Fact-checked by Grok 2 weeks ago

eFuse

An eFuse, short for electrically programmable fuse, is an irreversible electronic component integrated into semiconductor chips that can be programmed by applying a high electrical current to disrupt its conductive path, typically via electromigration or Joule heating, thereby permanently altering circuit connectivity for functions such as redundancy or configuration.^[1]^[2] Developed primarily by IBM in the early 2000s, eFuses originated as a solution for on-chip self-repair, allowing defective circuits in memory arrays or logic to be bypassed without external intervention, using standard CMOS fabrication processes without additional materials.^[2] The programming mechanism involves passing a controlled current pulse—often around 10-20 mA for microseconds—through a narrow polysilicon or metal wire (e.g., 0.12 µm wide and 1.2 µm long, capped with cobalt silicide), which causes atomic migration and creates an open circuit gap, ensuring one-time programmability with high reliability under subsequent thermal and electrical stresses.^[2]^[1] Key applications of eFuses extend beyond initial memory redundancy to include chip identification, authentication for security, analog trimming for precision calibration, and enabling autonomic computing features in advanced processors, where they support post-packaging or even in-field reconfiguration to optimize performance and yield.^[1] Compared to traditional laser-etched fuses, eFuses offer advantages such as smaller footprint, faster programming speeds, compatibility with automated testing flows, and the ability to program without physically handling the wafer, reducing manufacturing costs and enabling dynamic adjustments like voltage scaling to mitigate heat and enhance reliability in high-density integrated circuits.^[1]^[2]

Fundamentals

Definition

An eFuse is a microscopic, electrically programmable element integrated into semiconductor chips, enabling permanent reconfiguration by increasing its resistance through the process of electromigration.^[3]^[4] This technology functions as a one-time programmable (OTP) memory cell, where a high current pulse causes metal atoms in a narrow interconnect—typically silicide or copper—to migrate, creating voids or discontinuities that transform a low-resistance state (around 100 Ω) to a high-resistance state (over 10 kΩ), effectively "blowing" the fuse without external physical intervention.^[3]^[4] Unlike traditional fuses, such as mechanical wire fuses or laser-etched metal fuses, which necessitate post-fabrication physical alterations like cutting or vaporization that can damage surrounding chip structures, eFuses operate entirely electrically within the integrated circuit, allowing in-situ programming during operation or testing.^[3]^[5] This OTP characteristic ensures irreversible changes, providing robust, tamper-resistant storage for critical data like calibration values or repair information.^[3] A key advantage of eFuses is their ability to facilitate post-manufacturing customization of chip functionality, such as circuit rerouting or performance tuning, without invasively modifying the underlying silicon structure, thereby enhancing yield and adaptability in advanced semiconductor processes.^[6] The invention is credited to IBM researchers in 2004, who developed it as a foundational technology for dynamic chip adaptation in high-performance computing applications.^[6]^[5]

History

The invention of eFuse technology is attributed to IBM, which in 2004 introduced and patented a method leveraging electromigration in silicide structures to program microscopic electrical fuses, enabling dynamic chip self-management and adaptation as part of their "chip morphing" initiative.^[7] This approach marked a shift from traditional laser-based fuses to electrically programmable ones, allowing chips to monitor and adjust performance without external intervention.^[8] The first commercial deployment of eFuses occurred in IBM's System z9 mainframe processors in 2005, where they facilitated memory redundancy by replacing defective elements and supported autonomic computing through built-in self-repair mechanisms.^[9] This implementation improved yield at various testing stages and enabled runtime adjustments for reliability in high-end enterprise systems.^[10] By the 2010s, eFuse usage evolved from an emphasis on performance tuning and fault repair to a predominant role as one-time programmable (OTP) elements for chip serialization, unique identification, and calibration, driven by advances in semiconductor scaling that made them more compact and reliable.^[3] This transition broadened their adoption beyond mainframes into consumer electronics and mobile devices post-2010, where they provided essential non-volatile storage for authentication and configuration without additional hardware.^[11] In recent years, through 2023 and 2024, eFuses have maintained their utility in security applications, such as configuring secure boot and encryption parameters in Espressif's ESP-IDF framework for IoT chips, and enforcing hardware-rooted protections in Samsung's Knox platform for mobile devices. No significant paradigm shifts have emerged by 2025, though integration continues in advanced nodes like 5nm processes to support OTP functions in high-density SoCs.^[12]

