3D/SiP Advanced Packaging Symposium Session II: Wafer Level Integration & Processing April 29, 2008 Durham, NC

Similar documents
Fabrication Process For A Novel High Speed Coplanar-to-Coaxial Off-Chip Interconnect

Fabrication Process For A Novel High Speed CoPlanar-to-Coaxial Off-Chip Interconnect

Off-Chip Coaxial to Microstrip Transition Using MEMs Trench

NOVEL DIE-TO-DIE COAXIAL INTERCONNECT SYSTEM FOR USE IN SYSTEM-IN-PACKAGE APPLICATIONS. Christopher Michael McIntosh

Compensation for Simultaneous Switching Noise in VLSI Packaging Brock J. LaMeres University of Colorado September 15, 2005

Physical RF Circuit Techniques and Their Implications on Future Power Module and Power Electronic Design

Optimization of Wafer Level Test Hardware using Signal Integrity Simulation

Characterization Methodology for High Density Microwave Fixtures. Dr. Brock J. LaMeres, Montana State University

Chapter 2. Literature Review

Technology in Balance

Signal Integrity Modeling and Measurement of TSV in 3D IC

Microcircuit Electrical Issues

Evaluation of Package Properties for RF BJTs

Flip-Chip for MM-Wave and Broadband Packaging

Session 4: Mixed Signal RF

Chapter 2 Low-Cost High-Bandwidth Millimeter Wave Leadframe Packages

DesignCon Design of Gb/s Interconnect for High-bandwidth FPGAs. Sherri Azgomi, Altera Corporation

A Simulation Study of Simultaneous Switching Noise

Figure 1. Inductance

Texas Instruments DisplayPort Design Guide

FDTD SPICE Analysis of High-Speed Cells in Silicon Integrated Circuits

EQCD High Speed Characterization Summary

Considerations in High-Speed High Performance Die-Package-Board Co-Design. Jenny Jiang Altera Packaging Department October 2014

EVALUATION OF THE NEAR-FIELD INJECTION METHOD AT INTEGRATED CIRCUIT LEVEL

Cost-minimized Double Die DRAM Packaging for Ultra-High Performance DDR3 and DDR4 Multi-Rank Server DIMMs

License to Speed: Extreme Bandwidth Packaging

How to anticipate Signal Integrity Issues: Improve my Channel Simulation by using Electromagnetic based model

CHAPTER 4. Practical Design

Aries Kapton CSP socket

Suppression Techniques using X2Y as a Broadband EMI Filter IEEE International Symposium on EMC, Boston, MA

Optimization of Symmetric Spiral Inductors On Silicon Substrate

Introduction: Planar Transmission Lines

Demystifying Vias in High-Speed PCB Design

DesignCon Full Chip Signal and Power Integrity with Silicon Substrate Effect. Norio Matsui Dileep Divekar Neven Orhanovic

Substrate Coupling in RF Analog/Mixed Signal IC Design: A Review

Understanding, measuring, and reducing output noise in DC/DC switching regulators

1 FUNDAMENTAL CONCEPTS What is Noise Coupling 1

MMIC/RFIC Packaging Challenges Webcast (July 28, AM PST 12PM EST)

Introduction to VFTLP+

Custom Interconnects Fuzz Button with Hardhat Test Socket/Interposer 1.00 mm pitch

Comparison of IC Conducted Emission Measurement Methods

Design and Modeling of Through-Silicon Vias for 3D Integration

Impedance Matching: Terminations

AWR. SIP Flow White Paper UNDERSTANDING AVAILABLE TOOLS FOR RF SYSTEM-IN-PACKAGE AND MULTI-CHIP-MODULE DESIGN AND OPTIMIZATION

Wafer-scale 3D integration of silicon-on-insulator RF amplifiers

RECENTLY, interest in on-chip spiral inductors has surged

Microprobing with the Agilent 86100A Infiniium DCA

Taking the Mystery out of Signal Integrity

Low Jitter, Low Emission Timing Solutions For High Speed Digital Systems. A Design Methodology

Signal and Noise Measurement Techniques Using Magnetic Field Probes

--- An integrated 3D EM design flow for EM/Circuit Co-Design

Characterization of Alternate Power Distribution Methods for 3D Integration

The Challenges of Differential Bus Design

The Inductance Loop Power Distribution in the Semiconductor Test Interface. Jason Mroczkowski Multitest

Parallel vs. Serial Inter-plane communication using TSVs

GaAs MMIC Non-Linear Transmission Line. Description Package Green Status

High Speed Digital Systems Require Advanced Probing Techniques for Logic Analyzer Debug

SHELLCASE-TYPE WAFER-LEVEL PACKAGING SOLUTIONS: RF CHARACTERIZATION AND MODELING

DesignCon Effect of Power Plane Inductance on Power Delivery Networks. Shirin Farrahi, Cadence Design Systems

The Impact Of Signal Jumping Across Multiple Different Reference Planes On Electromagnetic Compatibility

Through-Silicon-Via Inductor: Is it Real or Just A Fantasy?

