IC Reliability Issues in VLSI Design

Technical Analysis, Effects, and Layout Mitigations

July 2025
 Table of Contents
1. Electromigration (EM)
2. Bias Temperature Instability (BTI)
3. Hot Carrier Injection (HCI)
4. Time-Dependent Dielectric Breakdown (TDDB)
5. Single Event Upsets (SEU)
6. Electrostatic Discharge (ESD)
7. Latch-up
8. Soft Metal Trap (SMT) & Aging
9. FinFET Reliability
10. 3D-IC Reliability
11. Self-Heating Effect (SHE)
12. Process Variation-Induced Reliability
13. Advanced NBTI in FinFET PMOS
 Electromigration: Introduction

• Electromigration (EM) is a failure mechanism in metal interconnects of ICs caused by momentum transfer from high-density
electron flow.
• Results in atomic diffusion, leading to formation of voids or hillocks in metal lines.
• Becomes increasingly critical in advanced technology nodes due to thinner interconnects and higher current densities.
 Electromigration: Physical Mechanism

• Two competing forces: electron wind force (drives atoms) and back stress (opposes it).
• Atoms migrate from cathode to anode under current stress.
• Leads to: open circuits (voids) or short circuits (hillocks).
 Electromigration: Black’s Equation

• Common empirical model to estimate Mean Time To Failure (MTTF):

MTTF = A * j^(-n) * exp(Ea / (kT))
where:
- A: constant based on material/process
- j: current density
- n: scaling factor (typically 1–2)
- Ea: activation energy
- T: temperature in Kelvin
• Strongly dependent on current density and temperature.
 Electromigration: Key Influencing Factors

• Current density: Higher current = faster EM degradation.

• Temperature: Increases atomic mobility, accelerating EM.
• Line geometry: Narrower lines fail faster.
• Material: Copper resists EM better than aluminum.
• Layout-induced stress gradients also impact EM.
 Electromigration: Layout Mitigation Techniques

• Use wider interconnects for high-current nets.

• Increase metal layer usage for critical nets (e.g., M4 instead of M2).
• Insert redundant vias to reduce local current crowding.
• Avoid sharp turns; promote uniform current flow.
• Perform EM-aware signoff with current density checks (e.g., in Cadence Voltus, Synopsys PrimeRail).
 Electromigration: Design Best Practices

• Follow foundry-recommended EM limits for each metal layer.

• Place high-current devices closer to power supply.
• Avoid tapering interconnects abruptly.
• Use EM monitoring test structures in silicon for data validation.
• In 3D-ICs: manage TSV current paths to prevent EM hotspots.
 BTI: Introduction

• BTI (Bias Temperature Instability) causes threshold voltage shift over time.
• NBTI affects PMOS under negative bias; PBTI affects NMOS under positive bias.
• Leads to timing degradation and eventual failure in digital logic.
 BTI: Physical Mechanism

• Caused by charge trapping at the Si-SiO2 interface under bias and elevated temperature.
• Stress results in formation of interface traps and oxide traps.
• Some degradation is recoverable; others permanent.
 BTI: Modeling & Equations

• Empirical models: ΔVth ∝ t^n * exp(-Ea/kT).

• Reaction-diffusion model: interface traps accumulate and diffuse over time.
• Finite recovery modeled with duty cycle-aware stress equations.
 BTI: Factors Affecting Degradation

• Temperature: accelerates degradation rate.

• Voltage: higher gate voltage increases trap generation.
• Duty cycle: intermittent stress leads to partial recovery.
• Fin pitch & gate dielectric composition in FinFETs.
 BTI: Layout-Level Mitigations

• Use of dummy gates to ensure uniform stress across layout.

• Cell placement to balance temperature and stress loading.
• Multi-Vt strategies to reduce high-Vth degradation hotspots.
• Avoiding dense clustering of high-Vth devices.
 BTI: Design Practices

• Aging-aware synthesis and timing closure.

• Implement path balancing to minimize margin hit.
• Predictive aging monitors embedded in silicon.
• Adjust clock skew proactively using adaptive clocking.
 HCI: Introduction

• HCI (Hot Carrier Injection) is a degradation mechanism caused by high-energy carriers.

• Affects NMOS more than PMOS in traditional planar CMOS.
• Can lead to gain degradation and timing shifts.
 HCI: Physical Mechanism

• Under high Vds, carriers gain energy and are injected into gate oxide.
• Causes interface state generation and trapped charges.
• Primarily impacts short-channel devices.
 HCI: Impact & Failure Modes

• Threshold voltage shift (ΔVth).