Semiconductor eFuses

Mechanism of Action

Semiconductor eFuses are constructed as thin strips of metal or semiconductor material, such as polysilicon or titanium silicide, embedded within the layers of an integrated circuit. These strips, often referred to as the fuse link, typically feature widths ranging from 0.2 to 1 μm and lengths of 1 to 10 μm, with the cathode and anode regions designed to optimize the temperature gradient for reliable operation. The structure leverages standard CMOS processes, utilizing silicided polysilicon gates over isolation regions to form the conductive path without requiring additional fabrication steps.^[13]^[14] The blowing process relies on electromigration, a phenomenon where high current densities cause metal atoms to migrate directionally under the combined forces of electron momentum transfer and thermal gradients. A programming current of 1 to 10 mA is applied for durations typically ranging from 100 μs to 10 ms, depending on the technology node and design, generating localized Joule heating and depleting material near the cathode while accumulating it near the anode, which forms voids and increases the fuse's resistance from approximately 0 Ω (intact state) to greater than 1 kΩ (blown state). Unlike mechanical rupture, this process avoids physical breakage, maintaining the chip's structural integrity while achieving a programmable open circuit.^[13]^[14]^[15] Detection of the blown state occurs post-programming through sense amplifiers or latches that measure the elevated resistance, distinguishing it from the low-resistance intact state with high reliability due to the large dynamic range—often exceeding several orders of magnitude. This electrical readout enables simple integration into control circuitry without invasive probing.^[14] These eFuses offer key advantages in scalability, integrating seamlessly into deep submicron CMOS nodes (e.g., 90 nm and below) where laser fuses become impractical due to alignment challenges in multi-layer stacks. Compared to antifuses, which rely on high-voltage dielectric breakdown and consume more energy, eFuses provide an energy-efficient alternative with lower programming voltages around 1.5 V.^[13]^[14]

Programming Process

In the pre-blow state, an eFuse operates in a low-resistance intact mode, typically around 100 Ω, where it is integrated into sense circuitry for reliable reading of its unprogrammed status.^[15] This configuration allows the eFuse to function as part of logic paths, such as in multiplexer trees, without disrupting normal chip operation until programmed.^[14] The programming process, or "blowing," activates the eFuse through controlled electromigration by applying a pulsed DC voltage of 1.5-5 V via on-chip drivers to the target element in the eFuse array.^[15]^[16] The current pulse, often delivered as a single or series of pulses with durations calibrated between 100 μs and 10 ms depending on the specific implementation, ensures complete electromigration while minimizing Joule heating to prevent overheating of adjacent structures.^[15]^[17] This sequence is managed by row and column decoders, akin to electrical programming in ROM arrays but relying on irreversible physical changes rather than charge storage.^[18] Following programming, verification confirms the blown state, where resistance shifts to over 10 kΩ, using built-in self-test (BIST) mechanisms or external testers that read via voltage dividers or current mirrors.^[15]^[19] In this high-resistance mode, the eFuse alters logic paths, such as redirecting signals in mux trees, enabling functions like redundancy allocation.^[14] Modern eFuse processes achieve blow yields exceeding 99.9%, supported by thermal isolation techniques that mitigate latent defects from localized heating during programming.^[15]^[14] These factors ensure high reliability, with post-programming resistance stable even under stress conditions like high voltage and temperature acceleration.^[20]