3D IC-Package-Board Co-analysis using 3D EM Simulation for Mobile Applications

Fuzz Button interconnects at microwave and mm-wave frequencies

High Data Rate Characterization Report

CIRCUITS. Raj Nair Donald Bennett PRENTICE HALL

Microcontroller Systems. ELET 3232 Topic 13: Load Analysis

DesignCon Impedance Matching Techniques for VLSI Packaging. Brock J. LaMeres, Agilent Technologies, Inc. Rajesh Garg, Texas A&M University

Introduction to On-Wafer Characterization at Microwave Frequencies

Compact Distributed Phase Shifters at X-Band Using BST

VLSI is scaling faster than number of interface pins

Objectives of transmission lines

Design Guide for High-Speed Controlled Impedance Circuit Boards

The Design Challenge to Integrate High Performance Organic Packaging into High End ASIC Strategic Space Based Applications.

Synthesis of Optimal On-Chip Baluns

Integration Techniques for MMICs and Chip Devices in LTCC Multichip Modules for Radio Frequencies

Advanced Transmission Lines. Transmission Line 1

insert link to the published version of your paper

The Advantages of Integrated MEMS to Enable the Internet of Moving Things

A Simulation Methodology for Wirebonds Interconnects of Radiofrequency Integrated Circuits

Digital Systems Power, Speed and Packages II CMPE 650

Infinity Probe Mechanical Layout Rules

Silicon Interposers enable high performance capacitors

DesignCon East Feasibility of 40 to 50 Gbps NRZ Interconnect Design for Terabit Backplanes

Design of the Power Delivery System for Next Generation Gigahertz Packages

Smart Devices of 2025

HV739 ±100V 3.0A Ultrasound Pulser Demo Board

2.5D & 3D Package Signal Integrity A Paradigm Shift

A Miniaturized Multi-Channel TR Module Design Based on Silicon Substrate

Design for EMI & ESD compliance DESIGN FOR EMI & ESD COMPLIANCE

Chapter 2. Inductor Design for RFIC Applications

Lecture 020 ECE4430 Review II (1/5/04) Page 020-1

DesignCon 2003 High-Performance System Design Conference (HP3-5)

Lecture 020 ECE4430 Review II (1/5/04) Page 020-1

TOP VIEW MAX9111 MAX9111

High-Speed PCB Design Considerations

PRELIMINARY PRELIMINARY

POSSUM TM Die Design as a Low Cost 3D Packaging Alternative

Gain Slope issues in Microwave modules?

A Technical Discussion of TDR Techniques, S-parameters, RF Sockets, and Probing Techniques for High Speed Serial Data Designs

Chapter 7 Introduction to 3D Integration Technology using TSV

Transcription:

3D/SiP Advanced Packaging Symposium Session II: Wafer Level Integration & Processing April 29, 2008 Durham, NC Off-Chip Coaxial to Coplanar Transition Using a MEMS Trench Monther Abusultan & Brock J. LaMeres Montana State University Bozeman, MT 1

Problem Statement System Performance in VLSI Designs is Limited by Package Interconnect 1) Signal Path Reflections - Unwanted Switching - Edge Speed Degradation 2) Signal Coupling - NE/FE Cross-talk - Power Supply Droop - Ground Bounce On-Chip Performance is outpacing Off-Chip interconnect 1) Emerging problem of getting high speed signals from chip-to-chip 2) This problem will continue as transistors keep getting faster 2

Why is packaging limiting performance? Today s Package Interconnect Looks Inductive - Long interconnect paths - Large return loops Wire Bond Inductance (~2.8nH) L I Today s Package Impedance is Not Controlled or Shielded 3