• Decrease in drain current (Idlin, Idsat).
• Gate leakage and transconductance reduction.
 HCI: Modeling Techniques

• Lifetime inversely proportional to substrate current and electric field.

• Empirical lifetime models: τ ∝ I_sub^−m, where I_sub is substrate current.
• Simulation models include time and stress voltage dependence.
 HCI: Layout-Level Mitigation

• Use LDD (Lightly Doped Drain) structures.

• Limit Vds across sensitive nodes.
• Use device mirroring and guard ring techniques to distribute stress.
• Increase channel length for critical paths.
 HCI: Advanced Practices

• Avoid abrupt switching in high-speed data paths.

• Use stress-aware standard cells and ring oscillator monitors.
• FinFETs significantly reduce HCI due to 3D channel geometry.
 TDDB: Introduction

• TDDB (Time-Dependent Dielectric Breakdown) refers to gradual breakdown of gate oxide under prolonged electric stress.
• Leads to permanent failure when the oxide becomes conductive.
• Key concern in thin oxide devices at advanced technology nodes.
 TDDB: Failure Mechanism

• Prolonged electric field across gate oxide causes trap accumulation.

• Eventually forms a conductive path, causing hard breakdown.
• Soft breakdown may occur first, increasing leakage before full failure.
 TDDB: Physical Models

• E-model: tBD ∝ exp(γ/E), where E is electric field across oxide.

• 1/E model: tBD ∝ exp(B/E), used for high-k/metal gate stacks.
• TDDB lifetime extrapolated using accelerated stress testing.
 TDDB: Factors Affecting Breakdown

• Oxide thickness: thinner oxides fail faster.

• Material quality: high-k oxides have different failure kinetics.
• Temperature: higher T accelerates trap formation.
• Interface roughness and mechanical stress.
 TDDB: Layout-Level Mitigation

• Use thicker gate oxides for I/O or high-voltage domains.

• Spread voltage stress across multiple transistors when possible.
• Avoid layout hotspots that concentrate electric field.
• Apply guard rings or dummy fill to reduce oxide field gradients.
 TDDB: Design Practices

• Derate operating voltages based on foundry TDDB specs.

• Use TDDB-aware PDK checks and DRC rules.
• Simulate oxide field using EDA tools (e.g., Ansys RedHawk, Synopsys Finesim).
• Reliability-aware floorplanning in sensitive IP blocks.
 SEU: Introduction

• SEU (Single Event Upset) refers to a change of state in memory or logic due to radiation-induced charge.
• Common in aerospace, automotive, and high-altitude applications.
• Does not cause permanent damage but leads to functional failure.
 SEU: Mechanism of Occurrence

• High-energy particles (neutrons, alpha particles) strike silicon and generate electron-hole pairs.
• If collected charge exceeds critical threshold (Qcrit), a logic flip occurs.
• SRAM, FFs, and combinational logic are vulnerable.
 SEU: Key Parameters and Models

• LET (Linear Energy Transfer) quantifies energy loss per unit path length.
• Qcrit: minimum charge to flip state; depends on node capacitance and Vdd.
• SEU Cross-section: σ = number of upsets per fluence per bit.
• Monte Carlo simulations and TCAD used for analysis.
 SEU: Impact on Circuit Behavior

• Bit flips in configuration or memory elements.

• Transient pulses (glitches) in logic leading to downstream errors.
• May propagate if not filtered or masked by logic depth or timing.
 SEU: Layout-Level Mitigation

• Use of guard rings and substrate ties to reduce charge collection.

• Increase node capacitance to raise Qcrit (e.g., via transistor upsizing).
• Physical separation of redundant logic (DICE, TMR).
• Use of dual-rail logic to sense differential upset.
 SEU: Design and System Practices

• Error Correction Codes (ECC) for memories.

• Triple Modular Redundancy (TMR) for logic voting.
• SEU-aware RTL simulation and fault injection.
• Use of radiation-hardened standard cell libraries (RHBD).
 ESD: Introduction

• ESD (Electrostatic Discharge) is a sudden flow of static electricity between two objects with different potentials.
• Can cause immediate or latent failure in ICs.
• ESD is one of the most common causes of yield loss in semiconductor manufacturing.
 ESD: Failure Mechanism

• Discharge generates a high current pulse (up to several amps) in nanoseconds.

