question-and-answer watchdog handshakes, which are external checks that software-only watchdog timers can’t replicate, enable the PMIC to verify proper MCU software execution. Similarly, if the PMIC has safe-state output capabilities, designers can configure it to signal external hardware directly in a fault scenario, thereby putting the system into a safe state. These systems can benefit doubly from processors with lockstep core architectures, in which multiple internal cores execute the same instructions simultaneously. By using a hardware comparator to monitor the core outputs in real time, lockstep processors can force a safe state if the core results diverge before any unsafe output reaches the system. Similarly, features like Memory Built-In Self-Test (MBIST) and Logic Built-In Self-Test (LBIST) let the system verify its diagnostic hardware. In that way, designers get a deeper level of confidence knowing that even the fault-detection systems themselves are functioning correctly. <img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785439356737-1785439356737.jpeg" width="624" height="351" class="fr-fic fr-dii" />Source: NXP<h2>Silicon-Level Safety With NXP</h2>With MCX E, customers get a qualified industrial part for SIL 2 under IEC61508. For higher safety needs where multicore or lockstep core is required, the S32K344 cluster of microcontrollers can help to close the gap by leveraging its ASIL D support.To reach these safety levels, NXP is certified by TÜV-SUD for functional safety management for safety-related projects up to ASIL D according to ISO 26262:2018 and up to SIL 3 according to IEC 61508:2010 following our <a href="https://www.nxp.com/webapp/Download?colCode=ISO-26262">BCaM process</a>. Importantly, NXP offers both microcontrollers with complete SEooC documentation, failure modes, effects, and diagnostic analysis (FMEDA), and a safety manual, meaning industrial designers can demonstrate IEC 61508 equivalence with only system-level analysis.With NXP’s MCX E and S32K3 family, industrial and robotics designers can achieve high levels of compliant functional safety using industrial-grade and automotive-grade parts.For example, the S32K3 features dual, 320 MHz lockstep-configurableArm® Cortex®-M7 cores, with an integrated hardware comparator to force safe states if core outputs diverge. The company also equipped this MCU with a Fault Collection and Control Unit (FCCU), a centralized hub that aggregates all internal and external error signals into a single, deterministic fault-handling pathway.NXP designed three diagnostic mechanisms into the FCCU. <ul><li>For memory faults, the FCCU monitors all SRAM and Flash memory using Error Correction Code (ECC) and Single-Error Correct, Double-Error Detect (SECDED) schemes, catching memory degradation before it produces unsafe outputs. </li><li>For logic and permanent faults, the system runs MBIST and LBIST at startup to verify the diagnostic hardware. </li><li>For core processing faults, Structural Core Self-Test (SCST) covers permanent hardware faults in the Cortex-M7 cores that transient-fault detection alone would miss. </li></ul>With these three features, designers can achieve high level Safe Failure Fraction during system-level analysis, which supports up to SIL 3.In addition to prioritizing standalone capabilities, NXP designed the S32K1 to pair with the FS23 safety PMIC and the S32K3 to pair with the FS26 and FS27. When used in conjunction, they offer question-and-answer watchdog handshakes and continuous FCCU monitoring to detect failures that the MCU can’t report on itself. Together, these devices unlock a dual-channel hardware architecture that satisfies PL e requirements before engineers write a single line of application code.<a href="https://www.nxp.com/design/design-center/development-boards-and-designs/frdm-development-boards:FRDM" target="_blank" rel="noopener noreferrer"><img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785439356749-1785439356749.png" width="819.1780821917807" height="460.0816479752236" class="fr-fic fr-dii" />FRDM Boards. Source: NXP</a><h2>Getting Started with FRDM Boards</h2>To help designers quickly get started with safety-first architectures based on automotive silicon, NXP offers FRDM development boards for both the <a href="https://www.nxp.com/products/processors-and-microcontrollers/s32-automotive-platform/s32k-auto-general-purpose-mcus:S32K-MCUS" target="_blank" rel="noopener noreferrer">S32K family</a> and its companion PMICs. A completely verified and documented hardware solution, the FRDM development board lets teams instantly implement and validate MCU-PMIC coordination, lockstep fault detection, and FCCU fault routing without building any custom hardware.In addition to the hardware, NXP also offers robust software tools that are ready for use out of the box. Pre-coded and validated examples are available through NXP's Application Code Hub, and the boards support the S32 Design Studio integrated development environment.  <h2>Safety Without Compromise</h2>As industrial and robotics designers strive for higher safety integrity levels, automotive-grade solutions offer a promising path forward. With the S32K3 family, they inherit NXP's proven ISO 26262 safety framework and provide work products helping customers to integrate ICs in IEC 61508 systems. Designers can shift their focus to what matters most: building the next generation of safer, higher-performance industrial and robotic systems.

Industrial and robotics designers strive for higher safety integrity with the NXP S32K microcontroller.

Extending Automotive-Grade Functional Safety into Industrial Systems with NXP S32K Microcontrollers

An instrumentation amplifier is a precision differential amplifier designed to amplify small voltage differences. This guide explains the three-op-amp architecture, and how to choose between INA128, AD620, LT1167, and a standard difference amplifier for sensor front-ends.