Applications

Chip Repair and Tuning

eFuses play a crucial role in redundancy implementation within integrated circuits, enabling the rerouting of signals around faulty memory cells or logic blocks to salvage otherwise defective chips. This technique, often referred to as sparing, is particularly vital for embedded memories like SRAM caches, where defects can otherwise render large portions of the die unusable. By programming eFuses during manufacturing test, redundant rows or columns can be activated to replace failed elements, thereby preserving functionality without requiring chip redesign.^[21]^[22] In analog trimming applications, eFuses facilitate precise calibration of circuit parameters post-fabrication, such as adjusting voltage references or clock frequencies by selectively blowing fuses to configure resistor or capacitor networks. This process compensates for process variations in mixed-signal ICs, where even small deviations can lead to performance failures. By storing trimming codes in eFuses, manufacturers can fine-tune analog blocks like ADCs or PLLs, enhancing overall yield and reliability in devices such as sensors and communication chips.^[23]^[3] Post-fabrication repair leverages eFuses during wafer-level testing to identify and disable defective cores in multi-core processors, allowing the chip to operate with reduced but functional resources. This approach isolates faults in one or more cores, enabling the sale of partially good dies as lower-binned products and significantly boosting overall manufacturing yield. The one-time programmable nature of eFuses ensures these repairs are permanent, supporting efficient production scaling in advanced nodes.^[24]^[22] A notable case study is the early adoption of eFuses in IBM mainframes, such as the Power5 processor and System z9, where they were employed for cache repair to address memory defects and improve system reliability. In these implementations, eFuses stored redundancy mapping data to bypass faulty cache lines, enabling autonomous defect handling during production. Modern applications extend this to high-performance processors, including GPUs, for defect mapping and redundancy activation to mitigate manufacturing variations.^[22]^[10]^[25]

Security and Identification

eFuses play a critical role in chip serialization by enabling the permanent storage of unique identifiers, typically ranging from 128 to 256 bits, which are created through specific blow patterns during manufacturing or provisioning. These identifiers facilitate device traceability and serve as a defense against counterfeiting, particularly in Internet of Things (IoT) applications where authenticity verification is essential. For instance, in ESP32 microcontrollers, eFuse blocks of 256 bits are allocated to store MAC addresses and custom serial numbers, ensuring each device has a non-replicable identity that can be read but not altered post-programming. Similarly, GigaDevice's GD32H7xx series utilizes 128-bit hardware unique keys in eFuses to establish a root of trust for device-specific operations.^[26]^[27] Version locking via eFuses prevents rollback exploits by storing firmware version information in a one-time programmable manner, allowing only updates to equal or higher versions once blown. This mechanism enforces monotonic progression of software versions, blocking the installation of older, potentially vulnerable firmware that could introduce security gaps. In Silicon Labs' SiWx917 SoCs, eFuses implement anti-rollback protection during secure boot, where version counters are incremented irreversibly to validate firmware integrity against known threats. Texas Instruments' AM26x devices similarly employ 64-bit eFuse fields in their Hardware Security Module (HSM) to preserve version data, ensuring compliance with secure update policies.^[28]^[29] In authentication processes, eFuses integrate with Physically Unclonable Functions (PUFs) to support secure boot mechanisms, where eFuses store PUF helper data or authentication parameters as one-time programmable (OTP) elements. This combination generates device-unique responses for cryptographic challenges, enabling root key derivation without exposing secrets. AMD's Zynq UltraScale+ devices, for example, use eFuses in PUF mode to register and store helper data for boot authentication, ensuring that only authorized images can execute by verifying PUF-derived keys during the boot chain. eFuses also function as OTP storage for root keys, providing a tamper-evident foundation for key hierarchies in secure provisioning; Qualcomm's platforms embed root-of-trust keys in eFuses to authenticate boot images and enforce chain-of-trust integrity.^[30]^[31] The inherent irreversibility of blown eFuses enhances tamper resistance, complicating reverse engineering efforts by permanently altering silicon structures that cannot be undone or mimicked without specialized equipment. This property makes eFuses ideal for configuring trusted execution environments (TEEs), where they lock security policies such as debug access or isolation modes post-deployment. In Qualcomm's TEE implementations, eFuses provide a tamper-resistant root of trust by storing immutable configuration bits that isolate secure operations from the normal world, preventing unauthorized modifications to the execution environment. Such usage ensures that once security features are enabled, adversaries face significant barriers to bypassing hardware-enforced protections.^[31]