• Causes local heating, junction breakdown, or metallization damage.
• Failure types: gate oxide rupture, thermal burnouts, or dielectric punch-through.
 ESD: Test Models

• Human Body Model (HBM): simulates static charge from human handling.
• Machine Model (MM): models discharge from equipment or machinery.
• Charged Device Model (CDM): emulates charge stored on the IC itself.
• Each has distinct current rise times and pulse widths.
 ESD: Protection Devices

• Diodes and resistor-capacitor (RC) clamps for core and I/O.

• Silicon Controlled Rectifiers (SCRs) for robust ESD protection.
• Use of GGNMOS (grounded-gate NMOS) for efficient current shunting.
• ESD protection integrated at pad-ring and power domains.
 ESD: Layout-Level Techniques

• Ensure wide and symmetric metal routing to prevent current crowding.

• Add ESD protection cells near I/O pads and power clamps.
• Maintain ESD discharge paths via guard rings and substrate taps.
• Avoid sharp corners and metal necking that concentrate current.
 ESD: Design Best Practices

• Use ESD-aware cell libraries and ensure clamp coverage.

• Validate ESD robustness using LVS and DRC checks with ESD rules.
• Simulate with ESD circuit models and parasitic-aware extraction.
• Apply on-chip system-level protection for sensitive blocks (RF, analog).
 Latch-up: Introduction

• Latch-up is a parasitic condition in CMOS where a low-impedance path is inadvertently created between power and ground.
• Causes high current flow, leading to thermal damage or IC failure.
• Triggered by overvoltage, transient noise, or ionizing radiation.
 Latch-up: Parasitic Structures

• Occurs due to inherent parasitic pnp and npn BJTs in CMOS substrates.
• Forms a Silicon Controlled Rectifier (SCR) structure.
• Once triggered, remains latched until power is cycled or the device is destroyed.
 Latch-up: Trigger Mechanisms

• ESD or I/O transients inject minority carriers into substrate.

• Forward biasing of p-n junctions initiates BJT action.
• UV exposure, alpha particles, or overvoltage at input pins.
 Latch-up: Detection and Analysis

• Simulation using latch-up extraction tools in EDA (e.g., Calibre LUP, StarRC).
• Use of test chips under controlled temperature and injection conditions.
• Latch-up robustness specified as holding voltage and trigger current.
 Latch-up: Layout-Level Prevention

• Use of guard rings to collect injected carriers and isolate devices.

• Deep n-well and triple-well isolation to separate PMOS and NMOS regions.
• Increasing spacing between sensitive devices to reduce parasitic gain.
• Adding substrate and well taps near transistors to lower resistance paths.
 Latch-up: Design Guidelines

• Follow foundry DRC rules for latch-up prevention (e.g., tap density, spacing).
• Ensure proper biasing of wells and substrate during operation.
• Avoid floating wells or underbiased regions in layout.
• Use latch-up hardened libraries for mission-critical blocks.
 SMT & Aging: Introduction

• Stress Migration (SMT) refers to metal atom movement under mechanical stress gradients, often due to thermal cycling.
• Aging includes all gradual parametric degradations that occur over operational life, impacting performance and reliability.
• Both effects are increasingly critical in advanced nodes and long-life products (e.g., automotive, aerospace).
 Stress Migration: Mechanism

• Caused by thermal expansion mismatch between metal and surrounding dielectric layers.
• Leads to formation of voids in metal lines, similar to electromigration but driven by stress gradients.
• More severe during power cycling and thermal variations.
 Aging: Mechanisms

• Includes BTI, HCI, SMT, and dielectric wear-out.

• Affects delay, leakage, threshold voltage, and drive strength.
• Typically modeled as a time-dependent variation in device parameters (e.g., ΔVth(t)).
 Modeling Aging in Circuits

• Foundries provide reliability models for aging-aware timing (e.g., SPICE models with BTI/HCI aging corners).
• Tools like Synopsys PrimeTime-Aging or Cadence Tempus support aging simulation.
• Used in high-reliability ASICs and safety-critical ICs.
 Layout Mitigation for SMT

• Use redundant vias and wide metal segments to reduce stress concentration.
• Avoid abrupt metal width changes and sharp corners.
• Place dummy fill and dummy metals to balance stress across layout.
• Ensure proper metal density to avoid local stress gradients.
 Design Guidelines for Aging

• Derate timing budgets to account for aging drift.