Instrumentation Amplifier: Theory, Three-Op-Amp Design, CMRR

A new edge AI processor becomes a practical design option once engineers can evaluate it inside a real system architecture. Engineers need to connect it to the host, define its power architecture, account for its package, and determine whether it fits the electrical, mechanical, and manufacturing constraints of the product.To support that evaluation, verified schematic symbols, PCB footprints, and 3D ECAD models for <a href="https://brainchip.com/wp-content/uploads/2025/10/AKD1500-Product-Brief-V2.4-Oct.25.pdf" target="_blank" rel="noopener noreferrer">BrainChip’s AKD1500 Akida chip</a> will become available across more than 25 design-tool formats, including Altium Designer, KiCad, and Autodesk Eagle. The AKD1500 is a neuromorphic edge AI co-processor designed to accelerate low-power inference alongside a host CPU using PCIe or microcontroller using SPI for event-based computation.This BrainChip-Supplyframe partnership gives engineers a more direct route from studying the AKD1500’s architecture to placing the device within a schematic and PCB layout. With the component definition already available, teams can focus earlier on the factors that determine if neuromorphic acceleration is suitable for the system, including host connectivity, board area, and prototype requirements.<a href="https://brainchip.com/brainchip-akd1500-neuromorphic-co-processor-chip-is-now-discoverable-in-supplyframe-design-modeler-software" target="_blank" rel="noopener noreferrer"><img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785343776936-BrainChipPR1.png" style="width:100%px" class="fr-fic fr-dib" />Read the full press release from BrainChip and Supplyframe.</a><h2>Moving a Processor from Datasheet to Design</h2>Before a processor can be assessed at system level, its datasheet information must be translated into design assets the ECAD environment can use. The schematic symbol defines how engineers interact with the device electrically, the footprint determines how it will be mounted and routed on the PCB, and the 3D model establishes its physical position within the assembled product.Creating these assets manually requires engineers or component librarians to interpret pin tables, package drawings, pad dimensions, pitch, orientation markers, and mechanical tolerances. Each value must then be entered correctly into the selected ECAD platform and checked against the manufacturer’s documentation.This work is especially consequential for a processor with numerous power, ground, control, and data connections. A symbol with an incorrect pin assignment, for instance, can propagate an electrical error into the schematic. Similarly, a footprint with the wrong pad geometry or package orientation can create assembly problems, while an inaccurate 3D model can lead to conflicts with an enclosure, heatsink, connector, or neighboring component.Some errors are identified during schematic review or design-rule checking; others remain undetected until PCB fabrication, assembly, or prototype bring-up, when correcting them becomes substantially more expensive. Even an accurate manually created model must often be reviewed internally before a design team is willing to use it in production work.<a href="https://supplyframe.com/designsense" target="_blank" rel="noopener noreferrer">Supplyframe’s DesignSense Models</a> service moves this preparation upstream by providing verified schematic symbols, PCB footprints, and 3D models in formats compatible with widely used ECAD tools. As engineers evaluate the AKD1500, the verified files establish a common component definition that can be used across schematic capture, board layout, mechanical review, and design collaboration.Engineers remain responsible for designing and validating the surrounding power, clocking, communication, and host-interface circuitry. The verified assets provide a reliable starting point for that work, allowing technical evaluation to begin with the device placed inside the intended hardware architecture.<h2>Adding Neuromorphic Acceleration to an Existing Host</h2>The AKD1500 is an edge AI co-processor based on BrainChip’s Akida neuromorphic processing architecture. It operates as a dedicated accelerator connected to an x86, Arm, or RISC-V host through PCI Express or lower-power serial interfaces.This allows engineering teams to add dedicated AI acceleration to systems built around an existing CPU or microcontroller architecture. BrainChip positions the Akida processor for applications including industrial sensing, robotics, medical devices, wearables, smart cameras, and other systems constrained by power consumption and heat dissipation.According to BrainChip, the AKD1500 delivers up to 800 effective GOPS and operates at less than 1 mW per GOP. The processor is manufactured on GlobalFoundries’ 22FDX platform and is intended for sub-watt edge inference workloads.Its architecture uses event-based computation and exploits sparsity within neural network activations and kernels. Computation is triggered by relevant activity rather than requiring every element in a dense representation to be processed during each operation. Reducing unnecessary data movement and arithmetic activity can lower the energy required for workloads that contain significant sparsity.Akida also supports on-chip learning for applications requiring local adaptation. BrainChip describes this capability in terms of one-shot and few-shot learning, where a deployed system can learn from a limited number of examples without transferring all sensor data to a remote training environment. Potential uses include personalization, anomaly recognition, and adaptation to changing operating conditions.Model development and deployment are handled through BrainChip’s MetaTF software environment. Developers can work from TensorFlow, Keras, and PyTorch workflows, then convert, optimize, quantize, compile, and deploy supported models to the Akida hardware.<h2>Evaluating Neuromorphic Hardware Earlier in the Design Cycle</h2>Making the AKD1500 available inside the ECAD workflow allows engineers to evaluate the device in the context of the complete product architecture. This is particularly relevant for edge AI systems, where processor selection affects power budgets, thermal limits, board area, sensor interfaces, host performance, and mechanical design.Earlier access to verified design data also makes feasibility studies more representative. Engineers can assess how the co-processor fits alongside the host, estimate the supporting circuitry and routing effort, and identify potential layout or enclosure constraints before committing to a prototype. This gives teams a clearer view of the integration cost associated with neuromorphic acceleration alongside the processor’s performance and efficiency specifications.The shared component definition can also improve coordination across electrical, PCB, mechanical, and manufacturing teams. Each discipline can work from the same verified package and design data as the architecture develops, reducing the risk of inconsistencies between schematic capture, board layout, mechanical modeling, and production review.BrainChip’s AKD1500 ECAD files will be available through Supplyframe’s component-search and schematic-selection workflow, as well as through BrainChip’s website. Products with available models will carry a verified-model badge, and engineers will be able to download the format required by their ECAD environment. According to the partnership announcement, this is intended to reduce the time between component discovery and schematic placement from hours to minutes.<h2>Lowering the Evaluation Barrier for Neuromorphic Hardware</h2>The practical value of the partnership lies in allowing engineers to evaluate a specialized AI architecture through the same component-selection and PCB design processes used for the rest of the hardware platform. Verified ECAD assets remove an early library-development task and allow teams to direct more effort toward power integrity, interface timing, thermal behavior, software compatibility, and workload performance.As teams explore neuromorphic acceleration, the AKD1500 can now enter the design workflow alongside the host processor, sensors, memory, and interface components. This brings the technology into the stage where architectural assumptions can be tested against the physical and electrical requirements of a deployable edge AI product.<a href="https://brainchip.com/brainchip-akd1500-neuromorphic-co-processor-chip-is-now-discoverable-in-supplyframe-design-modeler-software" id="isPasted">For more details on the partnership and the available AKD1500 design models, read the full press release</a> from BrainChip and Supplyframe.

BrainChip has partnered with Supplyframe to make its AKD1500 neuromorphic edge AI co-processor available as a design-ready component within the Supplyframe electronics design ecosystem.

Making Neuromorphic Edge AI Design-Ready with Verified ECAD Models

Every embedded design needs a clock, but not every clock needs the same source. This guide compares bare quartz crystals against packaged crystal oscillators, including TCXO, OCXO, VCXO, and MEMS variants, to help engineers pick the right timing component.