Performance Optimization

eFuses play a crucial role in semiconductor binning processes, where post-fabrication testing determines a chip's performance characteristics, and specific fuses are blown to configure clock multipliers or voltage rails accordingly. This allows manufacturers to sort chips into distinct speed grades, such as high-end variants capable of higher frequencies versus standard ones operating at lower speeds, maximizing yield and market segmentation without discarding functional silicon. For instance, in advanced processors, eFuses enable precise adjustments based on measured maximum operating frequencies, ensuring each chip is tailored to its optimal performance bin during production testing. In addition to static binning, eFuses support adaptive tuning in autonomic computing systems, where they facilitate reconfiguration to optimize for varying workloads by, for example, permanently disabling underperforming cores identified during initialization or stress testing. This approach enhances overall system efficiency by focusing resources on reliable components, particularly in multi-core architectures where variability in silicon quality can affect individual core performance. Such tuning is performed once at deployment or boot time, providing a hardware-level optimization that complements software scheduling.^[32] For power management, eFuses are utilized to trim critical analog components like oscillators and phase-locked loops (PLLs) in mobile system-on-chips (SoCs), fine-tuning operating voltages to reduce leakage current and dynamic power dissipation. By calibrating these elements to match process variations, eFuses minimize unnecessary power draw in low-power modes, thereby extending battery life in portable devices. This trimming is essential for achieving the tight tolerances required in battery-constrained applications, where even small reductions in voltage overhead yield measurable efficiency gains. Despite these benefits, the irreversible nature of eFuse programming introduces trade-offs, as modifications cannot be undone post-blowing, limiting adaptability to evolving workloads compared to dynamic software or reconfigurable hardware alternatives. However, this permanence ensures faster access times and lower latency during operation, as fuse states are read instantly at power-on without ongoing computational overhead, making eFuses preferable for scenarios demanding high reliability and speed over flexibility.^[33]

Implementations

Early Developments

The invention of eFuses in 2004 by IBM marked a significant advancement in semiconductor programmable devices, enabling electrical programming through controlled electromigration rather than physical severance.^[22] IBM's pioneering patent US7485944B2, granted in 2009, described a programmable eFuse structure consisting of an elongated semiconductor material, such as P+ polysilicon doped at concentrations of at least 10^17 atoms per cubic centimeter, overlaid with a metallic silicide layer like CoSi₂ or NiSi₂ to facilitate electromigration under a precise current pulse of approximately 10 mA for 150-350 μs.^[16] This design allowed for substantial resistance increase (over 5 kΩ post-programming) without physical rupture or collateral damage to adjacent structures, integrating seamlessly with CMOS logic circuits.^[16] The structure's asymmetry, with a wider cathode end, ensured directed metal migration, enhancing programmability for self-managing chip features when combined with software algorithms that monitor and adjust functionality in real time.^[16]^[34] The first major deployment of eFuses occurred in IBM's System z9 mainframe processors in 2007, incorporating thousands of eFuses per chip primarily for redundancy in memory repair and unique chip identification to support personalization and reconfiguration.^[9] This implementation utilized a methodology involving controlled high-voltage pulses to induce silicide electromigration, programmed post-wafer fabrication using standard logic processes, which avoided yield loss by eliminating the need for destructive testing or physical alterations during production.^[9] The eFuses were arrayed to enable on-chip self-repair, where defective elements could be bypassed without impacting overall chip integrity.^[9] The transition from traditional laser fuses to eFuses in the System z9 eliminated the requirement for specialized laser programming stations, streamlining high-volume manufacturing through simplified test equipment and processes.^[10] Early challenges included achieving uniform electromigration across large fuse arrays to prevent variability in programming thresholds, addressed via optimized silicide thickness (60-250 nm) and doping profiles that ensured consistent resistance changes.^[9] Reliability assessments demonstrated exceptional longevity, with accelerated testing of millions of programmed fuses at 125°C and 150°C for up to 2000 hours showing zero failures, confirming suitability for mission-critical applications.^[10]