• Use guardbanding in clock and critical paths.
• Apply dynamic voltage scaling (DVS) to minimize stress during operation.
• Design with awareness of worst-case duty cycles and toggling activity.
 FinFET Reliability: Introduction

• FinFETs are 3D transistors with improved electrostatic control and scalability beyond 20nm.
• While offering performance benefits, they introduce new reliability challenges.
• Issues include variability, self-heating, and enhanced electric field effects on narrow fins.
 FinFET: BTI and HCI Challenges

• NBTI is more pronounced in FinFET PMOS due to higher gate fields and fin geometry.
• HCI mechanisms are altered by vertical gate structures and increased velocity saturation.
• Aging models must consider 3D gate control and multi-fin variations.
 FinFET: Self-Heating and Thermal Issues

• Fins act as thermal bottlenecks, leading to localized heat buildup.

• Impacts carrier mobility, Vth stability, and accelerates wear-out mechanisms.
• Requires accurate modeling of thermal resistance in design tools.
 FinFET: Process-Induced Variability

• Height and width variability of fins introduce Vth shifts and timing unpredictability.
• Line-edge roughness and gate work-function variation further degrade reliability.
• Mitigated via restrictive design rules and variability-aware timing closure.
 FinFET: Layout Techniques

• Use uniform fin counts for critical paths to minimize Vth variation.
• Maintain consistent fin orientation across cells.
• Isolate high-current devices with dummy fins or thermal spacing.
• Leverage DFM guidelines for multi-patterning and fin quantization.
 FinFET: Design Best Practices

• Perform reliability-aware synthesis and place-and-route.

• Use foundry-provided FinFET aging and self-heating models in timing closure.
• Verify reliability corners with dynamic IR-drop and temperature simulations.
• Follow reliability DRCs for minimum spacing and fin alignments.
 3D-IC Reliability: Introduction

• 3D-ICs stack multiple dies vertically to improve performance, density, and bandwidth.
• Introduce unique reliability challenges due to thermal, mechanical, and electrical interactions.
• Key concerns include TSV-induced stress, thermal hotspots, and inter-die bonding defects.
 TSV-Induced Stress and Impact

• Through-Silicon Vias (TSVs) introduce mechanical stress during fabrication and operation.
• Affects transistor characteristics near TSVs due to stress-induced mobility changes.
• Requires stress-aware placement and keep-out zones (KOZ).
 Thermal Management in 3D-ICs

• Power density increases with stacking, causing severe thermal gradients.

• Can accelerate EM, BTI, and TDDB in upper layers.
• Requires thermal-aware floorplanning and use of heat spreaders or microfluidics.
 Bonding Defects and Delamination

• Reliability affected by Cu-Cu or hybrid bonding imperfections.

• Bond voids or delamination can lead to increased resistance or open circuits.
• Detected via X-ray, IR imaging, or post-bond testing.
 3D-IC Layout Considerations

• Minimize TSV count and use clustered TSVs for thermal control.
• Use redundant TSVs and guards for high-reliability paths.
• Space TSVs away from sensitive analog/memory blocks.
• Apply TSV shielding to reduce noise coupling.
 3D-IC Design Guidelines

• Simulate thermal profiles during early planning.

• Use reliability-aware 3D physical design tools.
• Ensure timing closure includes inter-die delays and aging effects.
• Follow foundry-specific rules for bonding interface design and TSV reliability.
 Self-Heating Effect (SHE): Introduction

• SHE refers to local temperature rise within a transistor due to power dissipation.
• Prominent in advanced nodes and FinFETs due to poor heat dissipation paths.
• Elevated temperature accelerates wear-out mechanisms like BTI, HCI, and EM.
 SHE: Causes and Impact

• Caused by reduced thermal conductivity in shallow channels and fins.

• Increases junction temperature (Tj), degrading mobility and Vth.
• Affects timing, leakage, and accelerates device aging.
 Modeling Self-Heating

• Modeled via compact thermal models (e.g., RC network or Joule heating-based).

• Foundry models provide thermal-aware SPICE corners.
• Temperature rise (ΔT) often included in reliability simulations.
 SHE in FinFETs and GAA FETs

• FinFETs have limited heat sinking due to narrow fins and buried oxide.
• Gate-All-Around (GAA) FETs suffer more due to complete channel isolation.
• Accurate modeling is critical for high-frequency and high-density designs.
 Layout Techniques to Mitigate SHE

• Increase fin spacing to enhance heat dissipation.