Crystal vs Oscillator: Choosing the Right Timing Component

When talking about large language models (LLMs), people usually imagine a general-purpose assistant: something that can answer questions about weather, politics, software, history, travel, cooking, electronics – and almost any other topic. The model is expected to know a little bit about everything, follow open-ended conversations, and respond to a very broad range of prompts. That’s the experience most of us are used to, since cloud-based AI tools have become so widespread.Embedded systems work in a much more “narrow” world. A robot does not need to discuss politics, an inspection system does not need to suggest vacation destinations, and a maintenance assistant installed near a machine does not need to explain ancient history. The system needs to understand the device, the task, the possible commands, the local data, and the actions that are safe to suggest or execute. The goal is to give an edge device enough language intelligence to become more useful, more understandable, and more independent from the network. <iframe id="6a3c254c6d19a31976afd60b" width="250" class="specs-iframe" height="440" src="https://wevolver.com/iframe/specs/arduino-uno-q-4gb?iframe=true" frameborder="0" scrolling="no"></iframe>This is the framework in which we can think about local LLMs on <a href="https://www.arduino.cc/product-uno-q?utm_source=content&amp;utm_medium=blogpost&amp;utm_campaign=unoq_is_the_answer&amp;utm_id=mktg-content" target="_blank">UNO Q</a>: a practical platform to explore this idea because it brings together a Debian Linux environment and the Arduino® hardware ecosystem. The Linux side can run local AI tools, command-line workflows, Python applications, web services, and inference runtimes. The Arduino side connects that intelligence to sensors, actuators, shields, <a href="https://store.arduino.cc/pages/modulino" target="_blank">Arduino</a><a href="https://store.arduino.cc/pages/modulino" target="_blank">®</a><a href="https://store.arduino.cc/pages/modulino" target="_blank"> Modulino</a><a href="https://store.arduino.cc/pages/modulino" target="_blank">™</a><a href="https://store.arduino.cc/pages/modulino" target="_blank"> nodes</a>, and real-world signals. This combination makes it possible to experiment with language models not as isolated chatbots, but as part of real embedded workflows. The most important question to consider is not how to force a large model to run, but what kind of useful intelligence can live close to the data, close to the device, and close to the physical action?<h2>Step 1: choose the right model for your use case</h2>The edge is where smaller, optimized models become interesting. On the cloud, a large general-purpose model makes sense because it is expected to answer almost anything. On the edge, a model that has been trained, fine-tuned, distilled, or quantized for a specific domain can be more practical. It carries less unnecessary weight, focuses on the type of language the device actually needs, and can be integrated into a controlled application flow.For example, in robotics the interaction can often be reduced to a limited set of useful instructions: move forward, stop, inspect this object, return to base, report battery level, explain the last error, switch to manual mode. The model can help interpret natural language, but the system should still map that interpretation to a controlled set of valid commands. This makes the behavior easier to test, easier to validate, and easier to trust.That narrower scope is one of the reasons local LLMs can make sense on embedded platforms.<h2>Step 2: understand your memory and storage constraints</h2>A large language model usually has many parameters, and every parameter represents data that must be stored, loaded, and processed during inference. Model weights are only part of the story. During generation, the runtime also needs working memory for the prompt, the intermediate computation, and the key-value cache used by transformer models to keep track of previous tokens. As the context grows, memory usage grows too.A 1B-parameter model in 4-bit quantization (such as Llama 3.2 1B Q4) occupies roughly 600–700 MB on disk and requires around 1 GB of RAM at runtime, including the KV cache for a short context window. A 3B model at the same precision pushes past 2 GB. These are numbers that matter on a board with fixed memory and storage, where the model must coexist with the OS, the runtime, and the rest of the application.Quantization is one of the techniques that makes this more realistic. Instead of storing model weights with high-precision numerical values, a quantized model uses lower-precision representations. This reduces memory usage and can make inference possible on hardware that would otherwise be too constrained. In practical terms, quantization helps move a model from “too large to run locally” towards “small enough to experiment with” – while accepting a trade-off in accuracy, fluency, or speed depending on the model and runtime. Model distillation is another important concept. In simple terms, distillation is a training approach where a smaller model learns from a larger teacher model. The goal is to keep useful behavior while reducing inference cost and memory footprint. A distilled model will not have the full breadth of the teacher, but it can be much more suitable when the application needs a focused capability on-device.This example of <a href="https://projecthub.arduino.cc/marc-edgeimpulse/running-local-llms-and-vlms-on-the-arduino-uno-q-with-yzma-74e288" target="_blank">running local LLMs and VLMs on UNO Q with yzma</a><a href="https://projecthub.arduino.cc/marc-edgeimpulse/running-local-llms-and-vlms-on-the-arduino-uno-q-with-yzma-74e288" target="_blank"> </a>expands the conversation beyond text chat and explores local LLM and VLM workflows using yzma and llama, pointing toward a wider class of edge AI experiments where language models can work together with images, local data, and device context.<h2>Step 3: identify where a local LLM adds real value</h2>Local LLMs become even more useful when they are combined with other edge workflows. OCR is a good example. A camera connected to an UNO Q may extract text from a label, display, document, or machine interface. A compact language model can then summarize that text, classify it, or turn it into a structured response. The model only needs to process the relevant context, which keeps the workflow lighter and more focused.The same principle applies to an UNO Q that collects logs, sensor readings, error states, or system events. A local model can turn that information into a short human-readable summary directly on the device. For a technician, this can transform raw data into something immediately useful – a compact explanation of the current status or a short description of the last error condition.<h2>Step 4: design the architecture and set your boundaries</h2>One of the most practical ways to think about local LLMs on UNO Q is to treat the model as an occasional reasoning layer. It can be called when language understanding, summarization, or interpretation adds value. Fast control loops, continuous monitoring, and timing-critical actions remain better suited to deterministic software running on the appropriate side of the system.When working with local LLMs on UNO Q, developers should take into consideration a few practical parameters. Memory usage comes first: the model must fit comfortably together with the runtime and the rest of the application. Response latency comes next: a model that runs may still feel too slow if the use case expects instant answers. Storage should also be planned carefully, because model files and dependencies can be large. The best entry point is the Arduino Project Hub tutorial <a href="https://projecthub.arduino.cc/robuinlabs/local-llm-ai-chatbot-on-arduino-uno-q-043aa9" target="_blank">Local LLM AI Chatbot on UNO Q</a>, which walks through installing a small LLM and running it offline. It is a useful starting point because it demonstrates the basic shape of a local LLM applicationThere is also a natural bridge toward local agents. Agentic workflows can move beyond a simple chat interface and start coordinating tools, files, scripts, and actions. On UNO Q, this direction is especially interesting when the agent is treated as an orchestrator on the Linux side. It can inspect logs, prepare files, call scripts, interact with local tools, or help drive development workflows, while the hardware-facing layer keeps direct control over physical I/O.This kind of setup requires clear boundaries. Giving an agent access to tools means giving it the ability to change things, so the environment should be designed carefully. A dedicated board can be a useful sandbox for this type of experimentation, with limited credentials, limited data access, and a specific set of allowed tools. This makes it possible to explore agentic workflows while keeping the system understandable and controlled.If you prefer a familiar developer workflow, <a href="https://blogm.tinivelli.com/installing-ollama-on-arduino-uno-q-d7b63a12b1e9" target="_blank">Installing Ollama on Arduino UNO Q</a> covers a practical detail that matters a lot on embedded Linux systems: how to efficiently manage the resources available on the UNO Q to get the most out of it.<h2>Step 5: run it, measure it, iterate</h2>Pick one model, run it on the board, and pay attention to memory usage and response time for your specific prompt. That real-world data will tell you more than any benchmark – and it will give you a much clearer picture of where a local LLM fits in your next embedded project.Local LLMs on UNO Q always balance power, cost, size, latency, privacy, reliability, and connectivity. The most interesting question is how much useful intelligence can be placed close to the data, the hardware, and the user. Because Edge AI is not about more power. It’s about smarter choices. With the right model, the right architecture, and the flexibility of UNO Q, you can test local AI where it matters most: on real hardware, in real projects.Start building with <a href="https://store.arduino.cc/products/uno-q-4gb?utm_source=content&amp;utm_medium=blogpost&amp;utm_campaign=unoq_is_the_answer&amp;utm_id=mktg-content" target="_blank">UNO Q</a> and bring your AI ideas closer to the real world.<pre>Arduino and UNO, and the Arduino logo are trademarks or registered trademarks of Arduino S.r.l.</pre>

The Arduino® UNO™ Q board is designed to bring AI to the edge. Unlock its potential to access a huge variety of applications.