Modern Examples

In gaming consoles, the Xbox 360 utilizes eFuses in its Xenon CPU to enforce hardware restrictions, such as preventing unauthorized modifications and ensuring system updates align with blown fuse counts during boot processes. These 768-bit eFuses are blown with each dashboard update, allowing the hypervisor to verify integrity and block incompatible software versions. Similarly, the Nintendo Switch employs eFuses to lock firmware versions and prevent downgrades; the bootloader checks the number of blown fuses on every boot, resetting the device if the count mismatches the expected value for the installed firmware, thereby supporting up to 32 irreversible updates across its lifecycle.^[35]^[36] In mobile devices, Samsung's Galaxy S22 series integrates eFuses within its Knox security platform to detect tampering and enforce version integrity. When the bootloader is unlocked or rooting is attempted, an eFuse is blown, tripping Knox and permanently disabling features like Secure Folder and Samsung Pay while displaying a warranty void status.^[37] Google's Pixel 6 similarly uses eFuses for anti-rollback protection, where upgrading to Android 13 or later increments the fuse state during boot, rendering downgrades impossible to avoid vulnerabilities in prior bootloader versions.^[38] Modern processors continue to leverage eFuses for critical functions like identification and repair. The Espressif ESP32, with updates through 2023, stores the base MAC address in eFuse block 3, enabling unique device identification via the eFuse Manager API for secure network operations without external storage.^[26] In high-performance computing, AMD and Intel CPUs fabricated on 5nm nodes employ eFuses for cache repair and binning; defective cache lines or cores are disabled by blowing fuses post-manufacturing, allowing yield optimization by reallocating resources to higher-binned SKUs.^[39]

Resettable eFuses

Resettable eFuses are standalone integrated circuits (ICs) designed for board-level power path protection, electronically monitoring and interrupting current flow to safeguard circuits from faults, in contrast to one-time programmable semiconductor eFuses embedded within chips.^[40] These devices, such as Texas Instruments' TPS series and Toshiba's TCKE9 series introduced in 2024, integrate protection functions into a single package for applications requiring reliable, reusable circuit safeguarding.^[41]^[42] The core mechanism relies on MOSFET-based switching to detect and respond to faults, including overcurrent conditions exceeding thresholds typically above 1 A, overvoltage events up to 28 V or higher depending on the model, and inrush currents during power-up.^[40]^[42] Upon fault detection, the internal MOSFET turns off to isolate the load, with voltage clamping to limit transients, followed by automatic reset once the fault condition clears, enabling repeated operation without manual intervention.^[40] This active control distinguishes resettable eFuses from passive fuses, providing precise interruption and recovery.^[43] Key features include diagnostic flags for real-time fault reporting, such as overcurrent or overtemperature alerts, and adjustable slew rate control to manage inrush currents during hot-plug events, preventing voltage dips on the power rail.^[40]^[43] Response times are notably fast, often under 1 μs for overcurrent protection, compared to milliseconds for polymeric positive temperature coefficient (PPTC) fuses, enabling quicker isolation of faults in sensitive systems.^[40] For instance, Littelfuse eFuses incorporate overtemperature shutdown with integrated sensing for thermal protection.^[43] The market for resettable eFuses is expanding, projected to reach $737.4 million by 2030 from $557.0 million in 2024, at a compound annual growth rate (CAGR) of 4.9%, fueled by demand in automotive electronics and USB power delivery systems.^[44] These devices offer advantages over traditional fuses, including reusability for over 1,000 cycles and programmability through interfaces like I²C for customizing limits, though they incur higher upfront costs due to integrated circuitry.^[40]^[42]