• Use wider diffusion taps and thermal vias.
• Distribute power-hungry cells to prevent hotspots.
• Apply thermal-aware cell placement during floorplanning.
 Design Practices for SHE

• Derate timing and IR-drop analysis for worst-case temperature rise.

• Include dynamic thermal simulation during power planning.
• Use sleep modes and activity scheduling to reduce peak power.
• Choose package types with better thermal properties.
 Process Variation-Induced Reliability: Introduction

• Process variations refer to deviations in device parameters during fabrication.

• These include variations in line-edge roughness, oxide thickness, and dopant concentration.
• Variations impact device performance, power, and long-term reliability.
 Types of Process Variations

• Die-to-die (D2D): Differences across different dies on a wafer.

• Within-die (WID): Variations within a single die, including random and systematic effects.
• Random Dopant Fluctuation (RDF), Line Edge Roughness (LER), and Gate Length Variation (GLV) are major contributors.
 Impact on Reliability Mechanisms

• Accelerates NBTI, HCI, and TDDB due to local electric field and Vth variation.
• Increases stress on certain paths, reducing overall chip lifetime.
• Requires guardbands or redundancy to ensure functionality over time.
 Statistical Aging and Timing Effects

• Timing pessimism arises due to compounding of process and aging variations.

• Statistical Static Timing Analysis (SSTA) helps model impact.
• Requires Monte Carlo simulations for worst-case aging scenarios.
 Layout Techniques for Process Variation

• Use common-centroid layout for analog blocks to average out variations.

• Employ dummy structures to balance pattern density.
• Increase spacing and symmetry in critical matched components.
• Use redundant vias and tracks in metal routing.
 Design Best Practices

• Include variability-aware design closure and margining.

• Validate design with process corners, skew models, and statistical tools.
• Apply DFM rules for litho-friendly layouts.
• Use post-silicon calibration and tuning techniques for high-accuracy circuits.
 Advanced NBTI in FinFET PMOS: Introduction

• NBTI remains a critical aging mechanism, even in FinFETs.

• FinFETs exhibit reduced but still non-negligible NBTI degradation due to high electric fields.
• Requires updated modeling due to different channel geometry and oxide interface properties.
 NBTI Physics in FinFETs

• Occurs due to charge trapping at Si/SiON or high-k/metal gate interfaces.

• Fewer interface traps due to 3D gate structure, but increased field concentration at corners.
• Stress-induced leakage current (SILC) also contributes to long-term degradation.
 NBTI Models for Advanced Nodes

• Use Reaction-Diffusion (R-D) and Trap Generation models tailored for FinFET structure.
• Includes physical effects like discrete trapping, fin width variation, and corner stress.
• Foundries provide updated reliability models for FinFET technologies (e.g., 7nm, 5nm).
 NBTI Impact on Performance

• Vth shift leads to increased delay and dynamic power.

• Impacts SRAM cell stability and read margin.
• Requires guardbanding and statistical modeling for critical paths.
 Layout Techniques for NBTI in FinFETs

• Use symmetric layout to minimize mismatch.

• Employ balanced current paths in critical logic.
• Insert dummy fins to maintain uniform stress and doping.
• Use minimum fin pitch recommended by the foundry.
 Design Mitigation Strategies

• Dynamic Vth compensation using body bias or back gate.

• Use adaptive voltage scaling to reduce stress under low workload.
• Alternate stress and relaxation cycles during simulation to improve recovery.
• Choose high-N reliability corners for worst-case timing analysis.
 Reliability Metric Comparison Table

Issue Worsens With Primary Effect Common Layout Mitigation

Electromigration (EM) High Current Density Metal Line Damage Wider Metal Lines,
Redundant Vias
TDDB Voltage & Time Gate Oxide Breakdown Thicker Oxide, Guard Bands

BTI Voltage, Temperature Vth Shift Stress-Aware Sizing

HCI High Vds & Temp Gate Damage Lower Vds Design, Buffer
Insertion
ESD Charge Build-Up Junction Damage ESD Protection Diodes

Soft Errors Radiation Bit Flip ECC, Shielding

Thermal Runaway Heat, Leakage Self-Destruction Thermal-Aware Floorplanning

…THANK YOU…