Running local LLMs on embedded systems: a practical guide

As integrated photonics expands into quantum systems, UV sensing, LiDAR, optical communications, and advanced spectroscopy, existing photonic integrated circuit (PIC) platforms are running into material limitations.Silicon photonics and silicon nitride, while widely adopted, are non-transparent below roughly 400 nm, restricting their use in ultraviolet and several quantum applications. Indium phosphide supports active optical functionality but comes with higher fabrication complexity. In many cases, system developers are forced to combine multiple material platforms to achieve the required optical performance, increasing cost, packaging complexity, and integration overhead.These challenges are also relevant for teams participating in the <a href="https://www.wevolver.com/article/global-photonics-engineering-contest-2026?utm_source=article&amp;utm_medium=aluvia&amp;utm_id=Photondelta" target="_blank" rel="noopener noreferrer">Global Photonics Engineering Contest 2026</a>, where application requirements often extend beyond the limits of certain PIC platforms.<h2>Aluminum Oxide (AlOx): A Platform Built for the Full Spectrum</h2>Aluminum oxide (AlOx) is emerging as an alternative platform designed to address these gaps. The material offers one of the broadest optical transparency windows in integrated photonics, spanning from approximately 200 nm in the ultraviolet to 3 μm in the mid-infrared. It also supports ultra-low propagation losses, high optical power handling, and rare-earth doping for on-chip optical amplification, capabilities not simultaneously available in conventional silicon-based photonic platforms. In addition, AlOx remains compatible with CMOS manufacturing workflows and scalable wafer processing.Aluvia, a leading AlOx PIC manufacturer, offers a unique platform spanning UV to mid-infrared wavelengths. Through Multi-Project Wafer (MPW) runs, a Luceda-based Process Design Kit (PDK), and dedicated wafer production, Aluvia makes AlOx more accessible. In the following discussion, Steve Stoffels, the CEO of Aluvia, shares how AlOx compares to existing PIC platforms and where it fits best.<img src="https://images.wevolver.com/eyJidWNrZXQiOiJ3ZXZvbHZlci1wcm9qZWN0LWltYWdlcyIsImtleSI6ImZyb2FsYS8xNzg1MzIwMzc1NTU0LTE3ODUzMjAzNzU1NTQuanBlZyIsImVkaXRzIjp7InJlc2l6ZSI6eyJ3aWR0aCI6OTUwLCJmaXQiOiJjb3ZlciJ9fX0=" width="741" height="185" class="fr-fic fr-dii" />Peter (PhotonDelta): Aluvia specializes in AlOx-based PICs. What differentiates the platform technically, and which application domains benefit most from its properties?Steve (Aluvia): Aluminum oxide stands apart from other photonic platforms through a combination of properties that no single competing material can fully replicate. Its optical transparency window spans from 200 nm in the ultraviolet all the way to approximately 3 μm in the mid-infrared, a range that exceeds what silicon photonics, silicon nitride, or indium phosphide individually offer. This is paired with exceptionally low propagation losses: around 5 dB/m at 1550 nm in the near-infrared and approximately 1 dB/cm at 369 nm in the UV. AlOx also handles optical power significantly better than silicon nitride, up to an order of magnitude higher, and exhibits superior linearity. Critically, AlOx is amenable to rare-earth doping, particularly with Erbium (Er³⁺), which enables on-chip optical amplification. This is a capability that silicon-based platforms cannot offer, and it opens up a category of applications like integrated optical amplifiers and active photonic circuits that are structurally inaccessible to other passive platforms. The application domains that benefit most include:<ul><li>Optical communications, where on-chip Erbium-doped waveguide amplifiers (EDWAs) can replace bulky fiber amplifiers</li><li>Quantum technologies, where UV transparency is essential for ion and atom trapping at the heart of next-generation optical clocks and quantum computers</li><li>LiDAR systems, where high power handling enables compact solid-state beam steering</li><li>Augmented reality, where efficient transmission in the blue wavelength range is critical for high-brightness display engines</li><li>Integrated UV lasers for spectroscopy, biomedical applications, and advanced manufacturing.</li></ul>Peter (PhotonDelta): Tell us something about your products and services. How do your building blocks and product modules integrate with customer-specific architectures? Are they intended as drop-in solutions or as starting points for customization?Steve (Aluvia): Aluvia offers a layered product portfolio designed to take teams from early-stage R&amp;D through volume manufacturing without forcing re-architecture at each step. Our MPW program allows teams to prototype cost-effectively through shared wafer runs, with quarterly cadence and two passive AlOx stack options: 100 nm for UV-visible applications and 400 nm for near-infrared. Design areas of 5×10 mm² or 10×10 mm² make the program accessible to research groups, startups, and university teams. For teams moving toward production, Dedicated Wafer Runs offer full-volume fabrication with stack-level customization, including active, rare-earth-doped AlOx layers, on both 100 mm and 200 mm wafer platforms. Our PDK, available through the Luceda Photonics Design Platform, gives designers a familiar and well-supported environment for working with AlOx from day one.<a href="https://www.wevolver.com/article/global-photonics-engineering-contest-2026?utm_source=article&amp;utm_medium=aluvia&amp;utm_id=Photondelta" target="_blank" rel="noopener noreferrer"><img src="https://images.wevolver.com/eyJidWNrZXQiOiJ3ZXZvbHZlci1wcm9qZWN0LWltYWdlcyIsImtleSI6ImZyb2FsYS8xNzg1MzIwMzc1NTU5LTE3ODUzMjAzNzU1NTkucG5nIiwiZWRpdHMiOnsicmVzaXplIjp7IndpZHRoIjo5NTAsImZpdCI6ImNvdmVyIn19fQ==" width="741" height="259" class="fr-fic fr-dii" /></a>Peter (PhotonDelta): What are the key volume applications of AlOx? Where do you see the strongest commercial momentum for AlOx photonics over the next few years?Steve (Aluvia): Near-term volume opportunities are concentrated in optical communications and quantum technologies. In communications, demand for integrated on-chip amplification is accelerating as AI-scale datacenters push the limits of conventional optical interconnects and their amplification solutions. AlOx-based EDWAs offer gain in the C-band in a fully integrated, wafer-scale format. We see strong commercial pull from the high-performance networking and datacenter connectivity segments, where power efficiency and miniaturization are rapidly becoming as important as raw throughput. In quantum technologies, AlOx’s UV transparency gives it a structural advantage for ion trap and atom trap photonics, a segment growing rapidly as quantum computing and sensing hardware matures from laboratory demonstrations toward deployable systems. LiDAR for mobility, robotics, and environmental sensing represents a third strong vector, where AlOx’s power handling enables the compact solid-state architectures next-generation systems require. Further out, augmented reality optics and chip-integrated UV lasers for spectroscopy and biomedical applications represent emerging markets that AlOx is uniquely positioned to serve.<img src="https://images.wevolver.com/eyJidWNrZXQiOiJ3ZXZvbHZlci1wcm9qZWN0LWltYWdlcyIsImtleSI6ImZyb2FsYS8xNzg1MzIwMzc1NTY0LTE3ODUzMjAzNzU1NjQuanBlZyIsImVkaXRzIjp7InJlc2l6ZSI6eyJ3aWR0aCI6OTUwLCJmaXQiOiJjb3ZlciJ9fX0=" width="741" height="493" class="fr-fic fr-dii" />Close-up fabricated AIOx PIC diesPeter (PhotonDelta): Could you share any recent or upcoming projects that Aluvia is working on?Steve (Aluvia): Our project portfolio reflects both the breadth of the AlOx platform and the growing European commitment to deep-tech integrated photonic infrastructure. Extended cavity UV laser, Nature Communications (2025), published in November 2025, is a landmark paper from the University of Twente that demonstrated the first integrated, extended cavity diode laser built entirely from UV-transparent materials, combining AlOx waveguide circuits with gallium nitride amplifiers. The laser achieves milliwatt-level on-chip output power in the violet range (403–408 nm), a record-low intrinsic linewidth of around 300 kHz, and better than 43 dB side-mode suppression. Critically, it tunes mode-hop-free to atomic transition frequencies, such as the strontium transition, which is precisely what portable optical clocks and ion-based quantum computers require. This result validates AlOx as the enabling material for a new class of UV-integrated lasers and directly underpins Aluvia’s quantum-facing product roadmap. P4Q, Photonics for Quantum: Aluvia is a partner in P4Q, a €50 million European quantum pilot line coordinated by the University of Twente and co-funded by the EU’s Chips Joint Undertaking and twelve national governments. P4Q bridges the gap between laboratory quantum photonics and industrial manufacturing, developing process and assembly design kits across multiple platforms, with AlOx as a key pillar. For Aluvia, P4Q is a critical step toward positioning AlOx as a production-ready platform for the quantum industry, embedded within a structured European supply chain. femto-i-Comb, EIC Transition: Building on the European Innovation Council (EIC) Pathfinder project FEMTOCHIP, femto-i-Comb aims to deliver the world’s first fully stabilized, integrated femtosecond laser-based frequency comb. By combining on-chip continuum generation and carrier-envelope locking, the project creates ultra-stable optical and microwave sources for applications spanning high-resolution environmental and health sensing, LiDAR, RADAR, and next-generation signal generation. Aluvia contributes its AlOx waveguide platform alongside partners including DESY, LIGENTEC, and Cycle Lasers. Qu-PIC, Horizon Europe RIA: The Quantum Universal PIC platform project has selected AlOx as its backbone technology for its low propagation losses and spectral coverage. Qu-PIC targets a comprehensive quantum PIC building block library: tunable lasers at 399 nm, 411 nm, and 935 nm; UVC external cavity lasers at 280 nm; squeezed photon sources; and single-photon detectors. Two demonstrators are in development, a source of GKP states for quantum processing and a ytterbium-ion-based atomic clock, all within a framework explicitly aimed at securing European quantum photonic supply chain sovereignty. PHANTOM, Photonic Gas Sensing: PHANTOM applies AlOx photonics to accurate identification and concentration measurement of gases for safety, sustainability, and efficiency. Target applications include detecting toxic hydrogen cyanide during manure storage, monitoring ammonia leaks in combustion engines and hydrogen fuel cells, and tracking ethylene in fruit storage facilities. Gas detection is achieved through laser absorption spectroscopy, where AlOx’s broad spectral window and ultra-low-loss waveguides are decisive advantages. The consortium, including OnePlanet Research, Radboud University, Bronkhorst, Epiphany, and Sensor Sense, kicked off in December 2025.Peter (PhotonDelta): If a team is considering an AlOx-based design, what technical constraints should they carefully evaluate?Steve (Aluvia): Aluvia’s AlOx technology provides a powerful and versatile platform for integrated photonics, but teams should approach it with a clear-eyed understanding of where it fits and what to plan around. The most important upfront decision is wavelength range. AlOx stacks are optimized for specific spectral windows. Our 100 nm layer targets UV-visible applications, while the 400 nm layer is designed for near-infrared. Selecting the right stack at the outset is essential: switching later means a new design and a new fabrication run, with both time and cost implications. Early-stage clarity on the target wavelength and integration architecture is therefore critical to fully take advantage of AlOx’s strengths. AlOx is an advanced photonic platform particularly well suited to low-loss light guiding, splitting, filtering and on-chip optical amplification across a broad wavelength range. For applications that also require high-speed electro-optic modulation or electronic integration, AlOx can be combined with complementary technologies through a hybrid integration approach. Planning this architecture from the outset allows each material platform to be used where it delivers the greatest performance, while avoiding the compromises that can arise when all functions are forced into a single material system.Peter (PhotonDelta): How can participants in the PhotonDelta Global Photonics Engineering Contest 2026 leverage Aluvia’s platform and design ecosystem?Steve (Aluvia): Aluvia is an industry partner in the Global Photonics Engineering Contest 2026. Our MPW program is the natural entry point for contest teams: affordable, rapid prototyping through shared wafer runs with no commitment to a fully dedicated wafer at the concept stage. The Luceda PDK provides a structured starting environment, and our team is available to support from early design through tape-out. The AlOx platform suits the contest’s core application domains, offering capabilities unavailable on conventional platforms. We encourage teams to reach out early to discuss whether AlOx is the right fit. Aluvia’s role as an ecosystem partner means the winning team has a credible, supported path from concept to fabricated chip.<h2>Advancing Innovation in Integrated Photonics with AlOx</h2>The PhotonDelta <a href="https://www.wevolver.com/article/global-photonics-engineering-contest-2026?utm_source=article&amp;utm_medium=aluvia&amp;utm_id=Photondelta" target="_blank" rel="noopener noreferrer">Global Photonics Engineering Contest 2026</a> is a real opportunity, not just visibility. With a network of industry partners ready to help turn a concept into a working device, what might seem ambitious becomes achievable. The path runs from idea to platform to fabrication to application. With the AlOx ecosystem in place and Aluvia as a partner, the support to walk that path is already there.