The Trend Toward System in Package (SiP) Moving more functionality on package reduces the amount of times a signal needs to traverse level 2 interconnect (package-to-pcb) Integrating functionality onto a single IC has limitations: - Reduced yield, suboptimal material selection (CMOS vs. GaAs vs. SiGe) Integrating multiple die onto the same package with wire bonds is an optimal balance However, we're back to the problem of unshielded, uncontrolled wire bonds www.amkor.com 4

Proposed Solution A New Chip-to-Chip Interconnect Technology Off-Chip Coaxial Launch - Exploit Advances in MEMS process technology - Target System in Package (SiP) applications 5

Proposed Solution A New Chip-to-Chip Interconnect Technology Application - High speed chip-to-chip signals require controlled impedance and shielding - Additional process step converts perimeter wire-bond pads to coaxial launch. - Step 1: Design, Model, and Fabricate interconnection between side-by-side die - Step 2: Investigate Vertically Stacked Die Interconnect 6

Proposed Solution A New Chip-to-Chip Interconnect Technology Processing - Etch a trench into the Silicon substrate to hold the coaxial cable - The center conductor is connected to a signal trace on-chip - A coplanar transmission line is used on-chip to provide connection to the signal and to the coaxial ground shield. 7

Coaxial to Coplanar Launch Geometric Dependencies - Coaxial Line - The coax outer diameter is the key dimension - Our design evaluations Semi-Rigid Coax's from Micro-Coax (UT-013, UT-020) - 50 impedance requirement sets coaxial dimensions - Extension diameters dictated by mechanical reliability 8

Coaxial to Coplanar Launch Geometric Dependencies - Coplanar Line - The ground separation is dictated by the outer diameter of the coaxial line G S G - 50 impedance set by material properties & signal trace width NOTE: On-Chip Coplanar Transmission Line has: 1) Imaginary Impedance due to Lossy Semiconductor Material 2) Potential for higher-order modes in addition to TEM 9

Coaxial to Coplanar Launch Geometric Dependencies - Trench - The trench must be wide enough to accept the coaxial outer diameter - The depth must place the coaxial center conductor on top of the coplanar signal trace - Using inscribed octagonal geometries sets width of trench - Anisotropic etch rate dictates angle of trench sidewall. 10

Coaxial to Coplanar Launch Geometric Dependencies Channel Spacing - Spacing of adjacent trenches must accommodate coax protrusion 11

Coaxial to Coplanar Launch Geometric Dependencies Transition Region - Length of sidewall dictated by anisotropic etch rate. - Overlapping lengths dictated by mechanical reliability 12

Coaxial to Coplanar Launch Summary of Dimensions - 2 Micro-Coax s are evaluated (UT-013, UT-020) - Each coax size influences the trench and coplanar transmission line dimensions Region Parameter Units Coaxial Line UT-013 UT-020 Coaxial Structure D oc m 330 584 Coplanar Structure D od m 254 419 D cc m 79 127 T sig m 1 1 T ox m 0.8 0.8 W sig m 239 446 W gnd m 100 100 S copl m 55 90 S ss m 634 916 Trench Structure W ttop m 349 626 W tbot m 150 228 W tsw m 100 169 H tsw m 141 239 Transition Region L trench m 1100 1170 L dext m 500 500 L sw m 100 170 L cext m 1000 1000 L ccov m 900 831 13

Coaxial to Coplanar Launch Impedance Discontinuities - Between the coax and coplanar T-lines, there are regions of impedance discontinuities - These add reflections and risetime degradation between the two ideal transmission line structures (i.e., the coaxial and coplanar lines) 14

Modeling Approach EM Field Solvers - Due to the complexity of the structure, a field solver is used to extract the characteristic impedance (Z 0 ) and propagation constant (g) -Z 0 andgare complex for signal propagation on the integrated circuit due to the use of a semiconductor substrate material. -Z 0 is real inside of the coaxial transmission line - We used Electromagnetic Design Systems (EMDS) from Agilent Technologies to perform 2D and 3D field simulations 15

Modeling Approach Our Approach 1) Extract Z 0 andg for each different Cross-Section within the transition using a 2D simulation 2) Import parameters into SPICE to perform transient simulations on the structures ability to transmit high speed signals 16

Modeling Results (XC1) Field Solder Results Z 0 = 52 + j26 g = 305 + j615 GND SIG Symmetry Axis GND SIG Symmetry Axis Electric Fields Magnetic Fields 17

Modeling Results (XC2) Field Solder Results Z 0 = 50 + j25 g = 299 + j604 GND SIG Symmetry Axis GND SIG Symmetry Axis Electric Fields Magnetic Fields 18