Antifuses

Antifuses serve as complementary one-time programmable (OTP) elements to eFuses in semiconductor devices, functioning by establishing permanent electrical connections rather than interruptions. Their basic structure resembles a capacitor, featuring a thin insulating dielectric layer—typically oxide-nitride-oxide (ONO) or amorphous silicon—sandwiched between two metal or polysilicon electrodes.^[45]^[46] In the unprogrammed state, this configuration exhibits extremely high resistance, often exceeding 1 GΩ, preventing current flow and acting as an open circuit.^[46] Programming an antifuse involves applying a high-voltage pulse, typically greater than 10 V and up to 15 V, across the electrodes, which induces dielectric breakdown in the insulating layer. This breakdown creates a conductive filamentary path through the dielectric, permanently shorting the structure and reducing its resistance to a low value, approximately 10–100 Ω depending on programming conditions such as current (e.g., 15 mA).^[47]^[48] Unlike eFuses, which start conductive and become non-conductive, antifuses transition from insulating to conductive, enabling their use primarily for forming routing connections in programmable logic.^[49] This mechanism ensures reliability in the programmed state, with the low-resistance path stable over time and resistant to further changes under normal operating voltages.^[50] Antifuses find key applications in field-programmable gate arrays (FPGAs), where they provide programmable interconnects for routing signals between logic blocks. Prominent examples include the Actel (now Microchip) Axcelerator family, which employs antifuse technology for one-time configuration of complex digital circuits.^[51] Their radiation tolerance makes them particularly suitable for space applications, as the permanent programming withstands single-event upsets from cosmic rays without configuration loss, unlike volatile SRAM-based alternatives. However, antifuses are less prevalent than eFuses overall due to the need for on-chip high-voltage circuitry, which increases programming complexity and power requirements.^[52] In comparison to eFuses, which operate by electromigration to open circuits for trimming and repair, antifuses close circuits via breakdown, offering complementary functionality in OTP memory hierarchies. Both are non-volatile and irreversible, but antifuses necessitate thicker dielectric isolation to handle programming voltages, which can limit integration density in advanced nodes compared to the lower-voltage eFuse processes.^[53]^[54] As of 2025, antifuses see hybrid integration in system-on-chips (SoCs) for targeted routing in security-critical or radiation-exposed subsystems, such as secure boot key storage in microcontrollers like the Raspberry Pi RP2350.^[55] However, in January 2025, the Raspberry Pi RP2350 Hacking Challenge revealed techniques to extract secrets from its antifuse-based OTP memory using semiconductor failure analysis methods like focused ion beam editing and passive voltage contrast, prompting recommendations to reassess security assumptions for such implementations.^[56] Despite this, eFuses remain dominant for general trimming and calibration due to scalability advantages in sub-16 nm processes, though antifuse FPGAs continue to grow in niche markets like aerospace, valued at $435.75 million in 2024 and projected to reach $798.60 million by 2032 at a CAGR of 7.5%.^[57]