Aluminum oxide PICs are gaining attention for their broad wavelength coverage, low optical losses, and support for on-chip optical amplification. In this discussion, Steve Stoffels, the CEO of Aluvia Photonics, explains where the platform fits and how teams can design with it.

Aluminum Oxide Photonic Integrated Circuits: Insights from Aluvia Photonics

Swiss machining excels at small, complex parts

What is Swiss machining? This guide explains how Swiss-type lathes work, how they differ from conventional lathes, and when to use them.

What Is Swiss Machining? Swiss-Type CNC Lathes Explained

The semiconductor industry is increasingly adopting a chiplet-based architecture as a practical solution to the growing complexities, costs, and scalability challenges associated with traditional monolithic integrated circuits.With the increasing price of advanced process nodes and longer design cycles, semiconductor companies are investigating new options for delivering higher performance, more flexibility, and faster product development. One of the most successful solutions has been the rise of chiplet technology, empowering complex systems to be built from multiple specialized dies in a single package.The critical enabler for this design shift is efficient communication between individual chiplets. High-performance/low-power die-to-die interconnects are essential to make a multi-chip package operate as a single system, not a set of separate components.This need has led to the advent of Universal Chiplet Interconnect Express (UCIe), an open industry standard that ensures high-bandwidth, low-latency, low-power communication between chiplets. <a href="https://forwardedgeasic.com/ucie-ip/">FEASIC’s UCIe IP solution</a> fulfills this need with a full implementation of the UCIe-SP standard package architecture, permitting developers to accelerate the deployment of interoperable chiplet-based systems.<h2>Chiplet Architectures on the Rise</h2>For decades, semiconductor innovation meant putting more functionality on a single die. This has resulted in large performance gains but has likewise become increasingly complex to maintain. Manufacturing large monolithic chips is also difficult and can result in low yield, high fabrication cost, long development time, and high design complexity.Chiplet architecture provides a fundamentally different approach. Instead of building a single large die, engineers can split a system into smaller functional dies. Each of them can then be tuned for a particular job. Processor cores, AI accelerators, memory controllers, networking interfaces, and analog functions can be developed independently and then integrated into a common package.There are several advantages to this modular design philosophy. Smaller dies generally improve yields and reduce overall costs of production. Individual chiplets can also be reused across multiple products, which shortens development timelines and improves return on investment. Moreover, designers can combine technologies fabricated with different process nodes, so that each function can be realized with the most appropriate manufacturing technology. Each chiplet is allowed to progress at its own desired pace, which allows upgrading specific or critical parts of the design independently. The overall packaged product can then improve incrementally at a much lower cost.But the success of this approach depends on how well these chiplets can communicate with each other. Without a common communication framework, chiplet ecosystems are at risk of fragmenting due to proprietary interfaces and limited interoperability.This modular approach is also reshaping the broader chiplet ecosystem. Reusable components can now be developed once and integrated across multiple platforms, echoing how standardization and modular design transformed other segments of the technology industry.Rather than designing and fabricating every subsystem in-house, companies can source specialized chiplets from different suppliers and integrate them into application-specific solutions, fostering innovation, reducing development costs, and speeding up product launches. UCIe addresses these problems directly, offering a standardized foundation for chiplet interoperability.<img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785253943231-1785253943230.jpeg" width="602" height="339" class="fr-fic fr-dii" />Engineers collaborating on a manufacturing project. Source: AdobeStock.<h2>UCIe: The Foundation for Chiplet Interoperability</h2>UCIe was created to address the interoperability issue that comes with chiplet adoption. The standard establishes a common die-to-die communication framework that allows chiplets from different vendors, foundries, and process technologies to cooperate in the same package.UCIe defines standardized physical, protocol, and software interfaces that eliminate many of the barriers that have traditionally made heterogeneous integration difficult. Instead of being forced to develop custom interconnects for each design, semiconductor designers can now leverage a common ecosystem supported by multiple industry players.This is particularly true for emerging applications that must move massive amounts of data between compute, memory, and accelerator resources. Efficient communication between multiple functional blocks is a key requirement for artificial intelligence workloads, high-performance computing systems, data center processors, and advanced edge-computing platforms. UCIe allows such systems to achieve high bandwidth density with low latency and power consumption.UCIe extends proven high-speed interconnect design principles to the package level, easing integration and allowing existing design methodologies to carry over to chiplet-based systems.<h2>FEASIC UCIe IP Solution</h2><a href="https://forwardedgeasic.com/">ForwardEdgeASIC (FEASIC)</a> has built a complete UCIe IP implementation that offers the basic infrastructure required for standards-compliant die-to-die communications. Instead of just offering individual building blocks, the company offers the complete UCIe stack, including the physical layer (PHY), die-to-die adapter functionality, and protocol support.This off-the-shelf solution enables customers to incorporate UCIe connectivity into their designs without the huge effort required to engineer a custom interconnect architecture. With a pre-validated solution, development teams can innovate on features specific to the application, while lowering the risk associated with implementations.The physical layer has been designed on the GlobalFoundries 12LP+ process technology and tuned for performance and power efficiency. FEASIC implementation supports x16 connectivity and provides transfer rates of up to 16 GT/s per lane to meet the high aggregate bandwidth required by today’s data-intensive applications.A key differentiator is that the solution has already been validated in the laboratory. And as the chiplet ecosystem matures, proven implementations gain value, as they reduce uncertainty in system integration and shorten qualification timelines.<img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785253943239-1785253943239.jpeg" width="602" height="339" class="fr-fic fr-dii" />Futuristic cloud computing and edge-processing technologies. Source: AdobeStock.<h2>Future Outlook</h2>One of the biggest trends in modern chip design has been the move from monolithic semiconductor devices to modular chiplet architecture. As demand increases for artificial intelligence, machine learning, cloud computing, networking, and edge-processing applications, the ability to efficiently integrate multiple specialized dies in a single package will become increasingly important.UCIe provides the standardized foundation to turn this vision into reality. The standard enables high-speed, low-latency communication between chiplets, permitting developers to realize the full benefits of modular semiconductor design.FEASIC’s UCIe IP solution is a practical implementation of this technology, offering designers a validated and ready-to-deploy interconnect platform. It supports high-bandwidth communication, heterogeneous integration, and scalable multi-die architectures to help accelerate the industry’s transition to a more flexible and interoperable chiplet ecosystem.With semiconductor complexity continuing to grow, standardized die-to-die connectivity will be critical to the success of future generations of computing platforms. Solutions like FEASIC’s UCIe IP are helping to transition chiplets from an emerging concept to a core technology for next-generation semiconductor innovation.

The ForwardEdge ASIC (FEASIC) UCIe IP advances  the deployment of interoperable chiplet-based systems.

Accelerating the Adoption of Chiplet-Based Semiconductor Architectures

This article explores the LiFePO4 voltage chart, covering SOC and voltage for 3.2V, 12V, 24V & 48V systems, charging and BMS cutoffs.