Modeling Results (XC3) Field Solder Results Z 0 = 114 + j3 g = 10 + j269 SIG SIG GND Symmetry Axis GND Symmetry Axis Electric Fields Magnetic Fields 19

Modeling Results (XC4) XC4 Field Solver Results Z 0 =128 + j1 g = 4 + j239 r1 r2 rms SIG SIG GND Symmetry Axis GND Symmetry Axis Electric Fields Magnetic Fields 20

Modeling Results (XC5) XC5 Field Solver Results Z 0 = 134 + j1 g = 3 + j229 r1 r2 SIG SIG GND Symmetry Axis GND Symmetry Axis Electric Fields Magnetic Fields 21

Modeling Results (XC6) XC6 Field Solver Results Z 0 = 111 + j1 g = 5 + j276 rc r1 r2 SIG SIG GND Symmetry Axis GND Symmetry Axis Electric Fields Magnetic Fields 22

Modeling Results (XC7) XC7 Field Solver Results Z 0 = 50 + j0 g = 0 + j296 rcrc r2 GND SIG Symmetry Axis GND SIG Symmetry Axis Electric Fields Magnetic Fields 23

Modeling Results (XC8) XC8 Field Solver Results Z 0 = 50 + j0 g = 0 + j296 rcrc SIG SIG GND Electric Fields GND Magnetic Fields 24

Modeling Results Field Solver Results Summary Region Z 0 g XC1 52 + j26 305 + j615 XC2 50 + j25 299 + j604 XC3 114 + j3 10 + j269 XC4 128 + j1 4 + j239 XC5 134 + j1 3 + j229 XC6 111 + j1 5 + j276 XC7 50 + j0 0 + j296 XC8 50 + j0 0 + j296 Electric Fields Magnetic Fields 25

Electrical Evaluation (Comparison to Wirebond) - Comparing to a chip-to-chip application where coplanar lines are used on-chip (35ps risetime) Signal Path 1: Using a G/S/G wirebond interconnect structure Signal Path 2: Using the new coaxial launch structure 26

Spatial Evaluation Wire bond Comparison - Is this interconnect comparable in size to that of the pads for wire bonding? - We evaluate against 100 m x 100 m pad requirements for wire bond in G-S-G configuration with 100 m spacing Results Region Parameter Units Coaxial Line UT-013 UT-020 Coaxial Structure D oc m 330 584 Coplanar Structure D od m 254 419 D cc m 79 127 T sig m 1 1 T ox m 0.8 0.8 W sig m 239 446 W gnd m 100 100 S copl m 55 90 S ss m 634 916 Trench Structure W ttop m 349 626 W tbot m 150 228 - Wire Bond Pads for G-S-G: = 3*(W pad ) + 2*(W space ) = 3*(100 m) + 2*(100 m) = 500 m - Coaxial Launch for G-S-G: W tsw m 100 169 H tsw m 141 239 Transition Region L trench m 1100 1170 L dext m 500 500 L sw m 100 170 L cext m 1000 1000 L ccov m 900 831 = W ttop + 2*W gnd = 349 m + 2*100 m = 549 m only 9.8% more area required 27

Electrical Evaluation (Parasitics) Electrical Evaluation - Interconnect comparison Coax length = 3mm Wire bond length = 3mm Results - Versus wire bond: Parameter Units Wire Bond Coaxial Line L' nh/m 569 242 C' pf/m 26 97 Zo 148 50 L 3mm nh 1.71 0.73 C 3mm pf 0.08 0.29 Inductance reduced by 57% Impedance reduced by 66% - Note: Interconnect is now Shielded and has Controlled Impedance 28

Electrical Evaluation (TDR/TDT) Wirebond = 34% reflection Coax Launch = 8% reflection 29

Electrical Evaluation (Eye Diagram) Eye Diagrams of a 5Gb/s, PRBS for a Load Terminated System Wire bond Coplanar to Coax 30

Summary 1) A new SiP interconnect was presented and compared to current technology - Coaxial to Coplanar launch using MEMS trench - Selective processing for high-speed nets 2) Spatially this interconnect takes similar area requirements for G:S:G 3) Electrically this interconnect has the potential to perform faster - Controlled impedance reduces reflections - Shielded interconnect eliminates signal coupling 4) Next Steps - Fabrication underway at Montana State - Measurements on prototypes expected during summer of 2008 31

Questions? 32