References

[1]
Electrically Programmable Fuse (eFUSE): From Memory ...
Electrical fuse (eFUSE) has become a popular choice to enable memory redundancy, chip identification and authentication, analog device trimming, and other ...
[2]
Electrical Fuse Lets Chips Heal Themselves - IEEE Spectrum
Oct 1, 2004 · The fuse is opened through a process called electromigration, in which current pushes the atoms in small wires out of place. Electromigration ...
[3]
The Growing Need For OTP - Semiconductor Engineering
Aug 4, 2016 · ... eFuse, invented by IBM in 2004. Using electromigration, IBM was able to program a fuse without damaging other parts of the chip. In this way ...
[4]
Electrically programmable fuse (eFUSE) using electromigration in ...
Electrically programmable fuse (eFUSE) using electromigration in silicides. Abstract: For the first time we describe a positive application of electromigration, ...
[5]
IBM's eFuse technology portends adaptable chips - EDN
Aug 2, 2004 · The technology, called “eFuse,” is said to combine software algorithms and microscopic electrical fuses to produce chips able to regulate and ...
[6]
IBM Introduces Chip Morphing Technology: Self-Managing Semiconductors System
### Summary of IBM's eFUSE Invention (2004)
[7]
IBM Introduces Chip Morphing Technology - HPCwire
Aug 6, 2004 · The patented technology, called “eFUSE,” combines unique software algorithms and microscopic electrical fuses to produce chips that can regulate ...
[8]
IBM Intros Chip Morphing Technology - EDN Network
Jul 31, 2004 · eFUSE works by combining software algorithms and microscopic electrical fuses, opposed to laser fuses, to produce chips that can regulate and ...
[9]
https://ieeexplore.ieee.org/document/5388714
[10]
IBM System z9 eFUSE applications and methodology - ResearchGate
Aug 9, 2025 · This paper describes a second-generation one-time programmable ... Electrically programmable fuse (eFUSE) using electromigration in silicides.
[11]
eFuse Market Size & Share Analysis Report by 2032
Oct 26, 2025 · With time, advances in the semiconductor design enabled eFuse to become smaller and more powerful, paving the way for broader use in consumer ...Missing: review | Show results with:review
[12]
Reduced area efuse cell structure - Google Patents
Electrical fuses (eFuses) are used to reprogram integrated circuit chips, such computer chips. For example, eFuses can be used to provide in-chip performance ...
[13]
https://ieeexplore.ieee.org/document/1028987
[14]
eFuse Design and Reliability
This paper will discuss selected eFuse technologies describing the design philosophy, electrical programming and characterization, the physics of failure, and.
[15]
Characterization of eFuse Programming for Varying RF BiCMOS ...
Apr 29, 2020 · This work demonstrates the compatibility of electrically programmable fuse (eFuse) technology across a range of process technology nodes, as ...
[16]
US7485944B2 - Programmable electronic fuse - Google Patents
A programmable device (eFuse), includes: a substrate (10); an insulator (13) on the substrate; an elongated semiconductor material (12) on the insulator, ...Missing: definition | Show results with:definition
[17]
US7345904B1 - Method for programming an electronically ...
A method for programming an electronically programmable semiconductor fuse applies a programming current to a fuse link as a series of multiple pulses.
[18]
US7710813B1 - Electronic fuse array - Google Patents
May 4, 2010 · The E-fuse memory array 200 includes a sense amplifier array 204, a bit line READ decoder/ multiplexer 206, a bit line PROGRAM decoder/ driver ...Missing: verification BIST
[19]
Method and circuit for implementing eFuse sense amplifier verification
A method and circuit for implementing Efuse sense amplifier verification, and a design structure on which the subject circuit resides are provided.Missing: column BIST
[20]
[PDF] Product specific sub-micron E-fuse reliability and design qualification
Figure 5a and Figure 9 displays the sensitivity to programming yield with respect to short times. The Figure 5a fuse has a programming yield of 100% for ea.
[21]
Low-power SRAM E-fuse repair methodology - Google Patents
A low power E-fuse repair methodology substantially removes system latency during memory and/or E-fuse farm module power-down in a device that employs ...
[22]
Electrically Programmable Fuse (eFUSE): From Memory ...
Sep 16, 2007 · We will review the evolution and applications of electrical fuse solutions for 180nm to 45nm technologies at IBM, and provide some insight ...Missing: electromigration | Show results with:electromigration
[23]
US12326473B2 - Electronic fuse device and operation method ...
For example, an integrated circuit (IC) may use the eFuse to store analog trimming data, system settings, a key for an encryption circuit, and/or other ...
[24]
IBM eFuse to yield self-repairing, self-regulating CPUs - The Register
Aug 2, 2004 · Indeed, "eFuse is part of a built-in self-repair system that constantly monitors a chip's functionality. If an imperfection is detected, this ...Missing: definition invention
[25]
Targeting Redundancy In ICs - Semiconductor Engineering
Jun 14, 2021 · With a hard Incremental repair, the device has spare efuse real estate to be able to record the repair programming.” This can improve ...
[26]
eFuse Manager - ESP32 - — ESP-IDF Programming Guide v5.5.1 ...
The eFuse Manager component is a collection of tools and APIs that assist with defining, burning, accessing eFuses parameters.Missing: 2023 2024
[27]
[PDF] GD32H7xx Security Introduction Application Note AN121