LiFePO4 Voltage Chart: SOC & Voltages for 3.2V/12V/24V/48V Systems

After a closet cleanout, what options are there for recyling our old threads? Not many. Apart from bringing used clothes to a donation center, there is no process for recycling textiles like there is for bottles and cans. And, the average American throws out around 81 pounds of clothing each year. That amounts to more than 11 million tons of textiles that end up in the landfill or incinerator. But MIT engineers hope to cut down on the growing mountain of textile waste, with a new, recyclable yarn. The team has designed a yarn made from a form of plastic that is commonly used in milk bottles and grocery bags. The new yarn, which has a feel similar to traditional sewing thread, can be woven into stretchy, lightweight clothing. The researchers say that at the end of its use, a yarn-spun garment could be melted down and redrawn into new yarn, and then woven into new clothing or even cast into buttons, belt buckles, and other plastic accessories.To demonstrate the yarn’s recyclability, the researchers spun a spool of yarn, melted the yarn down, and respun it into new yarn, multiple times. They found that even after 10 cycles, the yarn was as strong and flexible as conventional thread. They envision the new yarn could be an alternative to elastic spandex-polyester or spandex-nylon yarns, which are spun from a combination of fibers that cannot be recycled together. The team’s new yarn, in contrast, is made from a specific combination of plastic materials that mimics the tough and stretchy properties of spandex yarns, while also being easily recycled. <img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785237112025-1785237112025.png" class="fr-fic fr-dib" />A comparison shows that, like typical commercial fibers, the new yarn is just as stretchy. Credit: Courtesy of the researchers “Eighty percent of textiles on the U.S. market currently contain some amount of spandex, which makes them nonrecyclable,” says Svetlana Boriskina, a research scientist in MIT’s Department of Mechanical Engineering. “There’s no widely adopted technology now that recycles textiles into textiles. With our new yarn, we hope to change that.”Boriskina and her colleagues have published the details of the new yarn in a study <a href="https://pubs.acs.org/amlcef/article-abstract/doi/10.1021/acsmaterialslett.6c00337/5232302/Polyethylene-Based-Thermo-Mechanically-Recyclable" target="_blank">published in the journal ACS Materials Letters</a>. MIT co-authors include first author SeongHyeon Kim, Duo Xu, Volodymyr Korolovych, Domingo Flores-Hernandez, Kaniz Moriam, and Daniel Braconnier.The core of the problemSpandex is a polyurethane-based synthetic fiber that is springy but not very strong. A thread of an elastic yarn is made from two parts: a spandex-based core, surrounded by a sheath of tough polyester or nylon. The combination of these materials gives elastic yarns their unique stretch and strength. But this same material mixture makes elastic yarns nearly impossible to recycle. Yarns would first have to be chemically treated to separate the polyester sheath from the spandex core. The polyester-based sheath material could then be melted down and reused. But there is no way to recycle the yarn as a whole, without chemical separation.“Even though chemical separation technologies exist, they add extra cost and complexity, and usually require toxic chemicals that are harmful to the environment,” Boriskina says. “That’s why most stretchy garments go to the dump.”In 2021, Boriskina’s group developed a new type of <a href="https://news.mit.edu/2021/plastic-bags-recycle-fabrics-0315" target="_blank">yarn made from polyethylene</a>. Polyethylene is the most common type of plastic in the world, used to make everything from grocery bags, water bottles, trash bins, and toys to industrial pipes and plastic sheeting. Polyethylene is a thermoplastic, meaning that it can be melted down and remade, and thus recycled. And yet, polyethylene had never really been considered as a textile. In their previous work, Boriskina and her colleagues showed they could spin yarn out of polyethylene, which they then wove into various garments. In those experiments, they focused on the yarn’s moisture wicking, stain-resisting, and cooling properties. Spaghetti yarnIn their new study, the group aimed to tailor polyethylene yarn to mimic the strength and flexibility of spandex; they also sought to demonstrate the yarn’s recyclability. They first looked for formulations of stretchy, polyethylene-based copolymers that resemble a spandex elastic core. Separately, they engineered polyethylene yarns that can act as the sturdier sheath. Looking through the scientific literature and combing through industrial reports, the team evaluated many chemical variations of polyethylene.“The chemical structure of polyethylene is like Christmas garland — a backbone of carbon, carbon, carbon, and also these dangling ‘decorations’ of hydrogen atoms or short branches with the same structure as a backbone,” Boriskina explains. “How these chains are arranged can change the properties of the whole structure.”“Polyethylene can give us a wide range of properties, depending on how you make it,” adds first author SeongHyeon Kim.For the yarn’s core, the team used one polyethylene-based resin that results in a more stretchy fiber. They chose a second, stiffer resin as the basis for the yarn’s sheath. The researchers obtained pellets of each resin from a chemical manufacturer, and then put each type of pellet through a process of fiber fabrication, first pouring them into a hopper, then heating the pellets to about 350 degrees Fahrenheit, past their melting temperature. The melted polyethylene was then drawn through small extruders to make hair-thin fibers. “You just melt it in a barrel with a heater, and then you extrude and spin it into fibers,” Kim says. “It’s like a spaghetti machine.”The team used an industrial yarn spinner to wind the sheath fibers around a core fiber to make the final, elastic yarn. Because both the yarn’s core and sheath come from the same chemical family of polyethylene, Boriskina says the materials do not have to be separated before recycling, in contrast to spandex-based elastic yarns. The new yarn can be melted as is, and reformed into new yarn or other plastic products.<img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785237185318-1785237185318.png" class="fr-fic fr-dib" />“There’s no widely-adopted technology now that recycles textiles into textiles. With our new yarn, we hope to change that,” says Svetlana Boriskina. These two SEM images compare a conventional elastic yarn made of PET-spandex (left) with a new type of recyclable made from two types of recyclable polyethylene. Credit: Courtesy of the researchers“Because they are exactly the same chemistry, they play nicely together,” she says. “That’s what makes this yarn very recyclable.”As a demonstration, the team twisted an elastic core-sheath yarn, then melted it down and re-spun it, 10 times. Each time, they tested the yarn’s mechanical properties by precisely stretching a thread and measuring the pulling force at which the thread eventually broke. From these tests, they found that the yarn’s recycled versions were just as strong as the original sheath yarn. These recycled yarns can now be used to make new stetchy yarns by twisting them around a newly spun elastic core.“Now we have something that can be knitted and woven,” Boriskina says. “That is the next stage.”The team says their new recipe for polyethylene yarn can be scaled up into industrial-sized spools. Just like conventional spandex fibers, it would take kilometers of yarn to weave a single textile. But once woven and used, the team envisions that a polyethylene garment could conceivably be dropped in a recycling bin and sent to a facility to be melted down and respun, enabling a more sustainable, circular fashion and textile economy. “Hopefully it will prevent the need for making more and more textile materials, because you can keep recycling a large portion of it,” Boriskina says. This work was supported in part by the DEVCOM Soldier Center through the U.S. Army Research Office, the Office of Naval Research Global via Tecnologico de Monterrey, and the MIT Portugal Program.

Melt it, spin it, use it again. This new recyclable yarn is as strong and stretchy as spandex-based yarns.