Sep 13, 2023 · EFUSE ... Every device has a 96-bit unique identifier, which can't be modified. These bits can be read by the user and will not be repeated ...
[28]
[PDF] AN1442: SiWx917 SoC Secure Boot with Anti-Rollback Protection
Secure boot on SiWx917 executes the initial boot from immutable memory, authenticating code before execution, and includes anti-rollback protection.
[29]
Security in External Flash devices — AM26x Academy
For the same reason, the eFuses in HSM subsystem provide an option to preserve the version information and prevent anti-rollback. There are 64 bits provided ...
[30]
PUF Registration in Boot Header Mode - 2025.1 English - UG1209
The PUF registration software operates in a boot header mode or eFUSE mode. The boot header mode allows development without programming the OTP eFUSEs.
[31]
Qualcomm Linux Security Guide
Aug 6, 2024 · The Secure boot feature offers image authentication and tamper-resistant RoT in electronic fuses (eFuse). The debug security feature allows ...
[32]
[PDF] OpenSPARC T2 System-On-Chip (SoC) Microarchitecture ... - Oracle
... eFuse Array (EFA) 4–5. 4.2.1.2. eFuse Controller (FCT) 4–6. 4.2.1.3. TCU ... speed binning on the tester. Each CPU will have four signals driven to debug.v.
[33]
https://ieeexplore.ieee.org/document/9000162
[34]
IBM aims to make chips more autonomous | ZDNET
Jul 30, 2004 · eFuse adds the equivalent of numerous tiny electrical fuses to each chip that, when combined with special onboard software, can allow the chips ...
[35]
[PDF] Basics of eFuses (Rev. A) - Texas Instruments
If the fault lasts long enough to activate the eFuse's thermal shutdown (typically TJ = 150°C), then the fault will still cause the eFuse to turn OFF (similar ...Missing: programming | Show results with:programming
[36]
eFuses & hot swap controllers | TI.com - Texas Instruments
Improve the reliability of your system against overvoltage, short circuit, inrush current and reverse polarity events with our eFuses and hot swap ...Missing: trim oscillator PLL mobile SoC
[37]
Toshiba Releases TCKE9 Series Compact Electronic Fuses (eFuse ...
The TCKE9 series of 25V high input voltage eFuse ICs offers 2 types of product: an auto-retry type that allows the eFuse IC to automatically recover the ...
[38]
eFuse Overcurrent, Overvoltage, Inrush Current Protection - Littelfuse
An eFuse provides an all-in-one PCB-mounted overcurrent, overvoltage, overtemperature, and inrush current protection solution with real-time diagnostics.Missing: trim oscillator PLL mobile SoC
[39]
eFuse Market worth $737.4 million by 2030 - MarketsandMarkets
May 23, 2025 · The eFuse market is projected to reach USD 737.4 million by 2030 from an estimated USD 557.0 million in 2024, at a CAGR of 4.9% during the forecast period.
[40]
(PDF) Characterization and modeling of a highly reliable metal-to ...
A conventional antifuse consists of a dielectric layer sandwiched between two electrodes as shown in Fig. 2(a). Commonly used implementations include n+ ...
[41]
[PDF] A Synthesis Approach for Coarse-grained, Antifuse-based FPGAs
The antifuse element is formed by depositing a high resistance layer (> 1GΩ) of amorphous silicon above a tungsten via a plug that would otherwise bridge the ...<|separator|>
[42]
[PDF] FAQ's for Microsemi Antifuse FPGA Programming
How can antifuse resistance be determined? Is it possible that fuse resistance could be great enough to cause a delay greater than the worst case simulated ...
[43]
4.1 Antifuse Technology in FPGAs: Programming and Reliability ...
Increasing the programming current to 15 mA might reduce the average antifuse resistance to 100 W. Antifuses separate interconnect wires on the FPGA chip and ...
[44]
Anti-fuse memory provides robust, secure NVM option - EE Times
May 7, 2012 · Small quantities of e-fuse are ubiquitous in SoC designs to provide memory repair, for containing trimming data for analog mixed signal circuits ...
[45]
Long term storage reliability of antifuse field programmable gate arrays
The programming of the antifuse occurs when a voltage above a threshold value is applied across the structure. Upon programming, the resistance of the antifuse ...
[46]
Radiation-Tolerant FPGAs - Microchip Technology
High-reliability, radiation-tolerant antifuse-based FPGAs · Compact packaging for command and control applications · Flight heritage established on many programs ...
[47]
[PDF] Actel's Quality and Reliability Guide
The antifuse FPGA does lose some chip area because it needs on-chip high-voltage programming circuits. These circuits reside on the chip's edges, however; ...
[48]
Antifuse is the New Foundation of NVM Below 16nm - SemiWiki
May 8, 2015 · An antifuse is the opposite of an eFuse. The circuit is open (high resistance) to begin with and is programmed closed by applying electrical ...Missing: comparison | Show results with:comparison
[49]
[PDF] From Experiments to Test-chip Design
A comparison of eFuse and anti-fuse properties is summarized in Table. 1.1. The main disadvantage of the anti-fuse memory is the high amplitude of the ...<|separator|>
[50]
Anti-Fuse OTP | Technology | PUFseucurity - PUFsecurity
It can create a clear vulnerability and immediately compromise a secure boundary if, for example, the private keys are left unsecured in eFuse. Furthermore, ...Missing: serialization | Show results with:serialization<|control11|><|separator|>
[51]
Antifuse FPGA Market Size, Share, Growth | CAGR Forecast 2032
Rating 4.8 (88) Antifuse FPGA Market size was valued at USD 435.75 million in 2024 and the revenue is expected to grow at a CAGR of 7.5% from 2025 to 2032.