MIT engineers design recyclable elastic yarn

The extremely fast data-processing capabilities that power today’s telecommunications, large-scale AI data centers, quantum technologies and more, are enabled by photonic microchips – miniaturized semiconductors in which photons, rather than electrons, carry information through micrometer-wide waveguides. Using a computer algorithm, researchers at the <a href="https://seas.harvard.edu/">Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS)</a> and the <a href="https://mpl.mpg.de/" target="_blank">Max Planck Institute for the Science of Light</a> have developed three new functional components for photonic microchips that are each about 500 times smaller than conventional designs and offer a path toward higher-performance integrated light technologies. The results are published in <a href="https://www.nature.com/articles/s41467-026-73390-9" target="_blank">Nature Communications</a>.<img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785235912017-1785235912017.png" class="fr-fic fr-dib" />A photonic microchip, next to a 10-euro-cent coin for scale, containing hundreds of inverse-designed components. The close-up images show computer-designed nanostructures, wavelength splitters, mode sorters, and mirrors. The left illustration shows how such devices could be combined and densely packed into integrated photonic circuits on a chip. Credit: Tony Bi / MPLPhotonic microchips are made of many intricate parts, from grating couplers, which transfer light between fibers and waveguides, to ring resonators — tiny circular structures that temporarily store light and increase its intensity inside the chip. The design process for each component is time-consuming and tedious, with engineers typically starting with a familiar geometry and adjusting parameters as they go. A research team co-led by <a href="https://seas.harvard.edu/person/kiyoul-yang">Kiyoul Yang,</a> SEAS assistant professor of electrical engineering, used a technique called inverse design to take an unconventional approach. Flipping the traditional process upside down, they first specified what they wanted the light to do. A computer algorithm then searched a large space of possible nanostructures for that exact function. The shapes the algorithm finds often look like irregular patterns of holes and ridges, but they nonetheless precisely guide light in a chip area that’s a tiny fraction the width of a human hair. “Inverse design becomes practical when fabrication realities are built into the optimization,” said Yang, who co-led the work with Pascal Del’Haye, head of the microphotonics research group at <a href="https://mpl.mpg.de/news/article/beyond-human-intuition-algorithms-create-foundry-ready-photonic-circuits" target="_blank">Max Planck Institute</a>. “By including minimum feature sizes and robustness to manufacturing variations in the algorithm itself, we obtain designs that are not only compact but also compatible with a commercial foundry process.”<h2>Design and testing of devices</h2>The team applied inverse design to silicon nitride – a material with low optical loss that can also help produce clean, laser-like light of many different colors. Previous approaches have mainly focused on silicon and, more recently, on diamond, silicon carbide and lithium niobate. “Thick silicon nitride underpins most of the high-performance integrated photonics we work with, but until now its component library was limited to hand-engineered designs,” Del’Haye said. “These compact, computer-designed components are an important step towards more densely integrated nonlinear and quantum photonic circuits.”The team designed, fabricated and tested three classes of device: wavelength splitters, which separate the colors of light; spatial mode sorters, which separate light into spatial channels; and mirrors.As one of examples, inverse-designed mirrors, only a few micrometers across, reflect up to 98.5% of incoming light while blocking other spatial modes. Placed in pairs, they form on-chip optical cavities in which light bounces more than 100 times between the mirrors before escaping.<img src="https://s3-us-west-1.amazonaws.com/wevolver-project-images/froala%2F1785235936200-1785235936200.png" class="fr-fic fr-dib" />Progression of photonic design. The device structure evolves through successive iterations of the computational optimization process. Credit: Aditya Paul<h2>Future directions</h2>The next step is to combine the newly developed components with nonlinear optical circuits. In these building blocks, intense light circulating on a chip can generate optical frequency combs: precise sets of many evenly spaced colors of light used in precision measurement, telecommunications and quantum technologies.“Inverse design lets us define what we want light to do, and the optimization finds a structure that does it, often one no human would have drawn,” said Toby Bi, co-lead author of the study and researcher at the Max Planck Institute. “What is exciting is that the same framework can do three quite different jobs on the same chip: route light by wavelength, sort it by spatial mode, and act as compact mirrors that form on-chip optical cavities.”“<a href="https://www.nature.com/articles/s41467-026-73390-9" target="_blank">Inverse-designed silicon nitride nanophotonics</a>” was co-authored by SEAS graduate students Egemen Bostan, and Danxian Liu, SEAS fellow and MIT EECS graduate student Aditya Paul, and postdoctoral fellow Tianyi Zeng. The research was supported by the Defense Advanced Research Projects Agency (D23AP00251-00) and the Under Secretary of Defense for Research and Engineering (FA8721-05-C-0002). Device fabrication was carried out in part at the Harvard Center for Nanoscale Systems (National Science Foundation award No. 1541959).

Inverse design enables compact, foundry-ready silicon nitride photonic components

Algorithm-Designed Photonic Circuits Beyond Human Intuition

A typical cell pack of NMC batteries for an electrical vehicle

Understanding cell chemistry, safety, energy density, and manufacturing realism.

Solid-State Batteries vs Lithium-Ion Batteries

In this episode, we explore NASA’s successful test of a powerful new lithium-fed electric thruster that could one day power human missions to Mars. Learn how electric propulsion differs from traditional chemical rockets and why this breakthrough could make deep space travel faster, more efficient, and capable of carrying the massive payloads needed for astronauts. <hr /><iframe height="200px" width="100%" frameborder="no" scrolling="no" src="https://player.simplecast.com/be0eca65-df88-40f9-927d-8728007e5525?dark=true" class="fr-draggable"></iframe> <hr /> Farbod: Friends, welcome back to the Next Byte podcast. In the last week, SpaceX did IPO, but that's not even the coolest thing that's happened in space news in the past week because NASA just cracked a code on electric propulsion, which means that we're gonna be able to get to space faster, cheaper, and more practically than ever before. And yes, we can finally get to Mars. So if you've been waiting on buying your tickets, now's the time. Let's get into it.All right, friends, as you heard, today we're talking about propulsion. But before we get into today's propulsion topic, my god, tongue twisters already. It's gonna be a long one, we're gonna be talking about VTOLs. And this is actually a technical resource from today's sponsor, Mouser Electronics. So if you've been rocking with us for a while, you know we love Mauser. They kind of have the same mentality as us, that there's a lot of interesting things going on in the world. Especially given that they're at this intersection with like academia and industry, being one of the world's biggest electronic suppliers, they take whatever they're seeing, whatever they're learning, and deliver it to audiences like you. So one of those is about vertical takeoff and landing vehicles, VTOLs, right? And if you ever played Call of Duty in the two thousands like Daniel and I did, the Osprey was one of the more common military aircrafts that did the vertical takeoff and landing, and it was super cool. But as we think about the world today, VTOLs are becoming more and more common. How so? Drones, right? Like drone delivery is the big thing. Amazon wants to do it, Walmart wants to do it. And for the drones to be able to take off vertically instead of needing a runway, that's that's absolutely huge. However, there's this big elephant in the room being how do we like actually feel these things, right? The combustion engines that are used for the military aircraft obviously don't really make sense here.The world has changed, regulations have changed. We probably don't want more gas coming out of drones that are gonna be flying everywhere, dropping off packages. So electric power makes sense. Cool. What are we gonna do? Are we gonna use batteries? Well, technically we are today, right? But that has its own drawbacks, like the fact that after 20 to 30 minutes of operation, it kind of needs to spend the rest of the time on the ground being charged up again. So this article is all about what does this landscape look like? What alternatives are there? And this is gonna be a hot one already. It suggests that hydrogen fuel could be a great alternative because it has great power density, meaning that from the amount, you know, size, weight, whatever you put in, you can get a lot of power out of it.  Sorry, that's energy density. And then you have power density, which means that you can charge and discharge very quickly. The reason I'm saying it's a hot topic is because I think Elon Musk famously said hydrogen is for stupid people, something something along those lines. Am I paraphrasing here? It's like it's never gonna take off. But they're talking about how a hydrogen fuel cell approach will have less moving parts, it's quick for refills, and it has the energy density and the power density that we're looking for. So they're painting this future, this picture of what the future can actually look like. If you're interested about propulsion technology, aerospace, hydrogen fuel cells, and controversy, then you should definitely check out the article. It's gonna be in our show notes.Daniel: And if you want the verbatim Elon Musk quote, it's Tesla CEO Elon Musk calls hydrogen the most dumb thing I could possibly imagine for energy storage. So take that with a grain of salt and then go read this article and let us know what you think.Farbod: All right. Today's article. NASA, specifically NASA JPL, NASA Glen, and Princeton University. And I feel like we haven't talked about NASA in a while. And it's easy to forget about NASA given that there's so many interesting companies. I mean, SpaceX just went public. Blue Origin and SpaceX have been doing incredible work, and there's all these other rocket startups that are happening, which have kind of, I think, stolen the spotlight for young engineers.Are we young anymore? I don't know.  But all that to say, we haven't talked about NASA for a while. Regardless of us talking about them, they've been doing cool stuff in the background all this time. One of the big ones that they've been focusing on is electric propulsion. Before we get into that, I think it's worth giving some context about what chemical propulsion is, what electric propulsion is, and what does the landscape look like today? Is that fair?Daniel: Pretty much all of propulsion is, goes back to like Newton's laws of physics, right? To propel a mass forward and this mass forward being a spacecraft. If I'm sending something forward, I need to send an equal amount of energy backward to be able to force that thing to go forward. So all the basic principles of chemical rocket, electric propulsion, this new MPD thruster, all of those are, let me throw as much mass out the back as fast as I can to force my spacecraft to go forward. And they all have different methods of what the masses that's being thrown out of the back, what the propellant is, the thing that's speeding up that mass to throw it out the back, and then a different speed at which those things are exhausted out of the back. All of that adds up to how much energy can you emit out of the back of your spacecraft to force the rest of the spacecraft to fly forward. And the best example I can think of when it comes to propulsion is like, if you imagine you're sitting on a big giant sheet of ice and you're on skates and I've got a bunch of rocks in my pocket, I can throw a really hard, I can throw a boulder, I can push it really hard away from me and it will send the boulder flying backwards, but it will also send me flying forward. And that's the best example I can think of of like how this analogy works.All of these propulsion systems, chemical, electric, MPD, they're all different versions of that. Can I throw mass out the back to force my spacecraft to go forward? The only difference is where's that energy coming from and how fast and how much mass can you throw out the back? That changes how fast you can go forward.Farbod: Yep. And like traditionally, I think even to this day, the most common approach is the chemical format, right? I think it was the Saturn V rockets that use two tons of fuel, maybe solid state, maybe liquid, I don't remember, but I think the Raptor engines for SpaceX use liquid fuel, right? They're also b burning something to go up into space. The issue with that has always been that it's it's fine for getting out of Earth's orbit, but it's actually a lot of mass. Like that fuel is again two tons that you're sitting on to take with you. So, you know, if we're leaving Earth to just go do a quick tour around the moon and come back, it's like whatever. But if you have longer space missions, that starts to become a massive constraint of how far you can go, right? Just the principle of taking weight with you. So electric propulsion was kind of promising and the i idea started floating in the nineteen sixties. So it's not even that new, like this has been over sixty years in development. That, like you're saying, can we spit stuff out using electricity instead of requiring that chemical reaction? The answer has been yes, but it's just generally low amounts of thrust generated. Funny enough, the best stab at this that NASA or any lab I think in the world has been able to accomplish was NASA's Psyche mission. This was a drone that they sent out,  I think to Saturn? We we d we actually talked about it in episode fourteen, which is crazy.Farbod: Yeah, it was going to an asteroid Psyche 16. Yes, you're right. You're right. It was going to an asteroid. And even though the thrust is very small, over the course of however long it's been in flight, it ramped up to a hundred and twenty-five thousand miles per hour or kilometers an hour. So it's pretty impressive. Like the the acceleration is slow, but it can still reach the velocity that it needs to. Now, again, NASA has not stopped working on this. They've been tackling it in the background and they even say it in this article, you know, Mars has been in the back of their mind. And what they're talking about in this article is how they've been able to achieve a thruster that's 25 times the power of the one that went to Psyche 16. So the principle is still the same, you have a hot gas and basically an electromagnet that is pushing it out of the thruster. What is changing here is that this time around they're using vaporized lithium metal lithium vapor. There we go. Yeah, metalized lithium lithium vapor. Is that it? Yeah, that's it. So that burns hot enough, which allows them to I think five thousand degrees Celsius was the number they used to generate higher and higher levels of thrust that they've been able to before. The number they were able to achieve was an operation at 120 kilowatts, but they want to ramp it up to five hundred to five five hundred kilowatts to one megawatt is the goal that they're going for. And a spaceship that is actually able to make it to Mars would need two to four megawatts. So like they're slowly, slowly stabbing at the problem. And I think episode 14 was what four years ago at this point that we recorded? Twenty-five toDaniel:  No, I think this is exciting. One thing that I still like, I think I want to recap because it took me when we researching for this a couple of times to wrap my head around.  Just the walk from chemical to electric to MPD thruster. Chemical is pretty simple. It's like the most, I don't know, like the most stereotypical, yeah.You've got a bunch of fuel and you actually need an oxidizer there as well to help you burn the fuel. You basically create in a controlled explosion and the controlled explosion, the hot gas expanding from that explosion is the mass that's being ejected backwards to exit the nozzle and that creates thrust. That's great. But like you're saying, it typically requires up to 100 % of the weight of the spacecraft in fuel just to get out of the Earth's atmosphere and into space. If you want to then continue to use that for thrust in space, you need to carry a proportionally increased amount of fuel to be able to lift it off. You end up when you're trying to go to Mars, as an example, you end up in a situation where there literally is not enough fuel or there's not a strong enough booster to be able to carry all that fuel with us into space and drive us to Mars. The complete other end of the spectrum are these electric thrusters like the ones in that Psyche spacecraft. They use solar panels that generates electricity, they magnetically repel ions out of the back. That's great. Really, really mass efficient. Awesome continuous thrust can run for years and years and years with only a small amount of mass. The problem is the amount of thrust coming out of the back of that is about this comparable to the weight of a few sheets of paper. So it's like, great, the Psyche 16 spacecraft has sped up to this incredible speed due to this continuous thrust. But when it first started, it was probably crawling very, very, very slow. And that's because the thrust that's coming out of this very, very efficient electric thruster is very, very small. And so this NPD thruster is kind of like a best of both worlds scenario where you're not just emitting ions using electricity. You're actually heating them up with this massive electric current and the lithium, the metalized lithium actually becomes plasma.And then there's a magnetic field that accelerates that plasma and ejects it. It's much higher thrust because it enables them to accelerate a small amount of propellant very, very fast. So it's much faster than the propellant coming out of the back of the electric thrusters. That's where they're getting the extra force from. That's where they're getting the extra thrust. That's why engineers get excited about there being a lot more potential speed to these spacecraft. And this potentially being somewhere that's a happy medium between: let me blow up a bunch of rocket fuel or let me slowly emit ions like a drip out of the back of my spacecraft. This is like, can I control blow up using electricity, a medium amount of fuel and use it as a way of getting me a lot further, a lot faster in space.Farbod: And then instead of having, you know, solid state fuel or liquid fuel. And you know, you were talking about the example of going to Mars, that's not even considering coming back, right? So so you need to like, double the math there anyway. You can now have a nuclear reactor and like you said, a small mass of whatever your propellant is. In this case, it was solid lithium metal that's being vaporized and used as the propellant, to then take you there and bring you back. And it's totally feasible. And it's kind of not to get too nerdy, but Project Hail Mary, they were talking about the astrophage. It's kinda like making me think about the astrophage.Daniel: Well, no, I was going to say astrophage thruster would be like the actual solution to this. It doesn't exist in real life, but the astrophage, like they had, I think the thrusters on the back of the spacecraft wielded as much energy as like a small star. So, so we're not there yet, but the principle is actually very similar in Project Hail Mary, right? Can I use an energy source to accelerate mass and eject it out of the back of my spacecraft? And just so happened that the energy source they had in the fictional project, Homer, is like the highest energy density of any material they had ever discovered ever. Yeah.Farbod: Material known to man. Yeah. but yeah, I I think we actually kind of touched on the so what in our discussion of like what they were doing, but technologies like this make it a lot more feasible for us to think about you know, the the the human missions to Mars that have been in conversation since I was in high school, which apparently was a long time ago because I'm not a young man anymore. Okay. But it's demonstrating that like, you're saying, we don't we might not have to think about solid state chemical reactions as like one end of the spectrum and these like super low powered electric propulsion on the other side, but that there is actually a happy medium where we can have the best of both worlds as Hannah Montana one side. So pretty exciting if you're into aerospace or propulsion technology or you know, living on Mars, you've been waiting for that. Seems to be just around the corner.Farbod: You good for me to wrap it up? Alright. Folks, for decades we've been relying on these chemical rockets because they're really powerful, but they're also really fuel hungry. Up to two tons of fuel for the Saturn V. These electric propulsion technologies have been able to provide us some goodies when it comes to propulsion, but the trade off is actually different for them. They give you very little thrust, but they have very little mass. So do we really want that for long term? Probably not. Well, here comes NASA saving the day. They just took this massive leap towards an electric propulsion technology that can absolutely change that. They've increased the amount of propulsion generated by 25 times the next best option that they came out with four years ago. And these thrusters are only the starting point. It looks like in the next couple of years, they want to eight times the output of that, and they do have a clear roadmap to doing so. What they've been able to accomplish here has been great thanks to this liquid, sorry, solid lithium metal as the propellant, and still using the same technology that turns this lithium into plasma and ejects it out with electromagnetic technology. So, all that to say, this sustainable way to go into space and come back, like a Mars mission, is becoming more and more possible. And space technology and space travel in general will become faster and cheaper and a lot more practical with technologies like this.<hr />This episode was brought to you by Mouser, our go-to source for electronics parts for any hobby or prototype. Click <a href="https://resources.mouser.com/bench-talk-3/hydrogen-fuel-cells-power-the-future-of-evtols-3" rel="noopener noreferrer">HERE</a> to discover how hydrogen fuel cells are helping power the next generation of electric aircraft and making sustainable aviation a reality.Become a founding reader of our newsletter: <a href="http://read.thenextbyte.com/">http://read.thenextbyte.com/</a>As always, you can find these and other interesting &amp; impactful engineering articles on Wevolver.com.

Generics

Latest Posts